Home

User's Guide - solutionmetrics.com.au

Contents

1. 10 Intensity Intensity o o s ma oo o o C a o o ooma av aamemf ee 5 oxo com awam omn a _ _ of 8 CTAMTAMK AMT ADMDTAM CANMTAMK AMT ADMDTAM SEESSSR33333333 SSEEBS5532333333233 FETForcer HH OFS GCceS TITIOCVCTCMMNOTITIOCCES EEEZFZSZEEEEEESES EEEZZZEEEEEES23 Ss 3 t t 2 t t t 2 2 2 2 t t t BGRSEESESEGGEEE GARELESSSSGAGEEE eo0oss8s sss eooss8s sss ZZZNNN nnn ZZZNNN nnn OoOO oOoo0OoO0O Zz Z Z A Figure 2 21 Pre and post normalized boxplots of the swimming mice data Additional MvA plots are generated for the other arrays both before and after normalization but they are not displayed here For this example we ll look at the analysis in two ways The first applies a classical one way analysis of variance ANOVA across the different levels of the factor When setting up the ANOVA we specify the type contrasts to test so we can produce the comparisons described in Table 2 2 An alternate way to think of the analysis is as a set of two sample problems as described in Table 2 2 This allows us to apply two sample methods that may be more sensitive when there is little replication in the design The first analysis uses the ANOVA dialog
2. Cancel Apply k current Help Figure 3 10 The Normalization dialog Select Affymetrix CEL from the Show Data of Type drop down list and choose the cgAffyBatch data object Explore the normalization procedures available in the Normalization drop down list select the quantiles procedure to normalize both PM perfect match and MM mismatch intensities or the PM intensities only For this example select PM and MM to normalize both Save As The Save As field takes an object name for saving the normalized affyBatch probe level object with the default name cgAffyBatch norm Clicking OK creates the normalized affyBatch object and plots pre and post normalization boxplots for comparison The plot is on a logy scale but the expression intensities are saved on the original raw intensity scale Expression Summaries Two Sample Design Before quantiles Normalization After quantiles Normalization 4 o a o o Q o J x 4 nl No F r 1 T T i i ol aj il i l o4 e e e e e e e fe a el Sho ee ee tll i P od a ANN odl odi K Q Q Oo peo is Q N N vr r N N 4 D D N N D D N N S 9 g 8 8B Figure 3 11 Pre and post normalized probe level data The normalized int
3. Color Map J white black Ps Plot Type C Scatter plot Hex bin Saniple Size 2 fo Number of bins 20 M Other IV Intensity Boxplot Box Plot Type C By Aray By Print tip IV Prin Comp Plot Cancel Apply i curent Help Figure 4 20 Quality Control Diagnostics Two Channel dialog setup Two Way Reference Design Green Foreground for TP_01a gpr 61000 54000 48000 41000 34000 27000 20000 14000 6900 91 63000 56000 49000 42000 35000 28000 21000 14000 7000 51 Figure 4 21 Image plots of red and green foreground channels Hexbinned MvA Plots Ref LC 1 1 LC Ref 1 2 2 Hundreds al 13579 Tens 5 10 15 5 10 15 13979 A A Ones Ref LC 7 3 LC Ref 7 4 13579 Figure 4 22 Hexbinned MvA plots for four of the malaria arrays The lines overlaid indicate locally weighted mean value with confidence bounds 153 Chapter 4 Examples Two Color Data Print Tip Intensity Box Plot for Ref LC 1 1 s sil 3 2 a m H O 3 P ma pri r DoT i om i j ho H F i i i f i F E f F H H i R 7 o be ii g wt ae i L s r 3 i Ei Fs w e A i i 3 4 T ee ee a a T e b 2 a a a a a a a a x g x x 2 2 2 2 x g x x PrintTip Figure 4 23 Intensity boxplot by print tip
4. GOBPCHILDREN GOBPPARENTS GOCCCHILDREN GOCCPARENTS GOGO2LL GOMFCHILDREN GOMFPARENTS GOTERM Annotation data files for GO KEGGENZYMEID2GO KEGGEXTID2PATHID KEGGGO2ENZYMEID KEGGPATHID2EXTID KEGGPATHID2NAME KEGGPATHNAME2ID Annotation data files for KEGG lt species gt CHRLOC lt x gt START lt species gt CHRLOC lt x gt END lt species gt CHRLOCCYTOLOC lt species gt CHRLOCLOCUSID2CHR Chromosome location start and end for lt x gt 1 22 x y for lt species gt human rat and mouse Supporting cytoband and locus link information lt species gt LLMappingsGO2LL lt species gt LLMappingsLL2GO lt species gt LLMappingsLL2PMID lt species gt LLMappingsLL2UG lt species gt LLMappingsPMID2LL lt species gt LLMappingsUG2LL Locus Link mappings for lt species gt human rat and mouse homology Various homology annotation data files and mappings For non Affymetrix two channel chips one can construct similar S PLUS annotation objects for use in StARRAYANALYZER by using the lt layoutFileName gt rather than the Affymetrix chip name as the 337 Chapter 9 Annotation and Gene List Management Annotation from Graphical and Tabular Reports GenBank Sites 338 prefix in the annotation object For example if the layout file name used in the data import is AgilHuv2 the corresponding accession numbers could be simply inserted into an S PLUS data object AgilHuv2ACCNUM The format for
5. National PublMed ne Protein NCBI Search PubMed Nucleotide Genom Structure OMIM PMC hdj Limits Details Preview Index History De Summary Stow ENZ Sor Se Tee Items 1 20 of 55 Page 1 of 3 Next Related Articles Links Clipboard About Entrez Text Version F 1 Brigstock DR Regulation of angiogenesis and endothelial cell function by connective tissue growth factor CTGF and cysteine rich 61 CYR61 Angiogenesis 2002 5 3 153 65 PMID 12831056 PubMed indexed for MEDLINE Nishida T Kubota S Fukunaga T Kondo S Yosimichi G Nakanishi T Takano Yamamoto T Takigawa M CTGF Hes24 hypertrophic chondrocyte specific gene product interacts with perlecan in regulating the proliferation and differentiation of chondrocytes J Cell Physiol 2003 Aug 196 2 265 75 PMID 12811819 PubMed indexed for MEDLINE Utsugi M Dobashi K Ishizuka T Masubuchi K Shimizu Y Nakazawa T Related Articles Links Mori M Related Articles Links C Jun NH2 terminal kinase mediates expression of connective tissue growth factor induced by transforming growth factor beta in human lung fibroblasts Internet Figure 9 22 Pubmed articles for the four genes identified in the gene filtering analysis described above GO categories can be used to filter data to particular gene subsets of interest before differential expression testing and or other analyses For example in a study invol
6. OK Cancel Apply k current Help Figure 4 14 The Two Sample Test dialog for the Swir1MarrayRaw expression object The Options group allows you to specify the statistical test and alternative hypothesis a procedure for controlling the family wise error rate FWER or the false discovery rate FDR and the maximum error rate Additionally certain procedures estimate statistical tests and p values by permutation sampling For these procedures you can also set the maximum number of permutations and a random seed for reproducibility and testing FWER and FDR The Options group allows you to set the family wise error rate FWER or the false discovery rate FDR to maintain an overall Type I error rate false positive rate based on adjusting individual test p values to account for multiple tests In our swirl mutant example there are 8448 genes so the increase in Type I error is substantial without adjusting the p values 143 Chapter 4 Examples Two Color Data Tests Alternative Hypothesis Adjustment 144 You can chose the type of test you want to perform paired t the default for two color arrays Welch s t for unequal variance an equal variance t and the nonparametric Wilcoxon test For the swirl example leave the setting as the default paired t test The Alternative Hypothesis drop down list provides three options 1 Greater than 2 Less than 3 Not equal default These hypotheses refer to the alt
7. 1 The number of header lines to skip in each file before reading the data Normally this can be detected automatically but it is provided as an option for unusual cases where auto detection can not find the row with column names 2 The delimiter separating the fields in each line of the data files Normally this can be detected automatically but it is provided as an option for unusual cases where auto detection can not determine the field delimiter Press OK when you have completed the dialog and the data are imported It is now ready for use in S ARRAYANALYZER Quality Once the data is imported open the Quality Control Diagnostics Control dialog to do quality checks Open the dialog by clicking through the sequence ArrayAnalyzer Quality Control Diagnostics gt Diagnostics Affymetrix from the main S PLUS menu bar ArrayAnalyzer Import Data gt Quality Control Diagnostics Le Filtering Two Channel Affymetrix Expression Summary Normalization Differential Expression Analysis gt Cluster Analysis Annotation Gene List Management Figure 3 30 Opening the Quality Control Diagnostics dialog The resulting dialog provides a number of options for doing visual quality checks 1 Image Plot An image plot of each array For Affy CEL data each pixel corresponds to a different spot on the array 101 Chapter 3 Examples Affymetrix Probe Level Data 102 2 MvA Plot of replicates versus ea
8. ANOVA Dialog One Way Design We are now ready to do differential expression analysis for the one way design From the main menu open the ANOVA dialog by clicking ArrayAnalyzer gt Differential Expression Analysis gt ANOVA The procedures implemented in this dialog provide traditional ANOVA methods with a host of correction procedures to control the Type I Error rate For more details on the ANOVA dialog and Type I Error correction procedures see Chapter 7 Differential Expression Testing To set up the ANOVA dialog follow these steps 1 In the Show Data of Type field select Affymetrix 2 In the Data field select MouseSwimExprSet norm 3 The Array Name field should be automatically updated to mgu74av2 4 In the Contrasts group confirm that the Baseline check box is checked 5 From the Baseline Level drop down list select NoSwim4wks 6 In the Save As field in the lower right hand corner of the dialog type MouseANOVANoSwim4wks 7 Click OK to compute the ANOVA results Once the NoSwim4wks baseline analysis completes re run it with baseline set to NoSwim4wks1wk The additional analyses allow us to assemble all the significant genes for the contrasts listed in Table 2 2 for both of the baseline levels 41 Chapter 2 Examples Affymetrix MAS Data Options 42 amix m Data r Options Show Data of Type FWER FDR 0 05 att metriz zi IV Protected Data MouseSwimE xp Y Adjustment BH Array Name
9. Figure 2 4 The Create Modify Design dialog with one way experiment setup Step 2 Associate Once the design is complete click OK to copy it into the File Files With Design Selection tab of the Import Data From Affymetrix dialog Notice Points that the number of rows for the File Selection box is modified to match the number of arrays specified on the Create Modify Design dialog Furthermore values for the factor levels have been written into the Factor columns to facilitate associating files with design points If the experiment is balanced the factor level settings will be exactly as needed However the level values can be reset when the experiment is unbalanced or if you prefer an order different from the default 25 Chapter 2 Examples Affymetrix MAS Data 26 The next step is to associate files with each design point To do so right click in one of the file fields and browse to the location of your files Import Data From Affymetrix E x File Selection MIAME MAS Variables amp Filtering CEL Filtering Options r Step 1 Specify Design Read Existing Design Create Modify Design Save Design File r Step 2 Associate Files with Design Points Factor CondTime NoSwim4wks NoSwim4wks File Name C Program Files Insightf NoSwim4w1 txt NoSwimd w1 C Program Files Insightf NoSwim4w2 txt NoSwim4w2 C Program Files Insightf NoSwim4w3 tx
10. I Significant genes SurgeryE xprS ehd I Genes with maximum fold change Limit number of genes to j greater thani 2 a Recalculate I Genes with Expression values Number of genes selected by filtering exceeding fio fi 5 in at least experiment chips x Sort Order a E H Move Up Move Down OK Cancel Apply Id gt f current Help o Figure 3 49 Filtering Options settings for annotation of the genes identified by the SurgeryANOVABs YoungBH ANOVA Clicking OK now generates the annotation lists from the databases selected Figure 3 50 displays a summary table for the top significant genes identified by the ANOVA tests for the 1hr Old Young contrast 121 Chapter 3 Examples Affymetrix Probe Level Data 7 lt gt NCBI Entrez PubMed About Entrez Help FAQ Entrez Tools on LinkOut Submit to GenBank h for full length Display Summary Show 20 Send to Tex 7 One eo sh Nucleotide Nucleotide Protein Genome Structure PMC Taxonomy Books Search Nucleotide for A1836406 A1854614Xx68670 AV309347 L04503 Go Clear Limits Preview Index History Clipboard Details Items 1 5 of 5 page F 1 AI836406 Links UI M AP0 abi g 05 0 UIL s1 NIH_BMAP MST Mus musculus cDNA clone UI M AP0 abi g 05 0 UI 3 MRNA sequence 54706 19 gb A1836406 1 5470619 F 2 41854614 Links VI M BH0 ake d 06 0 UIL s1 NH_BMAP_M_51 Mus musculus cDNA clone
11. RNA Degrad Plot Plot of RNA degradation Only available for probe level CEL data Principal Components Plot Plot of the first two principal components using treatment combinations i e expression intensities for the entire array as variables in the principal components analysis One Way Design To generate the diagnostic plots select Affymetrix Summary in the Show Data of Type drop down list and select MouseSwimExprSet For the image plot set the Page Layout to 4 arrays per page Select the Genes Present Plot the Intensity Boxplot and the Prin Comp Plot check boxes Quality Control Diagnostics Affymetrix 5 oj x m Data r MvA Plot Bland Altman Show Data of Type IV Mv Plot aty meti Sumr z Formed From C Median Data MouseSwimE xp z Replicate Array Name mgu74av2 I Smooth Curves Image Plot Plot Type C Scatter plot IV Image Plot u p Hex bin Page Layout 4 arrays per pad Color Map blue yelow fad sample Size J100 i Number of bins 20 r Other Plots I Genes Present Plot IV Intensity Boxplot J ANA Degrad Plot IV Prin Comp Plot Cancel Apply K current Figure 2 12 Specifying options on the Quality Control Diagnostics dialog for the MouseSwimExprSet data set The following set of graphs are examples of the QC Diagnostic plots for MouseSwimExprset data set Figure 2 13 displays the image plots for four of the arrays in the study Each gene
12. defense response The complete ANOVA output for this analysis includes a summary page and parallel coordinates and volcano plots for each contrast 367 Chapter 9 Annotation and Gene List Management 368 APPENDIX A CREATING A DESIGN FILE Introduction Format Specification Special Keys ImportInfo valuenames FactorInfo valuenames Design valuenames Design Type valuenames Example Design File for Affymetrix MAS CEL Example Design File for Two Color Experiments 370 371 371 375 377 378 380 382 383 369 Appendix A Creating a Design File INTRODUCTION 370 Design files are used to describe the experimental conditions of a microarray analysis Typically you list the number of arrays to be read the number of factors in the experiment one or two are currently allowed and the name number of levels and level values for each factor In S ARRAYANALYZER you can enter your data files in the Import Data From Affymetrix or Import Data From Two Color dialogs and associate the files with a particular name level and factor but this may be inefficient if you have a large number of files or you plan on analyzing a number of experiments with a limited number of changes in each Creating a design file to import your data files is often simpler and you can create this design file in your favorite text editor and click the Read Existing Design button to read the design file Design files written by St ARRAYANALYZ
13. DEA M Dye Swap Reference Factor RefLC Z C Other Ref Leve Ref 7 C Rec3 C Rec Dye Swap Cancel Help Figure 4 17 The Create Modify Design dialog for the Malaria Parasite data Now click OK on the Create Modify Design dialog to copy the design into the file selection grid and associate the files which each design point You can find the malaria parasite example data by navigating to your splus62 module ArrayAnalyzer examples directory selecting GPR gpr as the Files of type and selecting TP_0la gpr Repeat for the other 10 gpr files entering one file per cell Alternatively read the design file TPDesignFile txt from the examples directory The Malaria Parasite data TP files are not installed by default You will only find the TP example files in the examples directory if you have done a Full install of S ARRAYANALYZER The TP dataset requires about 300MB of disk space Creating The Layout The create the layout Click the Create Array Layout button and browse to the examples directory for TPLayout gal Selecting Files of type GAL when you are browsing will make it easier to find TPLayout gal The other settings are as displayed in Figure 4 18 149 Chapter 4 Examples Two Color Data Selecting Expression Variables Selecting Filtering Variables 150 Create Layout File Selection Options Scanner Layout C Program Files Insightfulsplus62 module ArrayAnalyze
14. DISPLAYNAME c PF ID stringsAsFactors F Now we are ready to re run the analysis described in section Differential Expression Analysis on page 164 The significant genes in the summary table and the volcano plot are hyper linked to the DeRisi Lab Malaria Transcriptome Database Figures 4 39 4 40 and 4 41 display the results Summary Output for LPE Test with BH Adjustment Top 15 Genes Test Stat Rawp Val Adj p Val Fold Change POGUE 81 lt 0 001 lt 0 001 4 89 a 09 0 00 0 00 5 58 M56512_4 7 02 lt 0 00 lt 0 001 4 8 13638_1 6 72 lt 0 001 lt 0 00 4 87 C191 6 39 lt 0 001 lt 0 00 4 85 oPFM60552 6 31 lt 0 00 lt 0 001 4 31 F60188_2 6 31 lt 0 00 lt 0 001 4 31 aM43418_1 6 31 lt 0 00 lt 0 00 4 31 J152_17 6 31 lt 0 00 lt 0 00 4 31 E18606_4 6 31 lt 0 001 lt 0 001 4 31 C191 6 31 lt 0 00 0 00 4 31 oPFBLOBO111 6 31 lt 0 00 lt 0 00 4 31 N132_162 6 31 lt 0 001 lt 0 001 4 31 N134_130 6 31 lt 0 00 lt 0 001 4 31 N134_139 6 31 lt 0 001 lt 0 00 4 31 Figure 4 39 Hyperlinking from the Top 15 Gene List to the DeRisi Lab Malaria Transcriptome Database 171 Chapter 4 Examples Two Color Data Gene Name oPFGO060 Volcano Plot Probe Id PFCO186 T 3 gt a o D 3 a gt al o 2 T D z Mean Log2 Fold Change Figure 4 40 Hyperlinking from the Volcano Plot to the DeRisi Lab Malaria Transcriptome Database 172 Two
15. S ARRAYANALYZER functions need to access the named lists in these libraries when doing probe level operations If the lists are not available S tARRAYANALYZER attempt to load the needed library if it cannot find the library an error or warning message will appear The CDF and probe set information for Affymetrix chips is available on the StARRAYANALYZER Web site 233 Chapter 6 Pre Processing and Normalization Affymetrix Diagnostic plots 234 http www insightful com support Array Analyzer DataLibs README html There is a CDF zip file for each chip and each zip file unpacks to create a library These libraries include the probe set information listed above that is required by GC RMA Please refer to the section CDF and Probe Libraries for more details on loading additional libraries Both MvA plots and box plots are available from the expression summary and normalization dialogs under the S ARRAYANALYZER menu Affymetrix data uses one treatment condition per chip Comparisons can be made between treatments and within treatments To compare expression within treatment conditions the intensity log ratio M and the average intensity A are commonly defined as follows M log Ek Ek ae log EkiE k 2 where Eis the expression intensity for replication i i 7 i j 1 of replications of the kth treatment To compare expression between treatment conditions the intensity log ratio M and the average intensity A
16. Table 4 1 Experimental design and file association for the swirl cDNA experiment Cy3 Cy5 Array File Name swirl wild type 1 swirl L spot swirl wild type 3 swirl 3 spot wild type swirl 2 swirl 2 spot wild type swirl 4 swirl 4 spot 127 Chapter 4 Examples Two Color Data Importing Data Import Data From Two Channel Dialog 128 To import two color cDNA data go to the main S PLUS menu and select ArrayAnalyzer gt Import Data gt From Two Channel ArrayAnalyzer Import Data From Affymetrix Quality Control Diagnostics Filtering Affymetrix Expression Summary Normalization Differential Expression Analysis gt Cluster Analysis Annotation Gene List Management Figure 4 1 Menu selection to import two color cDNA data This launches the Import Data From Two Channel dialog with the File Selection page displayed The primary task of the import process is to associate data files with experimental conditions and select the variables and corresponding columns that are used in the data analysis Import Data From Two Channel xi File Selection MIAME Variable Selection amp Filtering Options r Step 1 Specify Design Read Existing Design Create Modify Design Save Design File r Step 2 Associate Files with Design Points Factor Level Cy3 cys Array Layout NIEHSLayout E Create Array Layout r Step
17. 0 0002952891 1 06673955 2 0 0004345541 1 0 6673955 3 0 0005307666 1 0 6673955 4 0 0005736545 1 0 6673955 5 0 0006389848 1 0 6673955 6 0 0007114647 1 0 6673955 7 0 0007657569 1 0 6673955 8 J 0 0009233798 1 D 6623955 9 1 0 0009371951 1 DO 6673955 10 0 0010859229 1 0 6673955 We create several objects that get passed to the plotting functions adjp lt testObj adjp index lt testObj index JHHF Set up for Bonferroni adjustment plot testDF lt data frame GeneName sw probes testStat testStat testStat foldChange foldChange stringsAsFactors F row names testDF lt sw probes testDF lt testDF Lindex testDFL rawp lt rawpLindex testDFL AdjPvalue lt adjpL Bonferroni JHHF Setup for BH adjustment plot testDFBH lt data frame gName sw probes testStat testStatL testStat foldChange foldChange stringsAsFactors F row names testDFBH lt sw probes testDFBH lt testDFBHLindex testDFBHL rawp lt rawpLindex testDFBHL AdjPvalue lt adjp BH From the Command Line Volcano plots volcanoPlot testDF fwer 5 volcanoPlot testDFBH fwer 5 Graphlets Now the Graphlets The dif fExpr object contain raw p values and adjusted p values for each procedure diffExpr lt data frame adjp row names diffExpr lt sw probes index dHHE Bonferroni adjustment MultSumm lt multtest graphlet M diffExprL 2 diffExpr 1 index testStat 1
18. 205 Chapter 5 Quality Control Diagnostics and Filtering The resulting Gene Filtering tab of the Filtering dialog is displayed in Figure 5 13 Figure 5 13 Filtering genes with non zero flags 206 PRE PROCESSING AND NORMALIZATION Introduction 209 Normalization 210 Technical Sources of Variability 210 Why do We Normalize Data 211 Normalizing in S ArrayAnalyzer 212 Ideas in Normalization 214 Normalizing to One Point 214 Normalizing to Many Points 215 Cautions in Normalizing 216 Workflow 217 Diagnostic Plots 218 Box Plots 218 MvA Scatter Plots 218 MvA Hexbin Plots 219 Chip Specific Plots 219 Normalization Methods for Two Channel Data 220 Normalizing with the GUI 220 Notes For Command Line Users 221 Two Channel Diagnostic Plots 221 Location and Scale Normalization Within Array Normalization 225 maNormMain Function 226 Normalization Functions maNorm maNormScale 228 Normalization Between Arrays 230 Pre Processing and Normalization for Affymetrix Probe Level Data 232 CDF and Probe Libraries 233 Affymetrix Diagnostic plots 234 207 Chapter 6 Pre Processing and Normalization Background Correction PM Correct Methods Normalization Summarization Methods Normalization Methods for Affymetrix MAS Data Normalization Methods Diagnostic Plots for Summarized Affymetrix Data References 208 237 239 241 245 251 252 254 256 Introduction INTRODUCTION Before differential expression testing ca
19. Note that when the identifiers from one source are linked to the identifiers from another there does not need to be a one to one relationship For example several different Affymetrix probe ids correspond to a single LocusLink identifier Thus when mapping Affymetrix to LocusLink we have no problem but mapping LocusLink to Affymetrix requires a mechanism for dealing with the multiplicity of matches Annotation and Gene List Management Functionality There is a great deal of annotation data available for any given gene Examples include LocusLink Unigene chromosome number chromosomal location cytoband or bp KEGG pathway information and the Gene Ontology GO categorizations The Annotation libraries for arrays in common use are included in the S ARRAYANALYZER installation media All of the Annotation libraries are available from the data libraries link in the S ARRAYANALYZER Web site http www insightful com support ArrayAnalyzer DataLibs README html 333 Chapter 9 Annotation and Gene List Management ANNOTATION LIBRARIES 334 The Affymetrix chip specific annotation libraries in S ARRAYANALYZER are as follows hgu133a hgu133b hgu133plus2 hgu95av2 hgu95b hgu95c hgu95d hgu95e hu6800 moe430a moe430b moe430v2 mgu74a mgu74av2 mgu74b mgu74bv2 mgu74c mgu74cv2 rae230a rae230b rgu34a rgu34b rgu34c The general annotation libraries in S ARRAYANALYZER are as follows GO KEGG humanCHRLOC humanLLM
20. This sequence of steps is available by simply selecting the RMA checkbox in the upper right corner of the Affymetrix Expression Summary dialog Open the dialog by selecting the Affymetrix Expression Summary selection from the ArrayAnalyzer menu Then select the SurgeryAffyBatch object in the CEL Data drop down list and select the RMA checkbox The result of the computation is an expression summary object Save the result as SurgeryExprSet rma by typing it into the Save As field A sequence of graphs is produced as output by the RMA procedure Figure 3 37 displays MvA plots created for all the replicates within one treatment condition The values in the lower left triangle of the matrix are the inter quartile ranges IQR of the values of M across all summarized expression values A small value indicates there is little 107 Chapter 3 Examples Affymetrix Probe Level Data difference on the log scale for the middle 50 of the expression values for the two chips For replicate chips there is no real differential expression so the IQR is expected to be small zox Data Options CEL Data SurgerAttyB ate x Summary Options C GCRMA S As Sur eryE xprSet rm EANA OAE ataia C Mix amp Match Summary avaditt z Normalization quantiles r Summary Graphics V MvA Plot IV Box Plot OK Cancel Apply current Help Figure 3 36 Specifying Robust Multichip Analysis with a single checkbox Figure 3 38 displays b
21. probe level microarray data The S PLUS script splus62 module Array Analyzer examples scripts normalization_chapter ssc includes the example code presented in this chapter Some of the example computations are time consuming 1 Affymetrix GeneChip microarrays represent each gene with an oligonucleotide 25 mer probe spotted at typically 16 20 pairs of spots 32 40 spots in all Each probe pair consists of a spot for the probe called a perfect match PM and a spot for a slight alteration of the probe called a mismatch MM The collection of the PM and MM spots for a specific gene are called a probe set 209 Chapter 6 Pre Processing and Normalization NORMALIZATION Technical Sources of Variability Two Channel Data Affymetrix Data 210 Many factors can modify spot intensity other than differential gene expression And each type of microarray chip has its own inherent systematic variations that need to be taken into account Normalization can be thought of as a series of corrections for these systematic effects In general normalization is needed to ensure that observed differences in fluorescence intensities are due to differential expression and not experimental artifacts and should be done before any analysis that compares gene expression levels within or between arrays Effects that have been consistently removed by normalization include differential nonlinear hybridization of the two channels in two color cD
22. rep 5 23 Note that this function and the function plclust2 fn which allows a dendrogram to be rotated and easily laid next to a heat map are included in the S tARRAYANALYZER module Results of the hierarchical cluster analysis are presented below in Figure 8 13 Overall we produce a clustering result similar to that of Alizadeh et al 2000 Note that the data we worked with are not the actual raw data We treat it as raw data to show the cluster methods in S PLUS but the resulting output should not be directly compared with figure 3a of Alizadeh et al 2000 We have ordered the columns based on the default rule in S PLUS namely that at each merge the subtree with the tightest cluster is placed to the left This is the opposite of the ordering used by the package Cluster Eisen et al 1998 which was used by Alizadeh et al 2000 The columns of our heat map are thus in approximately reverse order to those presented in Alizadeh et al 2000 Also note that individuals within nodes are paired by their original order in S PLUS while this ordering is at random in the package Cluster Eisen et al 1998 Further Alizadeh et al 2000 use a weighting function that is not well documented in the Cluster Eisen et al 1998 manual 325 Chapter 8 Cluster Analysis 326 E bane eens te Td m LD Figure 8 13 Heat map and dendrogram based on data from Alizadeh et al 2000 Note that this is not t
23. sample differential expression analysis They include the t test with or without the equal variance assumption the Wilcoxon rank sum test and several permutations tests In addition the LPE test procedure which produces improved error estimates when there is little replication in the design is implemented for two sample problems All these procedures are suitable for doing simple comparisons between two groups treatment versus control tissue 1 versus tissue 2 etc For more details about two sample procedures see Chapter 7 Differential Expression Testing A one way design is an extension to the two sample design Instead of comparing two samples you compare multiple samples For example treatment 1 treatment 2 and a control form a simple one way design The example we examine in this chapter is set up as a one way design but primary comparisons can also be thought of a two sample problems We ll demonstrate both approaches to the problem 17 Chapter 2 Examples Affymetrix MAS Data ONE WAY DESIGN Swimming Mice Data 18 In this chapter we step through the analysis of an experiment designed to improve understanding of the effect of chronic conditioning on the mass build up of the left ventricular muscle of the heart A study was conducted on eight week old mice which were regularly exercised by swimming Over the course of 10 days exercise was increased from 10 minutes twice a day to 90 minutes twice a day Conditioning of t
24. 308 310 310 311 316 317 320 329 301 Chapter 8 Cluster Analysis INTRODUCTION 302 Cluster analysis is the searching for groups clusters in the data in such a way that objects belonging to the same cluster resemble each other whereas objects in different clusters are dissimilar In two or three dimensions clusters can be visualized With more than three dimensions we need some kind of analytical assistance Generally speaking clustering algorithms fall into two categories 1 Hierarchical Algorithm A hierarchical algorithm describes a method yielding an entire hierarchy of clustering for the given data set Agglomerative methods start with the situation where each object in the data set forms its own little cluster and then successively merges clusters until only one large cluster remains which is the whole data set Divisive methods start by considering the whole data set as one cluster and then splits up clusters until each object is separate 2 Partitioning Algorithms A partitioning algorithm describes a method that divides the data set into k clusters where the integer k needs to specified Typically you run the algorithm for a range of A values For each the algorithm carries out the clustering and also yields a quality index which allows you to select the best value of k afterwards Algorithms of this type include k means and partitioning around medoids Clustering approaches have been wi
25. Dudoit S 2003 Bioconductor s marrayNorm Package Bioconductor marrayNorm library documentation January 23 2003 p 3 Wu Z Irizarry R A Gentleman R Murillo F M and Spencer F 2003 A Model Based Background Adjustment for Oligonucleotide Expression Arrays Johns Hopkins University Dept of Biostatistics Working Papers 258 DIFFERENTIAL EXPRESSION TESTING Introduction Statistical Tests Within Gene Two Sample Comparisons Local Pooled Error Test Raw P Values Controlling Type I Error Rates Controlling The False Positive Rate GUI for Two Sample Testing Two Sample Dialog Input GUI for LPE Testing LPE Testing Dialog Input GUI for ANOVA Testing ANOVA Testing Dialog Input Differential Expression Analysis Plots Common Plots Differential Expression Summary Table Output Top 15 List Complete Gene List References 260 261 261 262 263 265 266 270 270 276 276 280 280 286 286 294 294 294 298 259 Chapter 7 Differential Expression Testing INTRODUCTION 260 Differential expression testing in S ARRAYANALYZER is defined as statistical testing of the difference in expression intensities between the treatment conditions under study Effectively this means that a separate hypothesis test is computed for each gene or probe on the chip In two sample problems e g two tissue types treatment vs control this boils down to a t test or other two sample test for each
26. FDR 84 143 FWER 84 143 144 G GenBank 88 145 H Heat Map 89 146 Heat map 84 heat map 92 144 321 324 340 I INSTALL txt 6 installing I Miner 6 L LocusLink 9 92 145 340 341 360 M medianIQR 338 341 MIAME 75 100 129 MvA plot 141 P paired t 142 144 platforms supported 6 Q QQ Norm Plot 147 R requirements system 6 S support technical 7 supported platforms 6 system requirements 6 389 Index 390 T technical support 7 V Variance Plot 91 Variance plot 85 Volcano Plot 145 Volcano plot 84 189 volcano plot 2 88 339 WwW Wilcoxon test 142 144
27. Figure 4 19 Selecting variables for expression analysis and filtering Quality Once you ve read the data in you can generate an assortment of Diagnostics visual quality checks on the data These checks include Image plots of the entire array for any of the channels Multiple color schemes are available for the plots e MvVA plots as scatter plots or hexbin plots Hexbin plots display hexagonal points that are colored to give a sense of the density of intensity values at each location e Expression intensity boxplots by array or by print tip group for each array Principal components plot where points represent each array For this plot all the expression intensities for each array are considered the variables of interest 151 Chapter 4 Examples Two Color Data 152 For more information on the methods and interpretation of these visual quality diagnostics see Chapter 5 Quality Control Diagnostics and Filtering Select ArrayAnalyzer gt Quality Control Diagnostics and fill in the resulting dialog as displayed in Figure 4 20 Clicking OK or Apply generates the plots seen in Figures 4 21 4 22 4 23 and 4 24 Quality Control Diagnostics Two Channel Data Sf Show Data of Type Two Channel z Data TPMarayRaw z Array Name fiplayout m MvA Plot T MvA Plot I Smooth Curves m Image Plot IV Image Plot Signal Ratio M Channel both x Channel Type foreground z
28. GO 0008361 regulation of cell size 480 E GO 0016049 cell growth 438 GO 0006997 nuclear organization and biogenesis 1274 Annotation from the GUI Genelist Analysis Internet Figure 9 9 AmiGO results for Melanoma data with fold change set to greater than 2 and p value set to significant adjusted p lt 0 05 for LPE test Bonferroni FWER in this case top 10 genes based on p value followed by fold change We saw above the simple metadata lookup available for a gene list derived in S ARRAYANALYZER with on line databases at GenBank LocusLink Entrez Nucleotide Pubmed and the Gene Ontology Consortium GO AmiGO via the Annotation dialog Some other Web sites are available for statistical analysis of genelists for example functional gene enrichment where we ask the question Are there more genes at a given GO node e g the biological function GO tree in a gene list than we would expect by chance 349 Chapter 9 Annotation and Gene List Management Annotation Using Affymetrix IDs 350 S ARRAYANALYZER includes access and gene list upload to three sites for remote gene list analysis 1 Affymetrix NetAffx http www affymetrix com analysis index affx 2 NIH DAVID EASE http apps1 niaid nih gov david 3 Onto Express http vortex cs wayne edu ontoexpress onto htm These sites allow remote upload of gene list AffyIDs and or LocusLink IDs as enabled by StARR
29. SO0COCOOOO SSS5555 000000000555555555 oo0oo0ooooooo0oo0 ORe Re RERE ESRC PORE ee gt ea ae ae ai Figure 3 35 Pre and post normalized probe level data The normalized intensities are plotted on the log scale Once you import the data files you need to convert the raw probe level expression intensities to expression summaries before testing for differential expression This is usually done in a series of steps including some combination of the following e Background correction Probe specific background correction e g subtracting MM e Normalization Summarization of the probe set values into one expression measure Sometimes standard errors are computed as well RMA Summary RMA Output Two Way Design An assortment of procedures are available for completing these steps In addition to normalization in the context of summarizing raw expression intensities you can also normalize without the summarization step For more detail see Chapter 6 Pre Processing and Normalization In this section we focus on one sequence of steps referred to as robust multichip analysis or RMA for short This procedure completes the following steps 1 Probe specific correction of the PM probes using a model based on observed intensity being the sum of signal and background noise 2 Normalization of corrected PM probes using quantile normalization Bolstad et al 2002 3 Calculation of expression measures using median polish
30. The Gene Filtering tab is designed to create filtering expressions for dropping or keeping genes rows in the expression matrix For this example use the Flags column to create expressions for eliminating rows with non zero flags Do so in the following order 1 Select a data column Flags and click the Add button in the Data group to add the name Flags to the Expression field at the bottom of the dialog 2 Select the symbol double equal sign from the Logical field of the Operations group and click the Add button right under Logical label to add the symbol to the Expression field at the bottom of the dialog 3 Select the value 0 zero from the list in the Column Values group and click the Add button in that group to add 0 zero to the Expression field at the bottom of the dialog 155 Chapter 4 Examples Two Color Data 4 Select 1 one in the InAtLeast Value field of the Operations group and click the Add button right under the InAtLeast Value label to add InAtLeast 1 to the Expression field at the bottom of the dialog Note on InAtLeast The InAtLeast 1 operation allows you to keep any gene that has a Flag value of zero on one or more of the arrays Without this operation a gene will be eliminated even if the Flag value is non zero on only one array 5 Ensure that the Values Selected in the Expression group is set to Keep 6 Click OK to eliminate all genes except those with Flag 0 The resu
31. Vv Top 15 Genes Options 5 Output EWER FDR 0 05 Display Qutput in S PLUS Test t v I Save Output as HTML Alt Hypothesis Not equal E Save HTMLA myMultT est htm Adjustment Bonferroni I Display HTML Gutpu Permutation 10000 Save As myMultT est Random Seed Cancel ge current Help Figure 7 1 The Two Sample Test dialog Once a data object is selected the chip name is filled in the Chip Name field For custom 2 color cDNA or non Affymetrix oligonucleotide chips the chip name may be lt undetermined gt GUI for Two Sample Testing Options The Options group displayed in Figure 7 2 allows you to specify various options for the statistical tests 1 Select the statistical test default is Welch s t test 2 Specify the alternative default is Not equal 3 Input the FWER or FDR 4 Select the p value adjustment procedure 5 Specify the number of iterations along with a random seed if you select one of the permutation methods m Options EWER FDR 0 05 Test l v Alt Hypothesis Not equal X Adjustment Bonferroni i ations 10000 Figure 7 2 The Options group of the Two Sample Tests dialog Statistical Tests The statistical tests and p value adjustment procedures are all described in the section Statistical Tests and the section Controlling Type I Error Rates The key words or phrases used to select one of these options match those used in the descriptive text m Options FWER FDR
32. are commonly defined as follows M log Ek Ep ga E se y Box Plots MvA plots Pre Processing and Normalization for Affymetrix Probe Level Data where E pis the expression intensity for a given replicate and treatment i 1 of replications of the k th treatment and j 1 of replications of the th treatment Since Affymetrix probe level data is on the raw scale box plots from the GUI plot the logo of the intensities Please refer to section Box Plots on page 218 for a general discussion of box plots The function mva pairs or the MvA plots from the GUI show pairwise graphical comparisons e g between replicates of a treatment condition The axes of these plots are the log ratio intensities M between a replicate chip pair vs the average log sum A intensities of the chip pair The pairwise scatter plots are shown on the top right half of the graph and the inter quartile range IQR of the log ratios is shown on the bottom left half of the graphs The chip labels are given on the diagonal These plots can be particularly useful in diagnosing problems in replicate sets of arrays Figure 6 5 shows an MvA plot for one treatment condition of the Dilution experiment there are two replicates of this condition Please refer to the help files for information about the Dilution dataset Plots from the GUI show one 235 Chapter 6 Pre Processing and Normalization pairs plot per treatment condition From the
33. automatically hyperlinked to LocusLink and UniGene annotation databases Save Summary As Name used for saving the S PLUS data frame containing the complete gene list including test statistics and p values The default name is myMultTest Output Files There are two output files generated when you select Save Output as HTML 1 2 A significant genes summary table The name of the this table is generated by taking the name supplied in the Save Summary As field and then adding myMultTest html The default supplied name is myMultTest so the default output table name is myMultTest html An S PLUS Graphlet with selected graphs The name of the output files is generated by the name supplied in the Save Summary As field and then adding myMultTest html The default Graphlet name is myMultTest html GUI for Two Sample Testing Location of Output Files The location of these output files is determined by your S PLUS working directory To determine your working directory just type gt getenv S_ WORK D arrayanalyzer users lenk test The location of dumped files in general is the default S PLUS working directory If you specify no project folder when you start S PLUS your cmd directory is the default working directory gt getenv S_ WORK D Program Files Insightful splus62 cmd You should see two HTML files in your working directory when S PLUs has finished generating the output one for the summary tab
34. c Luce c E E ees c Pes eT E E a E C E E k L L L Pasong E L E OldOhr1 OldOhr2 OldOhr3 Old1hr1 Figure 3 38 Boxplot of log expression the composite RMA procedure Old1hr2 Oldthr3 Old4hr1 Old4hr2 Old4hr3 YoungOhr1 YoungOhr2 YoungOhr3 Youngthr1 Young1hr2 Young thr3 Young4hrt Young4hr2 Young4hr3 intensities for the 18 samples after applying 109 Chapter 3 Examples Affymetrix Probe Level Data Differential Expression Analysis ANOVA Dialog 110 We are now ready to do differential expression analysis for the two way design From the main menu open the ANOVA dialog by clicking ArrayAnalyzer gt Differential Expression Analysis gt ANOVA The procedures implemented in this dialog provide traditional ANOVA methods with a host of correction procedures to control the Type I Error rate For more details on the ANOVA dialog and Type I Error correction procedures see Chapter 7 Differential Expression Testing To set up the ANOVA dialog follow these steps 1 2 3 In the Show Data of Type field select Affymetrix In the Data field select SurgeryExprSet rma The Array Name field should be automatically updated to mgu74av2 In the Contrasts group confirm that the Baseline checkbox is checked Select Age as the factor for doing baseline comparisons From the Baseline Level drop down list select Young In the
35. selected graphs and the significant gene list to HTML files to view later Display HTML Output View the S PLUS Graphlet with selected graphs in a browser The displayed Graphlet has a hyperlink to the significant genes table Points on the Graphlet and entries in the significant gene list are automatically hyperlinked to LocusLink and UniGene annotation databases 278 GUI for LPE Testing e Save Summary As Name used for saving the S PLUS data frame containing the complete gene list including test statistics and p values The default name is myLPETest Output Files There are two output files generated when you select Save Output as HTML 1 A significant genes summary table The name of the this table is generated by taking the name supplied in the Save Summary As field and then adding myLPETest html The default supplied name is myLPETest so the default output table name is myLPETest html 2 An S PLUS Graphlet with selected graphs The name of the output files is generated by the name supplied in the Save Summary As field and then adding myLPETest html The default Graphlet name is myLPETest html Location of Output Files The location of these output files is determined by your S PLUS working directory To determine your working directory just type gt getenv S_ WORK D arrayanalyzer users lenk test The location of dumped files in general is the default S PLUS working directory If you specify no
36. transition metal ion binding o manganese ion binding 1 p 0 5 protein binding 2 p 0 5 102226 st told change 0 0 92759_at fold change 0 0 catalytic activity o hydrolase activity o peptidase activity 4 p 0 5 transferase activily 1 p 05 transferase activity transferring givcosyi groups 1 p 0 6 102836_at told change 0 0 moleculat_function unknown 1 p 0 5 102402_at told change 0 0 structural molecule activity 1 p 05 extraceliular matrix structural constituent 1 p 0 5 92759_at fold change 0 0 biological_process o biological process unknown 1 p 05 102402_at told change 0 0 cellular process o cell communication o celladhesion 2 p 05 102228 st fold change 0 0 92759_at fold change 0 0 physiological process o metabolism o biosynthesis o macromolecule biosynthesis o lipid biosynthesis o membrane lipid biosynthesis o Figure 9 19 OntoExpress java applet as launched from S ARRAYANALYZER showing tests for functional enrichment for the uploaded gene list using settings chosen from within S ARRAYANALYZER We saw above the simple metadata lookup available for a gene list derived from statistical analyses using the in the Annotation dialog We now show similar functionality from the S ARRAYANALYZER command line scripting environment JH Load libraries for annotation gt module arrayanalyzer gt library hgu95av2AnnoData JHF 1 Setup the data for analysis JHF Ma
37. x NULL y maM median absolute deviation MAD geo TRUE subset TRUE of intensity log ratios for a group of spots maNorm2D 2D spatial location normalization Defaults x maSpotRow Normalizes to the smoothed y maSpotCol z maM intensity surface loess surface by g maPrintTip w NULL print tip group at each x y subset TRUE span 0 4 coordinate Let s normalize the swirl data using a variety of methods in the maNormMain function The normalization methods will be applied to the set of chips given If you don t want to normalize across treatment conditions then the marrayRaw objects can be subset as shown below The swirl dataset For description type swirl or help swirl gt swirl Global median normalization for arrays 81 and 82 dt first two chips in set 227 Chapter 6 Pre Processing and Normalization Normalization Functions maNorm maNormScale 228 2 i gt swirl normGmed lt maNormMain swirl 1 2 f loc list maNormMed x NULL y maM Global median normalization over all chips in swirl swirl normGmed lt maNormMain swirl f 1loc list maNormMed x NULL y maM 2D spatial location normalization of array 93 swirl norm2D lt maNormMain swirll 3 f loc list maNorm2D A normalization that is a weighted average of the loess normalization over the chip and the loess normalization over the print tip groups swirl norm lt maNormMain swirl 1 f lo
38. xl ap internet Z Figure 9 8 Pubmed results for Melanoma data with fold change set to greater than 2 and p value set to significant adjusted p lt 0 05 for LPE test Bonferroni FWER in this case top 10 genes based on p value followed by fold change 348 E AmiGO Your friend in the Gene Ontology Microsoft Internet Explorer AL i lol x File Edit View Favorites Tools Help Annotation Libraries Hck gt amp A Al Asearch fyravorites Smeda lt 4 B S fe a Address amp http www godatabase org cgi bin amigo go cgi action replace _tree amp query G0 00047 1 3 amp query G0 00055248 amp query G0 0004674 amp query G Y co Links fed s FlyBase SGD COR Submit Query A aeolicus A fulgidus zi El GO 0003673 Gene_Ontology 146200 H GO 0008150 biological_process 96312 GO 0009987 cellular process 31905 E GO 0007 154 cell communication 8954 GO 0007165 signal transduction 6780 GO 0007242 intracellular signaling cascade 2345 Ho GO 0050875 cellular physiological process 25201 E GO 0008218 cell death 1837 Ho GO 0012501 programmed cell death 1631 GO 0006915 apoptosis 1353 E GO 0008151 cell growth and or maintenance 22177 GO 0016048 cell growth 438 Ho GO 0016043 cell organization and biogenesis 6226 GO 0000902 cellular morphogenesis 1532
39. 3 Save Output Save Data Set As myMarrayRaw I Display Report cont kr Figure 4 2 The Import cDNA Data dialog Two Sample Design The Import Data From Two Channel dialog has four pages e File Selection This page must be completed in order to create a data object for continued analysis e MIAME Completing this page is optional but highly recommended because information on the MIAME page is used for labeling tables and graphs e Variable Selection amp Filtering Red and green foreground colors are required Background colors are optional e Options This page provides options for specifying the number of header lines to skip and the delimiter used in the data file Data import is accomplished in three steps 1 Create the experimental design 2 Associate files with design points 3 Save data set as an S PLUS object Step I Create Before you associate data files with experimental conditions you The Experimental must to set up the experimental design in S ARRAYANALYZER Start Design by opening the Create Modify Design dialog Step 1 on the Import Data From Two Channel dialog For this simple two sample experiment most of the defaults work as they are The Number of Arrays and the Number of Factors have defaults of four and one respectively which correspond to this experiment The settings to change are 1 Set the Factor Name to Zebra Fish 2 Set the Level Values to Swirl and WildType L
40. AF045887 M23383 4F079565 AWw049359 Gol Clear Limits Preview Index History Clipboard Details Display Summary 7 Show 20 Send to Tex Items 1 5 of 5 L29454 Mouse fibrillin Fbn 1 mRNA complete cds gi 575509 gb L29454 1 MUSFBN1A 575509 AF045887 Mus musculus angiotensinogen precursor gene exon 5 and complete cds gi 2842773 gb AF045887 1 MMANGIO9 2842773 M23383 M musculus glucose transporter 2 mRNA complete cds gi 193706 gb M23383 1 MUSGT2A 193706 AF079565 Mus musculus ubiquitin specific protease UBP41 Ubp41 mRNA complete cds g 3386551 gb AF079565 1 4F079565 3386551 AW049359 UI M BH1 ane a 09 0 U1s1 NIH_ BMAP M 2 Mus musculus cDNA clone gi 5909888 gb AW049359 1 5909888 Figure 2 33 Annotation summary from Unigene 53 Chapter 2 Examples Affymetrix MAS Data FROM THE COMMAND LINE All of the analysis done through the GUI can be done from the S PLUS command line Having access to the command line adds great flexibility to the set of features available through the S ARRAYANALYZER GUI and opens the door to additional analyses The flexibility and feature rich S PLUS language make it an ideal platform for exploratory analysis statistical testing and modeling of gene expression data This section is designed to expose you to the critical functions for differential expression testing of microarray data If you have no interest in running your analyses from the command line you can skip this secti
41. Click the drop down button and select mgu74av2 as shown in Figure 2 6 RETEN X Figure 2 6 Selecting mgu74av2 as the Affymetrix array type S ARRAYANALYZER has pre loaded the gene annotation information for arrays hgu95av2 and hgu133a If you are using other arrays you may want to refer to the Chapter 6 Pre Processing and Normalization to see how to load the annotation information for your array To save the data object type a name in the Save Data Set As field near the bottom of the dialog Step 3 Save Output Remember this name It is used in other analysis steps such as quality checks filtering and normalization For our example enter MouseSwimExprSet as the object name The Display Report check box indicates whether or not to print summary information resulting from reading the data into an S PLUS report window Save Data Set As Mouses wim xprSet M Display Report Step 3 Save Output Figure 2 7 Saving the imported data as MouseSwimExprSet Saving the Design Once you ve entered all the information on this tab you can save it for later use by clicking the Save Design button at the top of the dialog A txt file is written to the directory of your choice with the number of factors number of levels repetitions and the full path file names and their associated factor levels 27 Chapter 2 Examples Affymetrix MAS Data MIAME Page 28 Reading Designs The design file can be reused for another expe
42. Clustering dialog by clicking ArrayAnalyzer Cluster Analysis from the main S PLUS menu bar On the General Options tab select Two Channel from the Show Data of Type drop down list and TPMarrayRawFiltered norm from the Data drop down list Leave the other fields are they are by default and switch to the Filtering Options tab 161 Chapter 4 Examples Two Color Data 162 Custer analysis TT General Options Filtering Options Data Partitioning Methods Show Data of Type T Partitioning Around Medoids Two Channel E Number of Clusters Data TPMarrayBaw fatto x Array Name Ttplayout 1 Clister ot gt Response Variable Response M A ae pey R Standardization Standard value x 11 Giuste an acne Hierarchical Methods T Model based Hierarchical Clusterign aines zj Weighting method average x E Output Dist metric euclidean z Display Output in S PLUS IM Names on Graph Save Output as HTML Saye HTML As Jrepclustechiml J Display HTML Output Save s inyCluster Cancel Apply i current Help Figure 4 30 General Options tab of the Cluster Analysis dialog ready for a hierarchical cluster analysis of the TP data On the Filtering Options tab in the Gene Sort Order Options group on the far right change the Limit number of genes to 500 This will allow us to cluster on the 500 most express
43. D 3 fw a Address http www ncbi nih gov LocusLinkjLocRpt cqi I 1490 2c707 1 2c3491 2c51421 hd C60 Links ke 3 NCBI LocusLink PubMed Entrez BLAST OMIM Map Viewer Taxonomy Search LocusLink 7 Display Brief Organism All Query Go Clear View MOGNEM One of 4 Loci ABCDEFGH IJKLMNOPQRSTUVWXYZ Structure Click to Display mRNA Genomic Alignments spanning 3126 bps Gene PUB OMIM ACEVIEW UNIGENE HOMOL GDB e UCSC Homo sapiens Official Gene Symbol and Name HGNC CTGF connective tissue growth factor LocusID 1490 Overview Locus Type gene with protein product function known or inferred Product connective tissue growth factor Alternate CCN2 NOV2 IGFBP8 oft Figure 9 20 LocusLink annotation for the four genes identified in the gene filtering analysis described above Note that the View field in LocusLink is populated with the four genes identified in the analysis There are several other simple metadata lookup queries we can run from the command line scripting environment for the gene list that we are holding in the S PLUS variable mel gnames We illustrate by obtaining accession number information and Pubmed articles S PLUS code for this annotation follows and results from the queries are presented in Figures 9 21 and 9 22 361 Chapter 9 Annotation and Gene List Management Entrez Nucleotide Microsoft Internet Ex
44. G Parmigiani E S Garrett R A Irizarry and S L Zeger Published by Springer New York Li C Wong W 2001a Model based analysis of oligonucleotide arrays Expression index computation and outlier detection Proceedings of the National Academy of Science U S A 98 31 36 Li C and Wong W 2001b Model based analysis of oligonucleotide arrays model validation design issues and standard error application Genome Biology 2 8 research0032 1 0032 11 Parmigiani G Garrett E S Irizarry R A and Zeger S L 2003 The analysis of gene expression data an overview of methods and software In The Analysis of Gene Expression Data Methods and Software Edited by G Parmigiani E S Garrett R A Irizarry and S L Zeger Published by Springer New York 257 Chapter 6 Pre Processing and Normalization S PLUS 2000 Guide to Statistics Volume 1 Data analysis Products Division MathSoft Seattle WA Yang Y H Dudoit S Luu P Lin D M Peng V Ngai J and Speed T P 2002 Normalization for cDNA microarray data a robust composite method addressing single and multiple slide systematic variation Nucleic Acids Research 20 4 Yang Y H Dudoit S Luu P and Speed T P 2001 Normalization for cDNA microarray data In M L Bittner Y Chen A N Dorsel and E R Dougherty editors Microarrays Optical Technologies and Informatics volume 4266 of Proceedings of SPIE May 2001 Yang Y H
45. IV Heat Map F Chromosome Plot IV Parallel Coords IV Top 15 Genes Figure 7 12 The Graph Options group in the LPE test dialog described in detail in the section Differential Expression Analysis Plots The Output group controls where the graphs are displayed and the gene list table is saved after the testing step is complete 283 Chapter 7 Differential Expression Testing 284 Display Output in S PLUS Displays the selected graphics in an S PLUS graphic device Save Output as HTML Saves the S PLUS graphlet with selected graphs and the significant gene list to HTML files to view later Display HTML Output View the S PLUS Graphlet with selected graphs in a browser The displayed Graphlet has a hyperlink to the significant genes table Points on the Graphlet and entries in the significant gene list are automatically hyperlinked to LocusLink and UniGene annotation databases Save Summary As Name used for saving the S PLUS data frame containing the complete gene list including test statistics and p values The default name is myANOVA Output Files There are two output files generated when you select Save Output as HTML 1 2 A significant genes summary table The name of the this table is generated by taking the name supplied in the Save Summary As field and then adding myANOVA html The default supplied name is my ANOVA so the default output table name is myANOVAhtml An S PLUS Graphlet with select
46. Import Data gt From Affymetrix ArrayAnalyzer Import Data tron Niece Quality Control Diagnostics gt Fromtwo Channel Filtering Affymetrix Expression Summary Normalization Differential Expression Analysis gt Cluster Analysis Annotation Gene List Management Figure 3 24 Menu selection to import Affymetrix data Figure 3 25 shows the Import Affymetrix Data dialog with the File Selection page displayed The primary tasks of the import process are 1 create an experimental design 2 associate data files with the experimental conditions and 3 save the resulting S PLUS data object 95 Chapter 3 Examples Affymetrix Probe Level Data for later use Secondary tasks include inputting meta MIAME data describing the experiment and specifying options for handling data marked as MASKS or OUTLIERS Import Data From Affymetrix xj File Selection MIAME MAS Variables amp Filtering CEL Filtering Options r Step 1 Specify Design Read Existing Design Create Modify Design r Step 2 Associate Files with Design Points File Type cundetermined gt Array Name crequired gt 7 CDF E r Step 3 Save Output Save Data Set As mySet IV Display Report Cancel Figure 3 25 The Import Affymetrix Data dialog 96 Step I Create The Experiment Design Two Way Design The Import Data from Affymetrix dialog has five pages File Selection This p
47. Import From File dialog You specify the name of the saved data object in the Data Set field If there is any header information in the file you need to specify Start row so the header information is skipped You do that on the Options page of the Import From File dialog Figure 2 35 shows the location of the Start row field See the position of the mouse cursor oix Data Specs Options Filter gt From File Name C Program Files Insightful splus62 moduleArrayAnalyzervex Browse File Format Excel Worksheet xl h To Data set xih v Create new data set Add to existing data set Start col END v Preview Rows fi 0 Probe Set Name Positive Negative Factor double double double 1 AFFX MurIL2_at 7 6 20 a AFFX MurILi0_at 6 6 20 3 AFFX MurIL4_at 7 6 20 of E OK Cancel Apply current Figure 2 34 Importing data through the S PLUS GUI The command line equivalent to the Import From File dialog is the importData function The critical arguments are file a character string specifying the name of the file to import 55 Chapter 2 Examples Affymetrix MAS Data type a character string specifying the type of file to import Possible values are listed here the case of the character string is ignored startRow an integer specifying the first row to be imported from the data file This argument is available only when importing data
48. Intensity shifted and scaled RNA Degradation Plot Probe Number Figure 3 33 RNA degradation plot Old Ohr Old 1hr Old 4hr Young Ohr Young thr Young 4hr Normalization Normalization Dialog Two Way Design Normalization procedures may be applied to both raw probe set intensities and to summarized expression intensities For examples of normalizing expression summary data see Chapter 2 Examples Affymetrix MAS Data and Chapter 6 Pre Processing and Normalization In this section we focus on normalizing probe set data without summarizing it Open the Normalization dialog by selecting Normalization from the ArrayAnalyzer drop down menu Normalization working E lol x r Data r Normalization Show Data of Type Normalization loess ha Attymetrix CEL VV MvA Plot Data SurgeryAtfyBatc Ne I Box Plot Save As SurgerafiyB atch When to Show Before amp After Probe Set PM C Only After PM and MM i OK Cancel Apply current Figure 3 34 The Normalization dialog Help Select Affymetrix CEL from the Show Data of Type drop down list and choose the SurgeryAffyBatch data object Explore the normalization procedures available in the Normalization drop down list The quantiles procedure allows you to normalize only the PM intensities or both PM and MM intensities Save As The Save As field takes an object name for saving the normalized affyBatch
49. LCG LCGLLi lt logb ifelse CG Li lt 1 1 CG i base 2 Removing Controls Normalization From the Command Line Now remove the control rows from the expression data frame JHH Remove controls gt LOG lt LOG L eontrols We normalize the intensity values by adjusting within array expression values to a common median and inter quartile range using the medianIQR norm function JHHF Normalize using medianIQR gt LCG N lt data frame medianIQR norm LCG Compute a summary of the resulting logged and normalized expression intensities as follows dHHE Summarize Non controls gt summary LCG N CGa CGb CG24a Min 0 5415185 Min 0 000000 Min 0 5071245 Ist Qu 3 2774186 Ist Qu 3 277464 Ist Qu 3 2121624 Median 6 6429682 Median 6 642968 Median 6 6429682 Mean 6 0522140 Mean 5 964245 Mean 6 0087211 3rd Qu 8 8044455 3rd Qu 8 804491 3rd u 8 7391893 Max 14 6903453 Mex 215 151559 Max 15 1604125 NA s 8 0000000 NA s 8 000000 NA s 8 0000000 CG24b Min 0 5725 788 Ist Quis 3 181388 Median 6 6429682 Mean 5 9970227 3rd Q 8 7084150 Max 15 1399289 NA s 8 0000000 Note the missing values in the summary output We ll have to take care of those before doing differential expression testing but first let s plot normalized and unnormalized data for comparison JHHF Before and after normalization boxplots gt par mfrow c 1 2 gt boxplot LCG style bxp att gt ti
50. Methods on page 304 and section Partitioning Methods on page 308 are available from the S PLUS command line The methods are provided in several flavors through different S PLUS functions To become fluent in clustering in S PLUS takes a bit of practice but it s well worth the effort The flexibility and richness of the S PLUS language extends the current capabilities of the GUI A variety of partitioning and hierarchical cluster analysis methods are available in S PLUS including a library of algorithms described in Kaufman and Rousseeuw 1990 Partitioning Methods The partitioning methods include K means partitioning around medoids and a fuzzy clustering method in which probability of membership of each class is estimated These methods are available through the following S PLUS functions kmeans K means clustering pam partitioning around medoids clara partitioning around medoids for large data sets The key difference between kmeans and pam is that kmeans uses a mean as the center of each cluster and pam uses an actual data point as the center Hierarchical Methods The hierarchical methods include agglomerative methods which start from individual points and successively merge clusters until one cluster representing the entire dataset remains and divisive methods which consider the whole dataset and split it until each object is separate The available agglomerative methods are hclust Performs hierarchical clusteri
51. Options are checked These include Volcano plot e Heat map e Chromosome plot e Variance plots e Top 15 genes table In the Output group Deselect the Display Output in S PLUS checkbox Select the Save Output as HTML checkbox Select the Display HTML Output checkbox and enter cgLPEBH html for the filename Enter cgLPEBon in the Save As field in the lower right corner of the dialog for saving the test result object 85 Chapter 3 Examples Affymetrix Probe Level Data The resulting dialog is displayed in Figure 3 17 Differential Expression Analysis LPE Test r Data m Variance Estimation Show Data of Type Smoother D F 10 Jattymeti Z Number of Bins fioo Data eae xprSet ima hd Trim 2 Factor Time x Sms ters Compare Level1 2hr Volcano Plot Compare Level 2 hpn x Y Axis Orientation Array Name frou5av2 negative zi m Options Fold Change Line zz FWER FOR 005 IC HeatMap Adjustment Bonfenoni x I Chromosome Plot Alt Hypothesis Notequal z F7 Variance Plots IV Top 15 Genes Output J Display Output in S PLUS IV Save Output as HTML Save HTML s fcgLPEBon html IV Display HTML Output Save As JoaLPEBon Cancel Apply if f current Help Figure 3 17 The Differential Expression Analysis LPE Test dialog ready to run the tests for differential expression for the melanoma data 86 Two Sample De
52. Options group select the BH Benjamini Hochberg FDR adjustment procedure In the SaveAs field in the lower right hand corner of the dialog Surgery ANOVABsYoungBH Click OK to compute the ANOVA results Two Way Design Import Data gt Quality Control Diagnostics gt Filtering Affymetrix Expression Summary Normalization Differential Expression Analysis LPE Test Cluster Analysis Two Sample Tests Annotation Gene List Management Figure 3 39 Opening the Differential Expression Analysis ANOVA dialog x m Data r Options Show Data of Type FWER FDR 0 05 attymetix z IV Protected Data SuraeivE xorSet Adjustment BH F Array Name mgu74av2 r Output Options Contrasts Volcano Plot Factor Age T Y Axis Orientation M Baseline negative gt Baseline Level Youn z Fold Change Line kho I Sequential I Heat Map I Linear Quadratic F Chromosome Plot Levels Order k af Parallel Coords aaa lt I Top 15 Genes r Output Move Up Move Down I Display Output in S PLUS I None IV Save Output as HTML Apply Contrasts Across Others Save HTML As SurgerwANOVABs Within Others IV Display HTML Output Save As Surgery NOVABs OK Cancel Apply I4 gt f current Help o Figure 3 40 The Differential Expression Analysis ANOVA dialog The Options group allows you to set the family wise error rate FWER an
53. Perou C M Rees C Spellman P Iyer V Jeffrey S S Van de Rijn M Waltham M Pergamenschikov A Lee J C Lashkari D Shalon D Myers T G Weinstein J N Botstein D Brown PO 2000 Systematic variation in gene expression patterns in human cancer cell lines Nature Genetics 24 3 227 35 Scherf U Ross D T Waltham M Smith L H Lee J K Kohn K W Reinhold W C Myers T G Andrews D T Scudiero D A Eisen M B Sausville E A Pommier Y Botstein D Brown P O Weinstein J N 2000 A cDNA microarray gene expression database for the molecular pharmacology of cancer Nature Genetics 24 3 236 244 Yeung K Y Fraley C Murua A Raftery A Ruzzo W L 2001 Model Based Clustering and Data Transformations for Gene Expression Data Technical Report 396 Department of Statistics University of Washington Seattle WA 329 Chapter 8 Cluster Analysis 330 ANNOTATION AND GENE LIST MANAGEMENT Annotation and Gene List Management Functionality 332 Annotation Libraries Annotation from Graphical and Tabular Reports GenBank Sites Annotation from the GUI GenBank and Other Browser Metadata Lookups Annotation from the GUI Genelist Analysis Annotation Using Affymetrix IDs Annotation Using LocusLink IDs Annotation Using OntoExpress Annotation from Command Line Scripting GenBank Metadata Lookups Filtering Genes Based on GO Categories 334 338 341 349 350 356 357 359 363 331 Cha
54. S ARRAYANALYZER menus and dialogs To obtain differential expression information from probe level cel file microarray data we perform the following basic steps 1 Import and filter the data 2 Normalization including e Adjustment for background noise e Mismatch correction e Distribution based normalization 3 Summarize 4 Differential expression analysis 5 Annotation In addition examining array quality and filtering out bad arrays and genes may be necessary and is typically done between import and normalization Clustering is also a normal part of gene expression discovery and may be performed between all the major steps of the analysis Because MAS 4 5 data has already been corrected for background noise mismatches and summarized we can skip most of step 2 and all of step 3 However if arrays have been analyzed with MAS 4 5 software only simple normalization has been done i e multiplying all expression values on a array by a single scalar such that the scaled mean expression values on each array are the same This simple normalization is not enough to account for much extraneous variability see Bolstad et al 2002 so we usually apply normalization procedures to MAS 4 5 data before analysis The following examples demonstrate steps 1 2 4 and 5 for Affy MAS data 16 Experimental Design EXPERIMENTAL DESIGN Two Sample Design One Way Design S ARRAYANALYZER has an assortment of procedures for doing two
55. Swim3wks NoSwim4wks Select the Significant genes checkbox in Contrast Filtering group Click the Recalculate button in the Gene Sort Order Options group You should see 15 listed Now on the General Options tab select the clustering method of choice and run the analysis The example output below results from partitioning around medoids clustering 317 Chapter 8 Cluster Analysis PAM PC Biplot Genes Component 2 T T j T T 6 4 2 0 S Component 1 These two components explain 98 79 of the point variability Figure 8 9 Biplot resulting from partitioning around medoids for only the significant genes of one contrast in the MouseANOVANoSwim4wksBH object PAM Silhouette Plot Genes r T T T T T 1 0 2 0 0 0 2 0 4 0 6 0 8 1 0 Silhouette width Average silhouette width 0 55 Figure 8 10 Silhouette plot resulting from partitioning around medoids for only the significant genes of one contrast in the MouseANOVANoSwim4wksBH object 318 Examples from the GUI PAM Parallel Coords Plot Genes Gene Expression Intensity ae T T T TT T T T T 1 2 3 4 5 Experimental Condition Figure 8 11 Parallel Coords plot resulting from partitioning around medoids for only the significant genes of one contrast in the MouseANOVANoSwim4wksBH object 319 Chapter 8 Cluster Analysis EXAMPLES FROM THE COMMAND LINE Available Methods 320 The methods described in section Hierarchical
56. Test Alt Hypothesis tequalvar tequalvar permute t permute wilcoxon wilcoxon permute Adjustment Permutations Random Seed Figure 7 3 Selecting a statistical test procedure in the Options group FWER and FDR Control The procedures for controlling the FWER and FDR are shown in the drop down list of Figure 74 The procedures correspond to those described in the section Controlling The False Positive Rate and the section FDR Procedures Both FWER and FDR procedures are 271 Chapter 7 Differential Expression Testing Cautionary Note 272 included in the drop down list For something other than the default Bonferroni correction with FWER 0 05 select an adjustment procedure from the drop down list and input the overall error rate in the FWER editable field m Options FWER FDR Test Alt Hypothesis Adjustment Permut Random Seed Not equal v BY Hochberg Holm None SidakSD SidakSS Figure 7 4 Procedures for controlling Type I Error rates Note that the minP and maxT procedures are only available for the permute versions of the test statistics When you use the permutation methods you can specify the number of permutations used in the p value estimation and provide a seed to the random number generator for repeatability of results in testing or validation studies The permutation and minP and maxT adjustment procedures should not be used in the con
57. Tests The dialog provides traditional Testing testing methods e g paired t test Wilcoxon test with a host of correction methods to control the family wise error rate and false discovery rate Type I Error To set up the Two Sample Tests dialog follow these steps 1 In the Show Data of Type field select Two Channel 2 In the Data field select SwirlMarrayRaw norm 3 For Compare Level 1 and Compare Level 2 set the values to Swirl and WildType respectively 142 The Options Group FWER Two Sample Design 4 The Array Name field should be automatically updated to lt undetermined gt 5 Deselect the Display Output in S PLUS and select Save Output as HTML and Display HTML Output checkboxes 6 To save the test results enter SwirlMultTest in the Save As field of the Output group m Data Dutput Options Show Data of Type Two Channel v Data SwitlMarrayRav Factor Zebrafish X M Volcano Plot Y Axis Orientation negative hd Fold Change Line 2 0 Compare Level 1 Swin ha V Heat Map Compare Level 2 wildT ype j TE Chromosome Plot Array Name D Devel splus_m IV Top 15 Genes Options T Output FWER FOR ois JT Display Output in S PLUS Test pairedt z MV Save Output as HTML Alt Hypothesis Not equal SaveHTMLAs SwirMultTest html Adjustment BH zji V Display HTML Output Permutations fi 0000 Save As Random Seed SwirlMultT est
58. The Gene List Management dialog Annotation Two Way Design SurgeryANOVABsYoungBH SurgeryANOVABsYoungBH Thr Old Young 4hr Old Young 16 Figure 3 47 Venn diagram comparing gene lists for two different contrasts of the Surgery data set In addition to the annotation described in section Graphical Annotation on page 117 there is a dialog with more general annotation capabilities To open the Annotation dialog click ArrayAnalyzer gt Annotation The resulting dialog has two tabs 1 General Options 2 Filtering Options There are many options available on the Annotation dialog but for this first example we ll focus on generating the annotation for the significant genes in Surgery ANOVABsYoungBH For a more thorough description of the Annotation dialog see Chapter 9 Annotation and Gene List Management To annotate only the significant genes identified in Surgery ANOVABsYoungBH select DiffExprTest from the Show Data of Type list and select Surgery ANOVABsYoungBH from the 119 Chapter 3 Examples Affymetrix Probe Level Data 120 Data list Note the public annotation databases available from the GUI in the General Annotation group For the example we leave the three defaulted databases checked Figure 3 48 displays the resulting settings oix General Options Filtering Options r Data m Use LocusLink IDs _ Show Data of Type J Save LocusLink IDs to File DiffExpr
59. The spatial 2D normalization fits a loess surface to the intensities at the x and y spot coordinates This surface is then subtracted from the pre normalized values to center the data This procedure is often used separately on each print tip group on a chip Caution Median Absolute Deviation Normalization MAD maNormMain Function Default Normalization Parameters For maNormMain 226 Loess normalization for two channel data is computationaly intensive and computations may take an extended amount of time Scale normalization attempts to align the variability of the expression intensity across chips Yang et al 2001 2002 suggest that for scale normalization a robust estimate such as median absolute deviation MAD may be used For a collection of numbers x Xp the MAD is the median of their absolute deviations from the median m median x x MAD median x m x p m The main function for location and scale normalization of two channel microarray data is maNormMain Normalization is performed for each chip independently in a given batch of arrays using location and scale normalization procedures specified by the lists of functions f loc and f scale Typically only one function is given in each list otherwise composite normalization is performed using the weights computed by the functions a 1oc and a scale When both location and scale normalization functions f 1oc and f scale
60. an existing gene list 4 Gene Sort Order Options options to sort filtering results 310 Clustering Affymetrix Summary Data Hierarchical Examples from the GUI Ciuteranaivss l General Options Filtering Options m Data r Partitioning Methods Show Data of Type J Partitioning Around Medoids cselect gt x Margencrcaster Number of Cluster x Auto X kundetermined gt Cluster on Data Array Name r Response Variable Response x Standardization Standard value x LASET OR Hierarchical Methods IV Hierarchical Weighting method average x I euclidean hid IV Names on Graph J Model based Genes z rup IV Display Output in S PLUS I Save Output as HTML Dist metric Save H myCluster html JE Display HTML Output Save As myCluster Help Cancel Appt K j current Figure 8 2 The default Cluster Analysis dialog To set up a clustering problem start by selecting the data type from the Show Data of Type drop down list Notice that you can choose Affymetrix or two channel expression objects or differential expression test objects For our first example we ll choose the swimming mice expression object created in Chapter 2 Examples Affymetrix MAS Data If you haven t worked through this example yet go to Chapter 2 and complete it before continuing with the example below Work through the Cluster Analysis dialog set up in the
61. and assigning each observation to the group with the closest centroid The k means algorithm alternates between calculating the centroids based on the current group memberships and reassigning observations to groups based on the new centroids Centroids are calculated using least squares and observations are assigned to the closest centroid based on least squares This use of a least squares criterion makes k means less resistant to outliers than the medoid based methods which will be discussed in later sections The partitioning around medoids PAM algorithm is similar to k means but uses medoids rather than centroids The method PAM is fully described in Chapter 2 of Kaufman and Rousseeuw 1990 Compared to k means clustering PAM has the following features a it accepts a dissimilarity matrix b it is more robust because it minimizes a sum of dissimilarities instead of a sum of squared euclidean distances c it provides novel graphical displays silhouette plots and cluster plots Partitioning methods are appropriate when distinct sets of subpopulations are hypothesized Results from using the partitioning methods are typically represented with cluster biplots and silhouette plots Cluster biplots show the subpopulations separated in the first two principal component dimensions whereas silhouette plots show how well individual samples are classified In silhouette plots for each object i a sample or experimental condition t
62. are passed location normalization is performed before scale normalization That is scale values are computed for the location normalized log ratios maNormMain operates on an object of class marrayRaw or possibly marrayNorm if normalization is performed in several steps and returns an object of class marrayNorm maNormMain accepts any of the normalization methods listed in Table 6 2 The default parameters for these methods are also listed The default normalization parameters can be changed by supplying the parameters as arguments in the normalization method call as follows Two channel normalization Default Within print tip group loess location normalization of first two arrays in the swirl Examples With maNormMain Normalization Methods for Two Channel Data dataset but change default span parameter from 4 to 6 gt swirl norm lt maNormMain swirl 1 2 f loc list maNormLoess span 6 Table 6 2 Two channel scale and location normalization methods performed through maNormMain Normalization Method and Default Settings Description maNormMed Location normalization using the Defaults x NULL y maM global median of intensity log subset TRUE ratios for a group of spots maNormLoess Location normalization to a fitted Defaults x maA y maM loess curve usually for M vs A z maPrintTip w NULL subset TRUE span 0 4 maNormMAD Scale normalization using the Defaults
63. be used Specifies the path to use for all files in the block Design START relativepath C Program Files Insightful splus62 module A rrayAnalyzer examples swirl 1 spot swirl 1 Swirl WidType swirl 2 spot swirl 2 Swirl Wid Type swirl 3 spot swirl 3 WidType Swirl swirl 4 spot swirl 4 WidType Swirl Design END 379 Appendix A Creating a Design File DesignType valuenames Table J 6 Table of DesignType valuenames Value Name Rules Example Type REQUIRED Defines the type of design used Can be one of the following values only Type 0 380 Only used if Type 0 two sample Determines whether dye swapping is done True is indicated by any of the following values Y Yes T True 1 All other values indicate false 0 1 2 3 where 0 two sample 1 loop 2 reference 3 other TWO_SAMPLE_DyeSwap OPTIONAL TWO_SAMPLE_DyeSwap Table J 6 Table of DesignType valuenames Continued Format Specification Value Name Rules Example LOOP Factor OPTIONAL Only used if Type 1 loop Specifies the factor to use This factor must be a valid factor name from the FactorInfo block LOOP_Factor ZebraFish LOOP_DyeSwap OPTIONAL Only used if Type 1 loop Determines whether dye swapping is done True is indicated by any of the following values Y Yes T True 1 All other values indicate false LOOP_Dye
64. do quality checks Open the dialog by clicking through the sequence ArrayAnalyzer Quality Control Diagnostics gt Affymetrix from the main S PLUS menu bar ArrayAnalyzer Import Data gt Quality Control Diagnostics Filtering Affymetrix Expression Summary Normalization Differential Expression Analysis gt Cluster Analysis Annotation Gene List Management One Channel Affymetrix Twokiannel Figure 2 11 Opening the Quality Control Diagnostics dialog 31 Chapter 2 Examples Affymetrix MAS Data 32 The resulting dialog provides a number of options for doing visual quality checks 1 Image Plot An image plot of each array For Affy MAS data each vertical bar or line when there are many genes corresponds to a different gene MvA Plot of replicates versus each other within treatment condition or versus the median expression computed for all replicates within each treatment condition Options for the MvA plot are a a scatter plot produced by drawing a random sample of genes and b a hexbin plot of all genes where the hexagonal points are colored to give a sense of the density of points at each location Genes Present Plot A simple barplot of percent of genes present for each array Available only for MAS 4 5 data when the Detection column has been selected for filtering during the data import stage Intensity Boxplot Boxplots of expression intensities for each array
65. exprSet ChipName mgu74av2 Test LPE t test Adjustment Procedure BH FWER 0 05 47 Chapter 2 Examples Affymetrix MAS Data Gene Lists Management 48 Number of Tests 7284 Number of Significant Expressions overall 40 Number of Significant Expressions for contrast Swim3wks NoSwim4wks 40 The Gene List Management dialog allows you to merge gene lists from testing different contrasts in an analysis We set up the Mouse Swimming experiment as a one way design and we want to test a subset of the contrasts listed in Table 2 2 We do so by running two different ANOVA s choosing a different baseline each time NoSwim4wks and NoSwim4wks 1wk The same strategy could be applied to the entire data set including NoSwim10min as a baseline as well When the ANOVA runs it compares all the levels to the baseline Consequently more contrasts are computed than needed The Gene List Management dialog allows you to select the contrasts you are interested in and merge the gene lists together Open the Gene List Management dialog by clicking ArrayAnalyzer gt Gene List Management Now in Data Group 1 select DiffExpr Test from the Show Data of Type drop down list Then select the MouseSwimANOVABs4wkBH ANOVA object from the Data drop down list and select one of the contrasts listed in Figure 2 2 We start with Swim3wks NoSwim4wks Repeat this selection for the other data group Pick the same data type in Data Group 2 and the same data
66. foldChange fwer 0 05 procedure Bonferroni chip name NULL volcano plot T heatmap plot T chromosome plot F html output F summary name MultSumm open browser F JHHF BH adjustment MultSumm lt multtest graphlet M diffExprL 3 diffExpr 1 index testStat 1 foldChange fwer 1 procedure BH chip name NULL volcano plot T heatmap plot T chromosome plot F html output F summary name MultSumm open browser F 189 Chapter 4 Examples Two Color Data 190 QUALITY CONTROL DIAGNOSTICS AND FILTERING Quality Control Diagnostics Diagnostic Methods Using the GUI Filtering Array Filtering Gene Filtering 192 192 200 203 203 204 191 Chapter 5 Quality Control Diagnostics and Filtering QUALITY CONTROL DIAGNOSTICS Diagnostic Methods 192 S ARRAYANALYZER provides an assortment of graphical tools for assessing the quality of your experimental data The tools allow you to consider quality in several ways so you can make judgements on the usefulness of the data in subsequent analyses and statistical inferences You can examine an entire array as well as the individual genes or spots on an array These tools range from color images of expression intensities where each pixel represents a different gene or spot on the array to sophisticated RNA degradation plots for Affymetrix CEL data and principal components plots In this section we describe the quality diagnostics tools available and how t
67. for selecting the database to query for annotation information Selecting either one opens an HTML page in your default web browser displaying a brief description of the gene with hyperlinks to more detailed information Figure 3 23 shows an example page from LocusLink with annotation for one of the differentially expressed genes in the melanoma example e NCBI LocusLink PubMed Entrez BLAST ON Search LocusLink y Display Brief Organism All hd Query J Go Clear j Map Viewer Taxonomy Structure View Hs RFC2 One of 1 Loci Save All Loci A BICER lal Tae IAMINI CEIC RSNA ee p O O A S O EEE O A O Click to Display mRNA Genomic Alignments spanning 22906 bps Homo sapiens Official Gene Symbol and Name HGNC RFC2 replication factor C activator 1 2 40kDa LocusID 5982 Overview RefSeq Summary The elongation of primed DNA templates by DNA polymerase delta and epsilon requires the action of the accessory proteins proliferating cell nuclear antigen PCNA and replication factor C RFC REC also called activator 1 is a protein complex consisting of five distinct subunits of 145 40 38 37 and 36 5 kD This gene encodes the 40 kD OMIM ACEVIEW UNIGENE Figure 3 23 Annotation information from LocusLink 92 Two Way Design TWO WAY DESIGN Mouse Surgery Data In addition to the two sample procedures described briefly at the beginning of this chapter S ARRAYANALYZER provides ANO
68. gene When computing tests for chips with many probes setting a usual Type I Error false positive rate for individual tests will result in many false positives One key ingredient to good expression testing is controlling the family wise error rate FWER or false discovery rate FDR There is a rich historical literature on this topic in statistical journals and texts see for example Hochberg and Tamhane 1987 Westfall and Young 1993 or Hsu 1996 The net effect of poor FWER control is wasted time and money in discovering many genes that really aren t differentially expressed This topic is discussed in more detail in the section Controlling The False Positive Rate Another key ingredient to good statistical testing is obtaining good estimates of the standard error of differential expression for each gene In some studies e g with few replicates specialized methods may be required to improve the power of the statistical test We describe one approach to doing this in the section GUI for LPE Testing Statistical Tests STATISTICAL TESTS Within Gene Two Sample Comparisons The S PLUS environment is rich in methods for statistical modeling and hypothesis testing Virtually all the traditional modeling and testing methodology is available in S PLUS through either its GUI or its Command line and many through both Furthermore because of the ease of programming S PLUS many new methods quickly find their way into the S PL
69. group for the malaria parasite data Note the trend within each print tip row indicating a need for applying a normalization method that remove this print tip group bias Principal Component 2 30 Principal Component Plot kg kg a E a RE g v o o a gt x Mo T T T T T T T T 58 60 62 64 66 68 70 72 Principal Component 1 OxXHIOX D0 Figure 4 24 Plot of first two principal components with taking all expression intensities for each array as the variables 154 Filtering Two Way Reference Design The QC plots indicate the need to normalize but before we do that we should eliminate problem spots We kept the Flags column for each of the arrays A non zero flag value indicates some problem occurred in calculating the intensity value or the spot was EMPTY during the experiment We can eliminate these spots using the Filtering dialog From the main S PLUS menu bar click to the filtering dialog by doing ArrayAnalyzer Filtering The Filtering dialog has two pages 1 Array Filtering for filtering out entire arrays and 2 Gene Filtering for filtering out genes On the first tab select Data of Type Two Channel and then select the TPMarrayRaw object for filtering In the Save As field of the Output group type in TPMarrayRawFiltered to save the resulting object Leave the rest of the first tab alone because we don t want to eliminate any array in its entirety Proceed to the Gene Filtering tab
70. included with the distribution of S ARRAYANALYZER The fundamental question posed by this study is Which genes were involved or expressed during the buildup of muscle in the heart during conditioning over a period of four weeks To answer this question data for the conditioned mice were collected at 10 minutes 2 5 days 1 week 2 weeks 3 weeks 4 weeks and 5 weeks The fifth week was a week of rest for the mice so it is referred to as 4 weeks of conditioning plus 1 week of rest or more simply 4 weeks 1 week Control data was collected only at 10 minutes 4 weeks and 4 weeks 1 week The suggestion is to match experimental mice Swim with control mice NoSwim as listed in Table 2 2 Because the control data is duplicated see Table 2 1 it is not appropriate to use a two way ANOVA for the analysis even though that might be tempting The reuse of data would produce inappropriately small error estimates for tests of conditioning in a Importing Data One Way Design two way analysis increasing the Type I error The proper setup for the analysis is a one way ANOVA which tests for differences between the pairs of treatment conditions given by the rows of Table 2 2 Table 2 2 Matching experimental mice with control mice to conditioning by reusing control data Test Swim No Swim 1 3 reps 10 min 3 reps 10 min 2 3 reps 2 5 days 3 reps 10 min 3 3 reps 1 wk 3 reps 10 min 4 3 reps 2 w
71. ja http david niaid nih gov david tools asp Go Links 2 Annotation Tool An automated method for the functional annotation of genome scale datasets GoCharts A visualization tool that graphically displays the distribution of differentially expressed genes among functional categories KeggCharts A visualization tool that graphically displays the distribution of differentially expressed genes among metabolic pathways DomainCharts A visualization tool that graphically displays the distribution of differentially expressed genes among functional protein domains EASEonline Provides statistical methods for discovering enriched biological themes within gene lists For detailed information please see the EASE FAQ H JETE a a Figure 9 14 DAVID EASE Web site analysis selections F http david niaid nih gov david easeresults asp Microsoft Internet Explorer File Edit View Favorites Tools Help Back gt O A A Gsearch GFavortes meda lt 4 B 3 a Address ja http david niaid nih gov david easeresults asp z Go Links ied EASE RESULTS Fisher System Category PH Exact cell growth and or maintenance biological_process 0357 biological_process cellular physiological process 0587 biological_process metabolism f 18 biological_process cellular process K 197 cellular_component lintracellular 214 molecular_function binding 247 biologi
72. of summarized data CEL ASCII CEL Raw CEL data CEL binary CEL Binary raw CEL data 386 Note that compressed binary files are not supported in this version of S ARRAYANALYZER ASCII Data Excel Data Introduction For either Affymetrix Summary MAS or two channel ASCII data the ASCII files must meet these requirements 1 The data for each array is in a separate file For two channel data data for both channels must be in the same file 2 The data must be delimited by a single common delimiter typically a tab or comma 3 The data can have a header the end of the header section must be a line of columns names for the data The column names must be unique If there is the no header then the first line is the column names row 4 From the column names row to the end of the file each line row must have the same number of delimiters 5 Each file must have the same number of lines If there is a header section each file must have the same number of lines in the header 6 A tail section is not allowed The import dialogs will attempt to auto detect the number of lines in the header and the delimiter used in the file In some case this detection may fail The user can explicitly set the number of lines to skip and the delimiter on the Options tab Note that opening a delimited ASCII file in Excel and then saving it as a delimited ASCII file can result in a file that violates rule 4 above if there are blank c
73. on the MIAME tab is used for labeling tables and graphs 3 MAS Variables amp Filtering This page has default settings depending on the type of data files e g MAS4 or MASS you select It also allows the selection of other variables which can be used for more general filtering by using the Filtering dialog 4 CEL Filtering This page allows a couple of options for filtering out spots not to be used in subsequent analyses when importing probe level data 70 Two Sample Design 5 Options This page provides options for specifying the number of header lines to skip and the delimiter used in the data file Import Data From Affymetrix x File Selection MIAME MAS Variables amp Filtering CEL Filtering Options m Step 1 Specify Design Read Existing Design Create Modify Design Save Design File r Step 2 Associate Files with Design Points Fle ame Factor AL File Type cundetermined gt Array Name krequired gt 7 CDF Step 3 Save Output Save Data Set As mySet IV Display Report Figure 3 2 The Import Data Affymetrix dialog Step I Create Before we can begin to associate data files with experimental The Experiment conditions we need to set up the experimental conditions in Design S ARRAYANALYZER The easiest way to do this is through the Create Modify Design dialog Open the Create Modify Design dialog by clicking on the Create Modify Design
74. pixel on the image corresponds to a different probe spot on the array These plots can reveal systematic irregularities due to poor hybridization or problems with the array Figure 3 32 displays a plot of the first two principal components for all the arrays Each point in the plot corresponds to a different array Different symbols correspond to different experimental conditions Two Way Design Figure 3 33 displays an RNA degradation plot Each line represents the sequence of average expression intensities at each probe location in a probe set averaged across all probe sets on one array Different colors represent different experimental conditions Trends in the lines indicate uneven hybridization over the probe set 600 1 500 300 200 T T T T T T T 0 100 200 300 400 500 600 Figure 3 31 Image plot for one of the arrays in the SurgeryAffyBatch dataset 103 Chapter 3 Examples Affymetrix Probe Level Data 104 Principal Component Plot o o v 34 N 8 x oq F S My ny v o o o a x oO xe c T ao A i o4 ki o SJ 4 N E T T T T 560 580 600 620 Principal Component 1 dox bD0 Old Ohr Old 1hr Old 4hr Young Ohr Young 1hr Young 4hr Figure 3 32 Principal components plot of the experimental conditions in the Surgery study Different points correspond to different arrays Plotting symbol indicates experimental condition Mean
75. probe level object Clicking OK creates the normalized affyBatch object and plots pre and post normalization boxplots for comparison The plot is on a logy scale but the expression intensities are saved on the original raw intensity scale Figure 3 35 displays the results from computing invariantset normalization on probe level expression intensities 105 Chapter 3 Examples Affymetrix Probe Level Data Expression Summaries 106 Before invariantset Normalization After invariantset Normalization HANI Sigua j i 4 NJ NJ z ce g TT ae gt S i 1 ae eo 2 a Ll non a IT Pll d dl PITT dd J DESEEN eT Pan geia joie MeL TE EE LEET LEIFER La LE of III t LIT dt 4 A Lge PE ela silt LiL l i My THT o4 T T T 5 o T T T TT TT riL eT uty ee TILL ET aba i dj itl L ddal TA dij il wih hy EVSLVSLVELVELVEecwVe TYUCULVLUCVLEVEEVLeVe let ait pl atta tt ate ag ap ay yl a ae an aia cocecococcococr coool ooorrryryTTooorrrTrittTtTt ooofcere Te eCooor sr HK ETT DODVVDVVVVY 290090900000 DOVDVVVVVBGIOBOOVIIVOOD
76. probe specific background correction subtracting MM normalization and summarization of the probe sets The input to this dialog is an AffyBatch class object created from the import dialog or at the command line The output is an object of class exprSet and graphical representation of the resulting summarized data The graphical output is discussed in section Affymetrix Diagnostic plots on page 234 Summarization and Correction Options This dialog allows you to choose either GC RMA RMA or to Mix amp Match correction summarization and normalization methods GC RMA as discussed in section Background Correction with gc rma on page 239 is primarily a background correction method From this dialog however the GC RMA option also normalizes the arrays by quantiles and summarizes the probe sets using medianpolish Choosing RMA will return data that has been corrected using rma background correction pmonly PM correction quantiles normalization and medi anpolish to summarize the probe sets 249 Chapter 6 Pre Processing and Normalization The Mix amp Match option allows you to be very creative in how the final summarized data is created However not all combinations of techniques make sense It is up to the user to know which options make sense to use at the same time 250 Normalization Methods for Affymetrix MAS Data NORMALIZATION METHODS FOR AFFYMETRIX MAS DATA Affymetrix data typically arrives as DAT CEL and CHP files Th
77. results from clustering on experimental conditions The legend at the bottom show the relationship between colors in the heat map and logged expression intensity values 313 Chapter 8 Cluster Analysis Partitioning 314 wim4w3 geninan lipocalin 2 exp ressed sequence C ngiotensino en erallothione n2 le en nhanced anti period homolog 2 PTOXISOME rolife ra S ongaipn tae RRA RIKEN RBS B tumor necrosis factor mitogen Smetne thioe per S methyl a IKEN cDNA RIKEN cDNA 061 inactive X specific t insulin receptor subs ReRe A cl RIRE BNA 3030857 hemokine C C motif cDNA sequence BC02 myosin lig ht polypep osin li jolype Re ght MAG a BP antigen delta po tedsarcoma onc histocompatibilit b en identifie pro Bap type R peuroblastoina myc re protein tyrosine phos aminglevuli acid s Bers 447480 00t ore OTZ oe tv0 vv0 oel ozz ore 00 Figure 8 5 Hierarchical clustering output for the swimming mouse example expression data The clustering dendrogram is displayed on the left side of the graph with gene tables on the right The dendrogram at the top corresponds to the experimental conditions The legend across the bottom indicates the values of the expression intensities displayed in
78. reveal nothing extraordinary about the arrays in the Mouse Swimming study so we ll move onto Normalization Now that the data has been imported and checked for quality we are ready to move to the next step of the analysis procedure normalization The Normalization dialog is designed to remove artifacts and systematic variation resulting from the measurement process The goal is to remove variability not due to differential expression so that differential expression is estimated accurately for each gene Note that we need to be careful not to normalize so aggressively as to wash out signal Typically this is accomplished by normalizing within experimental conditions although some forms of normalization may be comfortably applied across experimental conditions For our swimming mouse example this translates to normalizing within each conditioning level Swim or NoSwim and each observation time 3 wks 4 wks and 4 wks 1wk Normalization Dialog The Data Group One Way Design To normalize the expression data select Array Analyzer gt Normalization from the main menu ArrayAnalyzer Import Data gt Quality Control Diagnostics Filtering Affymetrix Expression Summary X Differential Exp Cluster Analysis Annotation Gene List Management ion Analysis gt Figure 2 18 Selecting the Normalization menu item Show Data of Type The Normalization dialog requires you to select the type of data you are
79. same value quantiles robust x weights NULL remove extreme variance n remove 1 approx meth FALSE Quantile normalization with options to Eliminate chips with high variability Eliminate chips with means too disparate from others Down weight particular chips in the computation of the mean vsn subsample 20000 niter Variance stabilizing normalization 243 Chapter 6 Pre Processing and Normalization An Example with Below we use the melanoma data set Fox et al 2001 to normalize 244 demonstrate various normalization procedures The melanoma dataset is discussed in section Melanoma Data on page 69 We first read in the data This can be done through the GUI as shown in section Importing Data on page 95 of Chapter Examples Affymetrix Probe Level Data gt directory lt paste getenv SHOME module ArrayAnalyzer examples sep gt cgnames lt paste directory c cg2a CEL cg2b CEL cg24a CEL cg24b CEL sep gt NCImelanoma lt ReadAffy filenames cgnames The data should be corrected for specific binding and background noise One way to do this is to simulate the Affymetrix MAS 5 0 software as follows j correct melanoma CEL data d background correct gt NCImelanoma lt bg correct NCImelanoma method mas Correct using MM as controls gt tmp lt pmcorrect mas NCImelanoma i Add the correct PM values back into melanoma
80. specified number A e g 5 log2 scale in at least k user specified experiments This is useful for filtering out genes that don t exhibit much expression in any experimental run Gene List Filtering e Most of the analysis functions in S ARRAYANALYZER can produce a gene list Gene lists are managed in the Gene List Management dialog where they can be intersectioned or unioned These gene lists can be used to filter any of the data objects being annotated Cluster Filtering The partitioning cluster analyses performed on genes return a class membership for each gene These class memberships can be used to filter any of the data objects being annotated 344 Annotation Libraries Gene Sort Order Options Limit number of genes which puts a cap on the gene list being annotated This is particularly important for the GenBank databases if you send too many queries in too short a time to GenBank your IP address may be recorded and you may be blocked from additional future use Recalculate button which allows you to monitor the number of genes in the gene list you are constructing through the filtering choices Sort order which allows you to sort the gene list you are constructing in order of the individual filters For example if you have the filtered gene list capped at 100 genes and the number of genes selected by the filtering is greater than 100 the genes are entered into the gene list in order of the filters in t
81. subset of 8448 spots Control spots There are 7681 types of controls control fb1l6a01 fb16a02 fb1l6a03 fb1l6a04 fb16a05 fb16a06 768 1 1 1 1 il 1 Notes on layout 175 Chapter 4 Examples Two Color Data Reading Experiment Information 176 C PROGRAM FILES INSIGHTFUL splus62 module ArrayAnalyzer examples fish gal Note that the column that had control indicators also had gene ID s we need to correct that in the swirl 1layout object We do that by using a couple of utility functions maNspots and maControls gt controls lt rep control maNspots swirl layout gt controls maControls swirl layout control lt N gt maControls swirl layout lt factor controls gt swirl layout Array layout Object of class marrayLayout Total number of spots 8448 Dimensions of grid matrix 4 rows by 4 cols Dimensions of spot matrices 22 rows by 24 cols Currently working with a subset of 8448 spots Control spots There are 2 types of controls Control N 768 7680 Notes on layout C PROGRAM FILES INSIGHTFUL splus62 module ArrayAnalyzer examples fish gal The functions maControls and maNspots return the control vector and the number of spots respectively for an object of class marrayLayout This step reads the file with the experimental design information For the example shown the informations is contained in the file SwirlSample txt located in the examples directory of the ArrayAnalyzer mo
82. working with Click the drop down button on the Show Data of Type field and select one of the choices For the melanoma example select Affymetrix Summary inii r Data r Normalization Show Data of Type Normalization E lt select gt z Between Aran Z Date EAE MvA Plot SaveAe Affymetrix CEL y FEMALE Two Channel IV Box Plot robe Se gt Ph Pigia Jai Jr When to Show Before amp After h and MM FM and C Only After Help Cancel Apply j current Figure 2 19 Selecting the data type for normalization Data Click the drop down button to right of the Data field and select the expression object created during the import step MouseSwimExprSet Save As Enter the object name for saving the normalized expression data in the Save As field By default this is set to MouseSwimExprSet norm the name of the object you select in the Data field with norm attached as a suffix 37 Chapter 2 Examples Affymetrix MAS Data The Normalization Group 38 In the Normalization group set the Normalization field to medianIQR select the MvA plot check box select the Box Plot check box and click the radio button to select Before amp After for pre and post normalization boxplots The normalization procedures for MAS 4 5 summary data are described in greater detail in Chapter 6 Pre Processing and Normalization For this example we select the default setting as medianIQR which adj
83. 0 gt maBoxplot swirl norm 3 main Post normalization srt 90 The resulting graph is display is shown in Figure 4 46 185 Chapter 4 Examples Two Color Data Pre normalization Post normalization Scale Print Tip MAD dee J00 ce ewe oc o eee qeeee eeeme oo 25d eal 3 s o a woos E wa oe u Hou ewe E ee Se ee E PrintTip PrintTip Figure 4 46 Before and after scale print tip MAD normalization of the swirl data Before we move onto differential expression testing note that the slots of the marrayRaw object are gt getSlots swirl raw maRf maGf maRb maGb maW matrix Matrix matrix matrix matrix maLayout maGnames maTargets maNotes marrayLayout marrayInfo marrayInfo character Each of the first four slots are raw intensity matrices with dimensions equal to number of genes x number of chips after controls have been removed For this example 7680 x 4 186 Differential Expression Testing Paired t test From the Command Line gt dim swirl raw maRf 1 7680 4 Once we apply the normalization procedures background correction is done the raw intensities are converted to M and A values and the and the normalized object has different slot names gt getSlots swirl norm maA maM maMloc maMscale maW maLayout matrix matrix matrix matrix matrix marrayLayout maGnames maTargets maNot
84. 0 35966_at 2509 5 368228 0 0 37420_i_at 2884 3 129978 0 0 37421_f_at 2885 3 034635 0 0 37484 _ at 2948 3 262691 0 0 38228_g_at 3214 3 169818 0 0 38555_at 3301 4 589305 0 0 30928 r t 3435 3 207205 0 0 40671_g_at 3980 3 205296 0 0 40755_at 4064 3 046220 0 0 The p values have been sorted from smallest to largest so printing the first 10 rows prints the 10 most statistically differentially expressed genes 63 Chapter 2 Examples Affymetrix MAS Data 64 It s worth noting that fold change values less than two in absolute value are significant if their standard errors are relatively small In this experiment amongst the 25 most significant genes there are three with very significant differential expression but with absolute fold change less than two These genes would have not made the cut using a straight fold change approach to gene discovery References REFERENCES Fox J W Dragulev B Fox N Mauch C and Nischt R 2001 Identification of ADAM9 in human melanoma Expression regulation by matrix and role in cell cell adhesion Proceedings of International Protelysis Society Meeting Lee JK and O Connell M 2003 An S PLUs Library for the Analysis of Differential Expression To appear in The Analysis of Gene Expression Data Methods and Software Edited by G Parmigiani ES Garrett RA Irizarry and SLZeger Published by Springer New York 65 Chapter 2 Examples Affymetrix MAS Data 66 EXAMPLES A
85. 1 of 962 Figure 9 12 Affymetrix NetAffx GO browser with list of IDs for Melanoma data with fold change set to greater than 2 and p value set to significant adjusted p lt 0 05 for LPE test Bonferroni FWER in this case top 10 genes based on p value followed by fold change 353 354 Chapter 9 Annotation and Gene List Management http david niaid nih gov david upload1 asp Microsoft Internet Explorer File Edit View Favorites Tools Help Back gt A A Asearch Ejravorites Meda S Eh 3 w a Address amp http david niaid nih gov david upload1 asp Po Links UPLOAD GENE LIST With the browse button choose an input file containing a gene list in the required format then submit the list with the Submit File button Alternatively paste a list into the text box below AFFYID LOCUSLINK UNIGENE GENBANK C Program Files insigh Browse Submit File PASTE GENE LIST Type or Paste a gene list into the text box choose the appropriate delimiter and submit the list with the Submit Text button C AFFYID LOCUSLINK UNIGENE GENBANK E OUE Ec eee Figure 9 13 DAVID EASE Web site launched by StARRAYANALYZER with list of IDs for Melanoma data Annotation Libraries vid niaid nih gov david tools asp Microsoft Internet Explorer File Edit View Favorites Tools Help Back gt amp A A Bsearch GgFavorites media B 3 fe a Address
86. 138 Differential Expression Testing 142 Two Way Reference Design 147 Malaria Parasite Data 147 Importing Data 148 Quality Diagnostics 151 Filtering 155 Quality Diagnostics Revisited 157 Normalization 158 Clustering Expression 161 Differential Expression Analysis 164 Gene List Management 166 Annotation 168 From the Command Line 174 Importing Data 174 Filtering Out Controls 180 Quality Diagnostics 180 Normalization 183 Differential Expression Testing 187 125 Chapter 4 Examples Two Color Data TWO COLOR DATA ANALYSIS WORKFLOW 126 The process of analyzing differential expression for custom cDNA arrays can be done through the S ARRAYANALYZER menu and dialogs To obtain differential expression test results from two color cDNA microarray data we go through four fundamental steps 1 Importing and filtering the data 2 Background adjustment and normalization 3 Differential expression analysis 4 Gene list management and annotation In addition S ARRAYANALYZER provides methods for quality diagnostics and clustering to round out its feature list The two examples of cDNA data we use are summarized means and medians across all spots with identical probes When we import the data we specify background intensity columns so adjustment for background intensity levels can be made prior to normalization and testing Two Sample Design TWO SAMPLE DESIGN Mutant Zebra Fish Data S ARRAYANALYZER has an assortment of procedu
87. 2 flan r General Annotation LocusLinkList tt 3 m Open OntoE xpress I LocusLink esr IV OntoExpress I Uni ae Distribution Binomial he J Pubmed Correction X I GO Website Username moconnell m Use Affymetrix FF Pees I Save Affy IDs to File Affpmetiy ID File ProbeList tet F Ope F Open DAVID EASE Browser metriz GO Browser Cancel Apply if j current Figure 9 18 The General Options page of the Annotation dialog Options chosen are from the Open OntoExpress group Annotation from Command Line Scripting GenBank Metadata Lookups Annotation Libraries OntoExpress Results Display Legend User Interactions B Unselected Function E Synchronized Function m Selected Function Searched Function Functional Catagories Observed M More Than Expected Search Function ana frotar gt 0 setect results ana p value lt 3y Total jj Search Display Molecular Function Sort By Pvalue Clear Search Input Save Onto Express Results rProgram Less Than Expected E Same As Expected Gene Regulation M Positive Negative m No Change Draw Selected RunOnto Design Run Onto Compare Gene_Ontology o molecular_function o binding o metalion binding o magnesium ion binding 1 p 0 5 102838 st fold change 0 0
88. 2544 0 33 561404 0 DA 102745_at 384 5 062317 0 39 989042 0 EN 104093_at 497 4 614482 0 48 593687 0 ee 98000_at 505 2 239734 0 17 949190 0 ere 101820_at 514 5 753799 0 63 326020 0 T 297 Chapter 7 Differential Expression Testing REFERENCES 298 Benjamini Y Hochberg Y 1995 Controlling the false discovery rate a practical and powerful approach to multiple testing Journal of the Royal Statistical Society Series B 57 289 300 Benjamini Y Yekutieli D 2001 The control of the false discovery rate in multiple hypothesis testing under dependency Annals of Statistics 29 4 1165 1188 Dudoit S Shaffer J P and Boldrick J C 2002 Multiple hypothesis testing in microarray experiments U C Berkeley Division of Biostatistics Working Paper Series Working Paper 110 Dudoit S Yang Y Callow M and Speed T 2002 Statistical methods for identifying differentially expressed genes in replicated cDNA microarray experiments Statistica Sinica 12 111 139 Efron B Tibshirani R Storey J D and Tusher V 2001 Empirical Bayes analysis of a microarray experiment Journal of the American Statistical Association 96 1151 1160 Hochberg Y 1988 A sharper Bonferroni procedure for multiple tests of significance Biometrika Vol 75 800 802 Hochberg Y and Tamhane A C 1987 Multiple Comparison Procedures New York Wiley Holm S 1979 A simple sequentially rejective multiple test proc
89. 33aAnnoData hgul33acdf hgu95av2AnnoData hgu95av2cdf KEGGAnnoData LPEtest marrayClasses marrayInput marrayNorm marrayPlots matchprobes mgu74av2AnnoData mgu 7 4av2cdf multtest ROC and vsn All except LPEtest are ported from the Bioconductor project In this release of S ARRAYANALYZER version 2 0 new libraries have been added including vsn gcrma and matchprobes Note that these libraries are loaded automatically when you load S tARRAYANALYZER The following shows a few examples of how S tARRAYANALYZER can process your data e Analysis of Affymetrix data uses the Biobase and affy libraries for reading and normalizing data e Differential expression analysis uses the multtest and LPEtest libraries and annotation is completed using the genefilter geneplotter and annotate libraries e Input and normalization of cDNA data uses the marrayClass marrayInput marrayNorm and Biobase libraries Welcome Most of the ported libraries in S ARRAYANALYZER are based on the Bioconductor 1 3 libraries with some based on 1 4 Information on these libraries is located at http www bioconductor org The other library LPEtest is provided by Insightful Chapter 1 Introduction To Microarray Data SUPPORTED PLATFORMS AND SYSTEM REQUIREMENTS Installing and Running S ARRAYANALYZER Online Help S ARRAYANALYZER is supported on the following platforms e Windows 2000 e Windows 2003 Server e Windows XP Professional The m
90. 4 5 6 7 8o gName mean Ohr mean 24hr foldChange testStat rawp adjp signif p 1 35704at 9 24 0 54 8 70 188 04 0 00 0 11 T Ea P 37023at 37023_at 8 74 0 54 8 20 174 87 0 00 0 11 T 3 33532_at 33532_a 7 78 10 90 3 13 3621 54 0 00 0 11 T 4 37712_g_at A 8 47 0 54 7 93 163 33 0 00 0 11 T 5 31979_at 31979_at 7 21 0 54 6 67 142 71 0 00 0 11 T 6 1837_at 1837_ati 7 44 0 54 6 90 164 14 0 00 0 11 il 7 41848_f_at 41948_f_a 8 59 0 54 8 05 150 70 0 00 0 12 T 8 1984s at 1984_s_a 8 43 0 54 7 89 199 13 0 00 0 12 T 9 41231_f_at 41231_f_a 12 93 13 80 0 87 115 31 0 00 0 13 T 10 36250_at 36250_at 8 62 9 07 0 45 109 52 0 00 0 13 T aza 37777_at 4 75 0 54 4 21 98 81 0 00 0 13 T Figure 7 22 The first few rows of the gene list summary table generated by the Two Sample Test dialog Open the S PLUS Object Explorer by clicking the Object Explorer tool bar button displayed in Figure 7 23 ArrayAnalyzer Graph Options Window Help rye 55 UB HE 34 BE iai Linear Object Explorer Figure 7 23 S PLUS Object Explorer tool bar button 295 Chapter 7 Differential Expression Testing From the Command Line 296 Under the Data tree in the Object Explorer double click the summary object of your choice See Figure 7 24 Object Explorer Contents of D Microarrays ArrayAnalyzer BHMultTestSumm E Data MIAME A aa adjpObj Agilent Agilent2 Agilent3 Agilent4 B BHMultTastSumm CG a CG controls
91. 4 Summary Data MAS Summary Data Probe Level CEL Affymetrix CHP ChipName REQUIRED ChipName mgu74av2 For Affymetrix imports used as the Array Name For Two Channel import used as the layout object name In this case the name specified must be the name of an existing layout object previously created using the Create Layout dialog in ArrayAnalyzer 375 Appendix A Creating a Design File Table J 4 Table of ImportInfo valuenames Continued Value Name Rules Example CDF Path OPTIONAL Only required for Affymetrix imports where FileType is Affymetrix CHP In this case the full pathname to the CDF file is specified Not required for Two Channel imports CDFPath d mycdfs cdf1 cdf SaveAs 376 OPTIONAL Specify the name of an S object to save the imported data WARNING If you specify the name of an existing S object it will be overwritten SaveAs MouseSwimExprsSet Format Specification Table J 4 Table of ImportInfo valuenames Continued Value Name Rules Example PrintOutput OPTIONAL PrintOutput 1 If not specified no output is printed Specify whether to display information about the import in a report window in S PLUS You may specify any of the following values for this valuename to indicate true Y Yes T True 1 all other values are interpreted as a false value FactorInfo The lines in a FactorInfo block are different th
92. AVID EASE New gene list management Allows you to merge gene lists from testing different contrasts in an analysis Enhanced data normalization methods Added vsn and quantiles to Affymetrix summary data vsn to Affymetrix CEL and a between array quantile normalization method to two color experiments Expanded differential expression testing This uses the LPEtest and multtest libraries and 1 way and 2 way ANOVA New quality control diagnostics Provides an assortment of graphical tools for assessing the quality of your experimental data Enhanced S PLUS Graphlets For annotation and exchanging information among researchers Chapter 1 Introduction To Microarray Data Goals Libraries Our goals are simple for S ARRAYANALYZER 1 Create rigorous statistical analysis that is easily initiated through either the menu or the command line 2 Generate interactive tabular and graphical reporting that targets biologists and non clinical statisticians 3 Develop solution and extensible environments for customized applications For the latest information and support on S ARRAYANALYZER go to http www insightful com support ArrayAnalyzer This contains information regarding Insightful efforts in the genomics and bioinformatics space There are 15 function libraries and nine data libraries included in S ARRAYANALYZER to assist in your analysis affy annotate Biobase edd gcrma genefilter geneplotter GOAnnoData hgu1l
93. AYANALYZER We also illustrate in this section remote access and gene list upload to the Stanford Source site information from Stanford Source is used in the Bioconductor construction of the annotation data used in S ARRAYANALYZER Figure 9 10 shows the General Options page of the Annotation dialog with options chosen from Use Affymetrix IDs group This writes out a file of Affymetrix IDs ProbeList txt by default corresponding to the genes selected according to the dialog options This ProbeList txt file can be uploaded to a variety of Web sites for annotation and gene list analysis Two options that are made easily accessible with St ARRAYANALYZER are the Affymetrix NetAffx site and the NIH DAVID EASE site These two sites may be opened from the StARRAYANALYZER Annotation dialog so that all the user needs to do is browse to the ProbeList txt file and upload it Figures 9 11 to 9 15 on the following pages show screen shots from the Affymetrix NetAffx site and the NIH DAVID EASE site as launched and uploaded from S tARRAYANALYZER General Options Filtering Options r Data Annotation Libraries Show Data of Type DitfExprT est x Data eal PEBon x r Use LocusLink IDs J Save LocusLink IDs to File LocusLinkList txt LocusLink File Browse Array Name fh 95av2 EA cata F Open Stanford Source r General Annotation Dpen OntoE z m Open xpress J LocusLink E I OntoExpress I Unigene Distribution Binomial g
94. CG N cg24a cg24b cga cgb diffExpr Fold change geneNames mgu 4 a a a a a a sa sa a a sa gm gm m a a a m m m ei LCG w D oO a Data Class Dimensi Agname IR mean Ohr IR mean 24hr foldChange testStat rawp adjp signif p Locus Link Acc Num rFPWUOON ANA UNE character numeric numeric numeric named numeric numeric logical integer character 12550 12550 12550 12550 12550 12550 12550 12550 12550 12550 Figure 7 24 The Object Explorer in S PLUS allows you to browse your data files and open them in a grid for viewing To access the gene list from the Command line you need the object name The default output names for the Two Sample Test LPE Test and ANOVA dialogs are myMultTest myLPETest and myANOVA respectively For an object with FDR set to 0 001 and Benjamini Hochberg adjustment the first 5 rows of one object named myLPETest looks as follows gt slotNames myLPETest 1 allData dataInfo means annoData gt myLPETest al1Data 1 10 GeneName GeneIndex foldChange Pvalue 63 4 208443 97763_at 97763_at T 160169_at 160169_at T 64 1 178213 contrData testInfo testStat AdjPvalue Signif p 0 44 932855 0 0 8 542566 0 Differential Expression Summary Table Output 101918_at 101918 _at 80 1 074805 0 9 563393 0 aie 97528_at 113 1 639526 0 11 371125 0 UEA 101475_at 284 1 186268 0 8 712748 0 re 102744_at 383 4 55
95. ENCE_Factor ZebraFish REFERENCE_RefLevel ZebraFishl REFERENCE_RefType 2 DesignType END Design START relativepath C program files insightful splus62 module arrayanalyzer examples swirl 1 spot swirl 1 ZebraFishl ZebraFish2 swirl 2 spot swirl 2 ZebraFishl ZebraFish2 swirl 3 spot swirl 3 ZebraFish2 ZebraFishl swirl 4 spot swirl 4 ZebraFish2 ZebraFishl Design END APPENDIX B IMPORTING DATA Introduction 386 Supported Affymetrix MAS CEL Data Formats 386 ASCII Data 387 Excel Data 387 Layout Information for Two Channel Data 388 385 Appendix B Importing data INTRODUCTION Supported Affymetrix MAS CEL Data Formats Table B 5 Supported Affymetrix MAS CEL data formats S ARRAYANALYZER can import many different microarray data file formats For example Affymetrix data files are available in several forms as shown in the table below The special binary formats are auto detected by the import dialogs Two channel cDNA data is typically in ASCII or Excel files This appendix provides guidelines for importing data successfully into StARRAYANALYZER The table below lists the various Affymetrix MAS CEL data formats supported by StARRAYANALYZER 2 0 Format Type Extension Notes MAS 4 ASCII txt csv Older Affymetrix format infrequently used MAS 5 ASCII txt csv Standard summarized Affymetrix data MAS 5 Excel xls Excel version of MAS 5 file CHP binary CHP Binary version
96. ER are always clear text files and record specific information from the first page of the data import dialogs The information is written between special keys that mark the beginning and end of blocks that contain data related to a specific area of the dialog Most of these blocks are optional however a block if present must have a START and an END key A specific format is required for S ARRAYANALYZER to identify and load the data from the file The following documentation can be used to decode the meaning of the keys and to help decipher the required format Format Specification FORMAT SPECIFICATION Special Keys The following describes a specification for the format of S ARRAYANALYZER 2 0x design files along with rules that must be followed when writing out or modifying these files Design files may have any extension but must be clear text files All the lines are read by S ARRAYANALYZER at one time These lines are then parsed to find keys and data blocks containing data used to populate the first import dialog page Blank lines are ignored The design file must begin with a line containing the S ARRAYANALYZER version key V2 01 This version key changes from version to version and represents the version used for the file The current version is 2 01 If this line is not found the file is interpreted as a version 1 1 or earlier file for backward compatibility Note that any line beginning with a semi colon is interp
97. FFYMETRIX PROBE LEVEL DATA Affymetrix Probe Level Data Analysis Workflow 68 Two Sample Design 69 Melanoma Data 69 Importing Data 70 Normalization 77 Expression Summaries 79 Differential Expression Analysis 84 Annotation 92 Two Way Design 93 Mouse Surgery Data 93 Importing Data 95 Quality Control Diagnostics 101 Normalization 105 Expression Summaries 106 Differential Expression Analysis 110 Gene List Management 117 Annotation 119 References 123 67 Chapter 3 Examples Affymetrix Probe Level Data AFFYMETRIX PROBE LEVEL DATA ANALYSIS WORKFLOW The process of analyzing Affymetrix probe level gene expression data can be done through the StARRAYANALYZER menu To obtain differential expression information from probe level microarray data we perform the following basic steps 1 Import and filter the data 2 Normalization including e Adjustment for background noise e Mismatch correction and e Distribution based normalization 3 Summarize 4 Differential expression analysis 5 Annotation In addition examining array quality and filtering out bad arrays and genes may be necessary and is typically done between import and normalization Clustering is also a normal part of gene expression discovery and may be performed between all the major steps of the analysis 68 Two Sample Design TWO SAMPLE DESIGN Melanoma Data S ARRAYANALYZER has an assortment of procedures for doing two sample differential expression analys
98. Format Specification ChipName mgu74av2 CDFPath SaveAs myExprSet PrintOutput 1 ImportInfo END FactorInfo START A CondTime Swim3wks Swim4wks Swim4wkst lwk NoSwim4wks NoSwim4wks 1wk FactorInfo END Design START relativepath C Program Files Insightful splus62 module ArrayAnalyzer examples Swim3wl txt Swim3wks1 Swim3wks Swim3w2 txt Swim3wks2 Swim3wks Swim3w3 txt Swim3wks3 Swim3wks Swim4wl txt Swim4wks1 Swim4wks Swim4w2 txt Swim4wks2 Swim4wks Swim4w3 txt Swim4wks3 Swim4wks Swim4wlwl txt Swim4wks 1lwk1 Swim4wkst lwk Swim4wlw2 txt Swim4wks lwk2 Swim4wkst lwk Swim4wlw3 txt Swim4wks 1lwk3 Swim4wks lwk NoSwim4wl txt NoSwim4wks1 NoSwim4wks NoSwim4w2 txt NoSwim4wks2 NoSwim4wks NoSwim4w3 txt NoSwim4wks3 NoSwim4wks NoSwim4wlwl txt NoSwim4wks 1wk1 NoSwim4wks 1lwk NoSwim4wlw2 txt NoSwim4wks 1lwk2 NoSwim4wks 1lwk NoSwim4wlw3 txt NoSwim4wks 1wk3 NoSwim4wks 1lwk Design END The following is designSwirl txt the design file used in the section Two Sample Design in Chapter 4 Examples Two Color Data The design format is quite similar to that used for importing Affymetrix MAS or CEL data V1 3 ImportInfo START ChipName Swirl Layout 383 Appendix A Creating a Design File 384 SaveAs Swir MarrayRaw PrintOutput 1 ImportInfo END FactorInfo START A ZebraFish No ZebraFishl ZebraFish2 FactorInfo END DesignTypeLSTART Type 0 TWO_SAMPLE_DyeSwap 1 LOOP_Factor ZebraFish LOOP_DyeSwap 1 REFER
99. GUI a maximum of 2000 genes are plotted If the chips have more than 2000 genes then a random sample of 2000 genes are plotted MVA plot 20A 0 22 20B A Figure 6 5 MvA plot of one treatment group of Dilution experiment 236 Plots from the Command Line Background Correction Pre Processing and Normalization for Affymetrix Probe Level Data Table 6 5 lists the diagnostic plots available in S ARRAYANALYZER from the command line The functions box plot hist and image are methods which work on AffyBatch objects See the AffyBatch class help file The input to mva pairs is the matrix of expression measures usually the log intensity matrix is used Table 6 5 Exploratory data analysis plots available from the command line for Affymetrix probe level data Function Name Description boxplot Box plot of log base 2 of intensity matrix hist Calls plotDensity Plots the non parametric density estimates of the given matrix mva pairs MvVA plots image Raw image plots can be used to detect spatial artifacts plotAffyRNAdeg Requires object returned from Af fyRNAdeg RNA degradation plots aid in assessment of RNA quality Expression intensity measurements are summaries of the fluorescence intensities for the pixels contained within each chip spot The background of the chip contributes to this signal and the background noise levels may not be consist
100. Genes Present Plot If you want to do a Genes Present Plot available on the Quality Control Diagnostics dialog you must select the Detection variable from the Extra Variables list and move it to the Keep list by clicking the button Other Options The last page on the Import Data From Affymetrix dialog is the Options tab The tab provides two options used during data import 1 The number of header lines to skip in each file before reading the data Normally this can be detected automatically but it is provided as an option for unusual cases where auto detection can not find the row with column names 2 The delimiter separating the fields in each line of the data files Normally these are tabs but alternative choices include comma space colon and semi colon Normally the delimiter can be detected automatically but this option is provided for unusual cases where auto detection can not determine the field delimiter 30 One Way Design Press OK when you have completed the dialog and the data are imported It is now ready for use in S ARRAYANALYZER Import Data From Affymetrix x File Selection MIAME MAS Variables amp Filtering CEL Filtering Options Import Options Header lines to skip Auto X Data delimiter auto E Figure 2 10 The Options tab of the Import Data From Affymetrix dialog Quality Once the data is imported open the Quality Control Diagnostics Diagnostics dialog to
101. HTML graphical output pops up a menu for selecting the database to query for annotation information Selecting either one opens an HTML page in your default web browser displaying a brief description of the gene with a hyper link to more detailed information Figure 3 45 shows an example page from LocusLink with annotation for a differentially expressed gene ETN NCBI LocusLink PubMed Entrez BLAST OMIM ETS r Taxonomy Structure Search LocusLink y Display Brief v Organism All hd Query Go Clear View Mm Rom1 z One of 1 Loci Save All Loci ABCDEFGHIJKLMNOPQRSTUVWXYZ Top of P Click to Display mRNA Genomic Alignments spanning 1796 bps f cone T us Tumoene MARET var nomor uct ct Ka Mus musculus Official Gene Symbol and Name MGI Roml rod outer segment membrane protein 1 LocusID 19881 Overview Ti Locus Type gene with protein product function known or inferred Product rod outer segment membrane protein 1 Alternate Rom 1 Symbols Function Submit GeneRIF All Pubs Gene Ontology Term Evidence Source Pub e G protein coupled photoreceptor activity TEA MGI e cell adhesion TEA MGI Figure 3 45 Locus Link annotation page resulting from graphical annotation query The Gene List Management dialog allows you to merge gene lists from testing different contrasts in an analysis or from different analyses The Surgery experiment is a two way design with several contrasts of
102. Insightful S ARRAYANALYZER 2 0 User s Guide June 2004 Insightful Corporation Seattle Washington Proprietary Notice Copyright Notice Trademarks Insightful Corporation owns the StARRAYANALYZER software program and its documentation Both the program and documentation are copyrighted with all rights reserved by Insightful Corporation S ARRAYANALYZER provides access to the Bioconductor R packages for microarray analysis which are free software The affy annotate Biobase edd gcrma genefilter geneplotter LPEtest marrayClasses marrayInput marrayNorm marrayPlots matchprobes multtest ROC and vsn libraries are copyrighted 2004 by Insightful Corporation These libraries are free software that are redistributed and modified under the terms of the GNU Lesser General Public License as published by the Free Software Foundation version 2 1 of the License The S ARRAYANALYZER software is covered by a separate license agreement The correct bibliographical reference for this document is as follows S ARRAYANALYZER 2 0 User s Guide Insightful Corporation Seattle WA Printed in the United States Copyright 1987 2004 Insightful Corporation All rights reserved Insightful Corporation 1700 Westlake Avenue N Suite 500 Seattle WA 98109 3044 USA Insightful Insightful Corporation Insightful intelligence from data S PLUS S S PLUS Graphlets Graphlets and InFact are registered trademarks
103. Intensity shifted and scaled 15 5 lt gt 3 Probe Number Figure 5 7 Example RNA Degradation plot indicating differential labeling as a function of location in the probe set The Principal Components plot is a plot of the eigen vectors resulting from a principal components analysis taking each chip as a variable with values equal to the expression intensities for all genes present on the array The plot uses different plotting symbols for each experimental condition so association between treatment conditions is indicated by relative position in the graph Replicates with wildly deviant points indicate potential problems with one or more arrays The example shown in Figure 5 8 shows very different expression for arrays taken later in the study 27 and 31 compared to arrays taken earlier in the study 1 7 and 11 199 Chapter 5 Quality Control Diagnostics and Filtering Using the GUI Diagnostics for Affymetrix Data 200 Principal Component Plot Principal Component 2 Principal Component 1 Figure 5 8 Example Principal Components Plot oxkRIOXK D0 1 Ref ELG 7 Ref 7 LC 11 Ref 11 LC 27 Ref 27 LC 31 Ref 31 LC The Quality Control Diagnostics dialog is available by clicking ArrayAnalyzer QC Diagnostics lt Affymetrix Two Channel gt from the main S PLUS menu bar The use of lt gt indicates that one of Affymetrix or Two Channel is selected but not both Figure 5 9 shows t
104. NA arrays biases in DNA spotting due to eroded print tips and spatial variability of signal in regions of an array Two channel microarrays developed within research organizations are subject to variability in all of the preparation phases e g amplification purification and concentration of DNA clones the amount of DNA spotted the binding of the DNA to the array the shape size of the spot and dye quality and labeling There are several environmental factors at play during hybridization and scanning including temperature humidity non specific binding and washing conditions The scanning process is complex with higher intensities giving higher signals but leading to saturation at the high end while lower intensities remove saturation but miss signal on the low end Imaging algorithms are likewise complex with significant segmentation issues involved in the separation of signal from background see Yang et al 2001 Commercially manufactured oligonucleotide arrays have their own variability issues including those described above Affymetrix the market leader has made a considerable effort to minimize and track variability in their arrays As in many assay formats the vendors of Why do We Normalize Data Normalization microarray technology compete based on how well their manufacturing and deployment processes control extraneous variability and provide reproducible results Let s look at the swirl data set that has already b
105. NALYZER has two dialogs for two sample problems The LPE Test dialog and the Two Sample Tests dialog See Chapter 7 Differential Expression Testing for more details To apply the LPE test procedure to this experimental open the dialog by clicking ArrayAnalyzer gt Differential Expression Analysis gt LPE Test Select Affymetrix data and the SurgeryExprSet rma data set Now choose the Age factor creating comparisons Running the LPE Test creates a volcano plot a heat map plot and variance plots as well as a summary table of most significant genes Two Way Design Differential Expression Analysis LPE Test Chromosome Fiat save HTML As El Display HTML Output SurgeryLPETest ok crea av do Figure 3 43 The SurgeryExprSet rma object used in the LPE Test dialog used in the two sample problem using the mouse surgery example data 115 Chapter 3 Examples Affymetrix Probe Level Data Two Sample Test Dialog 116 Volcano Plot qo 10 0o oo Q o0 o 2 ie qo ie pe Log10 Unadjusted p Value Mean Log2 Fold Change Figure 3 44 LPE test comparing Young and Old ignoring Time for the SurgerExprSet rma data set Similar results may be obtained from the Two Sample Tests dialog but for different analysis methods Graphical Annotation Gene List Management Two Way Design Clicking one of the hyper linked points in either the volcano plot or the heat map when you are view the
106. R in this case top 10 genes based on p value followed by fold change 346 Entrez Nucleotide Microsoft Internet Explorer File Edit View Favorites Tools Help Annotation Libraries eck gt OA Qsearch Gyravorites media 4 B S fei a Protein Genome Structure PMC Search Nucleotide gt for U10564 W27675 AB019987 U83981 Y07909 AL0 About Entrez Help FAQ Entrez Tools LinkOut equence 1 for Genes Limits Preview Index History Clipboard Details Display Summary gt show 20 _Sendto i One Items 1 10 of 10 page M1 U10564 Links Human CDK tyrosine 15 kinase WEE1Hu Wee1Hu mRNA complete cds l699 107 gb U 10564 1 HSU10564 699107 W27675 Links 36b3 Human retina cDNA randomly primed sublibrary Homo sapiens cDNA MRNA sequence gi1307623 gb W27675 1 1307623 AB019987 Links Homo sapiens mRNA for chromosome associated polypeptide C complete cds 409284 5 dbj AB019987 1 4092845 U83981 Homo sapiens apoptosis associated protein GADD34 mRNA complete cds 32586 17 gb U83981 1H5U83981 3258617 Y07909 E Figure 9 7 Entrez Nucleotide results for Melanoma data with fold change set to greater than 2 and p value set to significant adjusted p lt 0 05 for LPE test Bonferroni FWER in this case top 10 genes based on p value followed by fold change TT tenet FG 347 Chapter 9 Annotation and Gene List Management Entrez PubMed Micro
107. S Agglomerative Nesting Linkage Methods 304 Partitioning algorithms are based on specifying an initial number of groups and iteratively reallocating observations between groups until some equilibrium is attained In contrast hierarchical algorithms proceed by combining or dividing existing groups producing a hierarchical structure displaying the order in which groups are merged or divided Agglomerative methods start with each observation in a separate group and proceed until all observations are in a single group Divisive methods start with all observations in a single group and proceed until each observation is in a separate group The basic hierarchical agglomeration algorithm starts with each object in a group of its own At each iteration it merges two groups to form a new group the merger chosen is the one that leads to the smallest increase in the sum of within group sums of squares The number of iterations is equal to the number of objects minus one and at the end all the objects are together in a single group This is known variously as Ward s method the sum of squares method or the trace method The hierarchical agglomeration algorithm can be used with criteria other than the sum of squares criterion For example in the connected also known as the single linkage or nearest neighbor method the distance between two groups is defined to be the smallest distance between any two members from different groups and a
108. Swap 1 REFERENCE Factor OPTIONAL Only used if Type 2 reference Specifies the factor to use This factor must be a valid factor name from the FactorInfo block REFERENCE Factor ZebraFish 381 Appendix A Creating a Design File Table J 6 Table of DesignType valuenames Continued Value Name Rules Example REFERENCE_RefLevel OPTIONAL Only used if Type 2 reference Specifies the factor level to use This level must be a valid factor level name from the FactorInfo block for the factor specified in REFERENCE Factor REFERENCE_RefLevel Swirl REFERENCE_Reflype OPTIONAL Only used if Type 2 reference Specifies the reference type Allowable values are 0 1 2 where 0 REF_CY3 1 REF CY5 2 REF_DYESWAP REFERENCE Reflype 2 Example The following is swimming txt the design file used in the section Design File for One Way Design in Chapter 2 Examples Affymetrix MAS Data F This file can be found in the examples directory of Affymetrix S tARRAYANALYZER This experiment has two factor levels and 6 MAS CEL replicates of each level As an example let s only read four of the Oweek chips and all six of the 4week chips The two lines with blank file names are required to give six replicates for each level 3V1 3 ImportInfoLSTART FileType MAS 5 Summary Data 382 Example Design File for Two Color Experiments
109. Test Data Surgen NOVARA Array Name mgu74av2 r General Annotation M LocusLink r Open OntoE xpress I OntoExpress IV Unigene J Pubmed V GO Website Use Affymetrix IDs _ I Save Affy IDs to File Affpmetny ID File JProbeLi ttxt B F Open Affymetiy GO Browser F Open DAVID EASE Browser Cancel Apply j current Figure 3 48 General Options settings for annotation of the genes identified by the SurgeryANOVABs YoungBH ANOVA Now on the Filtering Options tab in the Contrast Filtering group select the Lhr Old Young contrast and check the Significant genes checkbox Clicking the Recalculate button in the Gene Sort Order Options group will show you how many genes are selected by the filtering Note the Limit number of genes to field which puts a cap on the number of gene ID s that will be sent to the databases for annotation extraction Figure 3 49 displays the resulting settings Two Way Design General Options Filtering Options m Contrast Filteing gt r Gene List Filtering Data on which to Filter Data on which to Filter SurgeyANOVA z GeneListLPEAr x Contrast ihr 0ldYoung x I Filter on Gene List I Genes with fold change r Cluster Filtering greater than p Data on which to Filter myCluster v J Filter on Cluster Summary r Expression Filtering Data on which to Filter Elster ian fi i r Gene Sort Order Options
110. The Genes Present plot is available for Affymetrix MAS data only if Plot you have selected the Detection variable during import See the section Keeping Extra Variables on page 30 in Chapter 2 Examples Affymetrix MAS Data for more details on keeping extra variables The Genes Present plot displays the percent of genes detected on each array and gives a sense of the quality of the hybridization across the entire array Figure 5 5 displays an example Genes Present plot Genes Present Plot Present 60 80 100 1 40 1 20 x v i X ise ise E z z a an Swim3wks3 Swim4wks1 Swim4wks2 Swim4wks3 Swim4wks 1wk1 Swim4wks 1wk2 Swim4wks 1wk3 NoSwim4wks1 NoSwim4wks2 NoSwim4wks3 NoSwim4wks 1wk1 NoSwim4wks 1wk2 NoSwim4wks 1wk3 Figure 5 5 An example Genes Present plot displays a bar plot of the percent of genes present on each array 197 Chapter 5 Quality Control Diagnostics and Filtering Intensity Boxplot Box and whisker plots display distribution summaries for a set of numbers The Intensity Boxplot displays a set of box and whisker plots one for each array in the experiment so you can compare distributions A box and whisker plot displays the median the dot or band in the center of the box the25 and 75 percentiles the lower and upper shoulders of the box and the extremes minimum and maximum unless outliers are present When outliers are present the whiskers are drawn at the nea
111. The fundamental questions of the study were 1 Which genes are active or expressing at any point in time and 2 How did gene expression change over time To answer Question 1 we need to compare the life cycle samples to the reference samples at each time point To answer Question 2 we need to compare the life cycle samples across time points Importing the malaria parasite data is similar to importing the mutant Zebrafish data that we did at the beginning of this chapter The main difference is in the design setup and specifying variables for later filtering We ll focus on these two aspects of the import process and refer you back to the beginning of the chapter for more complete information on the data import process Open the Import Data From Two Channel by clicking ArrayAnalyzer gt Import Data From Two Channel Then open the Create Modify Design dialog by clicking on the Create Modify Design button at the top of the dialog Here we specify 10 arrays two factors and factor names of LCRef with values of LC and Ref and Time with values of 1 7 11 27 and 31 The resulting dialog is displayed in Table 4 17 Creating the Design Two Way Reference Design Create Modify Design xj m Factors Number of Factors 2 Number of Arrays fio of Levels Across Chip Level values Factor Name Time 5 Yes S i 7112731 No ziuc ref B RefLC 2 m Design Type Two Sample C Loop V Dye Swap Factor RefLC
112. TipMAD f loc list maNormLoess x maA y maM z maPrintTip w NULL subset subset span span f scale list maNormMAD x maPrintTip y maM geo TRUE subset subset Normalizes to the loess curve of M vs A within each print tip group followed by within print group scale normalization using the median absolute deviation 229 Chapter 6 Pre Processing and Normalization Table 6 4 The norm parameter of maNormSca 1e results in the following normalization methods and settings being passed to maNormMain Examples With maNorm and maNormScale Normalization Between Arrays 230 Normalization Method floc Value Summary globalMAD f loc NULL Scale normalization over each f scale chip using the median absolute ties i Ne ob Eei ae deviation MAD this allows geo subset between slide scale subset normalization printTipMAD f loc NULL Within print tip group scale f scale normalization using the list maNormMAD x ee maPrintTip y median absolute deviation maM geo geo subset subset Let s look at some examples using maNorm and maNormScale scalePrintTipMAD performs both location and scale normalization gt swirl PrintTipMAD lt maNorm swirl norm scalePrintTipMAD print tip loess gt swirl ptloess lt maNorm swirl norm printTipLoess dF globalMAD gt swirl gMAD lt maNormScale swirl nor
113. UI M BH0 ake d 06 0 UI 3 MRNA sequence gi 5498520 gb A1854614 1 5498520 3 X68670 Links Mus musculus mRNA for terminal deoxynucleotidyltransferase Tdt gene 25187909 emb X63670 2 MMTDNTYL 25187909 I 4 AV309347 Links AV309347 RIKEN full length enriched 8 days embryo Mus musculus cDNA clone 5730578005 3 similar to X82786 M musculus mRNA for Ki 67 MRNA sequence gil6362382 dbj AV 309347 1 6362382 F 5 L04503 Links Mus musculus uteroglogin mRNA complete cds gi 202313 gb L04503 1 MUSUTEROG 202313 Figure 3 50 Annotation summary from Entrez Gene 122 References REFERENCES Bolstad B M Irizarry R A Astrand M and Speed T P 2002 A comparison of normalization methods for high density oligonucleotide array data based on variance and bias Bioinformatics 19 2 185 193 Irizarry R A Bolstad B M Collin F Cope L M Hobbs B and Speed T P 2002 Summaries of Affymetrix GeneChip Probe Level Data Nucleic Acids Research Vol 31 No 4 e15 Irizarry R A Hobbs B Collin F Beazer Barclay Y D Antonellis K J Scherf U Speed T P 2003 Exploration Normalization and Summaries of High Density Oligonucleotide Array Probe Level Data Biostatistics 4 249 264 123 Chapter 3 Examples Affymetrix Probe Level Data 124 EXAMPLES TWO COLOR DATA Two Color Data Analysis Workflow 126 Two Sample Design 127 Mutant Zebra Fish Data 127 Importing Data 128 Normalization
114. US environment This makes S PLUS ideal for doing microarray analysis where many traditional methods e g t test Wilcoxon test ANOVA are used but where the advantage of using cutting edge methods loess normalization invariant set normalization local pooled error test may provide a big pay off by reducing false positives and negatives The focus of this chapter is on methods primarily supported through the GUI for differential expression testing However there are many techniques not covered here that are accessible through other sections of the GUI or through the Command line See the Chapter 8 Cluster Analysis for examples on clustering and mixed effects models The focus of the S ARRAYANALYZER GUI is two sample problems For two sample problems it is quite easy to do the following 1 Read the data 2 Summarize probe level Affy data 3 Normalize 4 Test differential expression and 5 Annotate differentially expressed genes from the GUI The within gene two sample comparisons implemented through the GUI include the following methods e paired t Paired t test t Welch s t test unequal variance e tequalvar Student s t test equal variance e wilcoxon Wilcoxon signed rank sum non parametric test e t permute Welch s t test null distribution and p value estimated by permutation 261 Chapter 7 Differential Expression Testing Cautionary Note Local Pooled Error Test 262 t equalvar per
115. VA methods for analyzing one and two way designs as well The statistical methodology for two sample and multi sample one and two way designs is discussed in detail in Chapter 7 Differential Expression Testing In this section we focus on a single example a two way design to study the differences in healing processes for young and old mice after an injury In this section we step through the analysis of an experiment using mice with surgery induced injuries to discover genes active in the healing process The experimental design involves two factors 1 the age of the mice recorded as young or old and 2 the time of observation taken at 0 0 5 1 2 and 4 hours following the surgery Gene expression was measured in three replicate sets i e three mice for each time point and each age combination The main hypothesis of interest involves discovering genes showing differential expression between the two age groups and over time To demonstrate StARRAYANALYZER we focus on three time points 0 1 and 4 hours The arrays and data files are listed in Table 3 2 Table 3 2 Experimental design and file association for the mouse surgery study Age Time Rep Array label File name Old Ohr 1 Old0hr1 OldOhr1 Old Ohr 2 Old0hr2 Old0hr2 Old Ohr 3 Old0hr3 Old0hr3 Old lhr 1 Old1hr1 Old1hr1 Old lhr 2 Old1hr2 Old1hr2 Old lhr 3 Old1hr3 Old1hr3 Old 4hr 1 Old4hr1 Old4hr1 93 Chapter 3 Examples Affymetri
116. Variable Selection Green Foreground Red Foreground Rmean x Green Background bgGmean X Red Background bgRmean X Weights lt none gt hd m Extra Variables for filtering later All Variables Keep Cancel Figure 4 9 The Variable Selection amp Filtering page allows you to set the variable and row selections In this example select Gmean as the Green Foreground and Rmean as the Red Foreground to complete the required fields Optionally select bgGmean as the Green Background and bgRmean for the Red Background The Weights field is for specifying a column of spot quality weights These weights are used in subsequent computations to down weight poor quality spots during normalization See Chapter 6 Pre Processing and Normalization for more detail 137 Chapter 4 Examples Two Color Data Clicking OK Other Data Formats Single Grid Arrayers Normalization 138 Once you complete the Variable Selection amp Filtering page click OK to begin importing the files The object resulting from the import step is of class marrayRaw which is saved as an S PLUS object with the name you entered on the first page of the Import Data From Two Channel dialog SwirlMarrayRaw Some scanning equipment generates layout information as part of the data file For example some scanners generate expression intensity files with columns containing Row and Column layout information for the s
117. Way Reference Design HOME June 16 2004 2 Maximum Minimum Amplitude Score Phase CGH Avg Med OligoID Status Hour Hour log2 Pito Pi 3D7 Intensity opfg0060 UNIQUE 0 68 104 550 95 1 12 24 36 46 Chromosome 7 623292 53449_1 62846_1 218212 218211 43801 opfblob0161 494391 154104 154102 opt 30060 306631 99740 to 199740 52 147240 to 152240 Log Ratio CyS Cy3 30663_2 174731 54854_2 54854_1 opfblobi12 495361 53854_2 e8769_2 711621 149240 to 150240 T T T gi 20 30 40 opf30060 abil Time Hours ti kb OLIGo PlasmoDB ID Description MAL7P1 16 hypothetical protein Oligo Sequence BLAST PlasmoDB S GTAGTTATAATAAGAGCCGTATAAATTGTAAGAGGGTATTATTTGAATACATATGTAAACATGTATTGAA Automated Predictions chr gen_288 22 identity to 72 of hypothetical protein CO820w malaria parasite Plasmodium falciparum chr glm_37 23 identity to 95 of hypothetical protein CO760c malaria parasite Plasmodium falciparum chr phat_46 23 identity to 95 of hypothetical protein CO760c malaria parasite Plasmodium falciparum Description Figure 4 41 Annotation page from the DeRisi Lab Malaria Transcriptome Database 173 Chapter 4 Examples Two Color Data FROM THE COMMAND LINE Importing Data 174 All of the analysis done through the GUI can be done from the S PLUS command line Having access to the command line adds great flexibility to the set of feature
118. YNAME column name used on HTML gene list output The easiest way to do this is to read in the layout file in a separate S PLUs data frame You can do that either through the command line or through the GUI via the S PLUS menu bar We demonstrate the example through the GUI e Select File gt Import Data gt From File to open S PLUS Import From File dialog Two Way Reference Design Browse to the TPLayout gal file in the examples directory for S ARRAYANALYZER and select it Set the File Format field to ASCII File whitespace delim asc dat txt prn Type in the Data set name TPLayoutComplete in the To On the Options tab type in 6 for Col names row e 6 for Start row to skip the header rows and e deselect the Strings as factors checkbox Click OK to read in the GAL file The completed Data Specs and Options tabs are displayed in Figures 4 37 and 4 38 Import From File El m xl Data Specs Options Filter gt From File Name C Program Files Insightful splus62 module ArrayAnalpzersex Browse File Format fasci file whitespace delim asc dat txst pr x m To Data set TPLayoutComplete v Create new data set Add to existing data set Start cal END X Preview Rows fi D Name factor PFE0192 PFE0272 PFE0225 5 aff OK Cancel Appt K j current Help Figure 4 37 Data specs for importing gene ID data 169 Chapter 4 Example
119. a medianIQR normalization scales the summarized chip data so that they have the same inter quartile range as the maximum IQR for the set and the median of each chip s data is shifted to the maximum median of the chip set medi anIQR takes as input an expression intensity matrix each column is one chip s values and returns a matrix of the same dimensions one column for each chip in the set medianIQR can be used from the command line as follows Normalizing each treatment group separately gt DilutionEsetNormTmtl lt medianIQR norm Dilution exprSet 1 2 exprs gt DilutionEsetNormTmt2 lt medianIQR norm Dilution exprSetLl 3 4 exprs 1 An object of class exprSet contains information for experiments where the probe level data has already been summarized into one expression value for each gene Please refer to the Biobase documentation for more details splus62splus62 library Biobase Biobase pdf or the exprSet class help file Normalization Methods for Affymetrix MAS Data The data can be plotted by typing the following pre normalized data box plot log transform the data for nicer plots gt boxplot data frame log2 Dilution exprSet exprs ylim c 0 15 style bxp att post normalized data gt boxplot data frame log2 cbind DilutionEsetNormTmt1 DilutionEsetNormTmt2 style bxp att ylim c 0 15 Note When creating box plots from the normalization dialog the log intensity is use
120. a Examples of these procedures from the GUI can be found in Chapter Examples Affymetrix Probe Level Data The key task is to convert probe level data to one expression value for each gene transcript which can then be used to test for differential gene expression This is typically achieved through the following sequence of steps 1 Exploratory data analysis and diagnostics Background correction Probe specific background correction e g subtracting MM 2 3 4 Normalization 5 Summarizing the probe set values into one expression measure and in some cases a standard error for this summary As discussed in the section Workflow on page 217 normalization can be done before and or after summarizing probe level data Steps 2 5 above can be done using separate functions or together using functions such as expresso These functions as well as functions for plotting probe level data for exploratory data analysis are discussed in the next sections In S tARRAYANALYZER the expresso function provides many options to handle the tasks in steps 2 5 above Examples are given in section Summarization in StARRAYANALYZER Command Line on page 247 232 Pre Processing and Normalization for Affymetrix Probe Level Data CDF and Probe In order to compute expression summaries and or normalization of Libraries Affymetrix probe level data you will need to have the Affymetrix CDF information available DNA sequence information i
121. a ea Sample 24h A xls Gene 41126_at Accession Number EEEN aE Summary Volcano Plot Variance Plot Gene List Significant Genes Figure 7 15 Heat map plot for differentially expressed genes This graphlet may be displayed through a Web browser or an S PLUS Java graphics device Pixels colored red signify positive expression values those colored green signify negative expression values The brighter the color the larger the intensity in absolute value 289 Chapter 7 Differential Expression Testing Chromosome Plot A chromosome plot displays the human genome for Affymetrix s HG U95A chip Differential expression is marked for up regulation and down regulation for each gene represented on the chip The top 10 differentially expressed genes are highlighted with color orange to indicate their location on the chromosome Hovering the mouse over one of the colored active points displays the gene ID in the upper right hand corner of the graph as shown in Figure 7 16 Graph Window 3 File View Options 37420_i_at a o E 5 a 6 E 2 2 o A gt Page 1 Summary Volcano Plot e ene aca Plot Figure 7 16 Chromosome plot of the human genome for Affymetrix s HG U95A chip with the 10 most differentially expressed genes displayed in color Hovering the mouse over the colored spots displays the gene ID in the upper right corner 290 Differential Expression Analysis Plots Two Sample T
122. a 0 60 aag 051 0 31 023 0 18 7 GENEM OX 034 oes 024 033 om ao aa o oir 137 oog 6 aek 0 12 0 19 a9 0 46 oa 0 38 CE 024 0 64 049 0 34 9 GENES 0 35 0 09 O05 03 0 40 0 20 Oc oa 0 92 063 003 10 GEIM oad E oz oan os 053 016 oa NA ore 013 11 GeL 0 28 o2 00A Di 0s oss no Ei 0 01 ose 0 05 12 GEX 004 0 23 045 0 08 as 0 44 ax a 0 78 018 0 20 13 ENED ong Des 043 a 12 2 04 022 oad 1 60 019 ong 14 an 008 aa oa O98 146 oa oa Lie nag 0 19 15 GIE 0 27 0 23 ons 0 07 073 1 52 O81 ona 1 24 Ose 018 is nex 2a 121 021 aa ong pe as oH m 27A 034 17 ameak 3all Le 00H 3 71 D7 402 cr ooa 335 da 19 GENEL2X a 0 00 031 0 45 O78 dai ai CE 0 82 ona 19 aea oal ace 00A asl 103 0 68 noe 038 1 02 022 20 eeizx oad Ls oul 157 La 1 09 an 093 13 KEI 21 GENE127K O41 0 99 0 01 Lg 123 0 95 O28 06 1 10 ool 22 GENE ond 0 56 oo 133 14 0 68 oa ow 0 75 024 23 Cenex oo ce O10 155 027 0 00 ne Od Lad CE 24 GENELOK 021 0 71 013 195 05 O66 0 28 a7 CELI na z 027 ma i renean E7 vr a eT eT rr Sad 3an Y Aad ry z m Figure 8 12 Imported data from Figure 3a of Alizadeh et al 2000 Note that this is not the actual raw data but rather data as summarized by Cluster Eisen et al 1998 and prepared for viewing in Tree View Eisen et al 1998 We treat it as raw data to show the cluster methods in S PLUS but the resulting output should not be directly compared with fig3a of Alizadeh et al 2000 We then standardize this data frame calculate
123. a perfect match PM and mismatch MM probe The oligos are 25 mers and the MM probe uses the Watson Crick complement at the 13th position A key data operation is the summary of the 11 20 probe pair set intensities into a single value for each gene transcript that faithfully represents the expression of that gene transcript The Affymetrix MAS4 0 software did a poor job at this summarization by simply taking the average difference of the PM and MM values for each probe pair set Affymetrix MAS5 0 software does a better job of this summarization this is described below Several other summary methods for probe pair sets have emerged most notably those of Li and Wong 2001 and Irizarry et al 2003b This is an active area of research and as stated by Parmigiani et al 2003 there is mounting evidence that alternative summarization to the defaults currently implemented by Affymetrix may provide improved ability to detect biological signal The available summary methods can be obtained by typing gt express summary stat methods 1 avgdiff liwong mas medianpolish playerout 245 Chapter 6 Pre Processing and Normalization avgdiff liwong mas medianpolish playerout 246 Note that avgdiff and mas methods refer to the methods described in the Affymetrix manual versions 4 0 and 5 0 and the Affymetrix Statistical Algorithms Description Document SADD available from Affymetrix The avgdiff method i
124. accession number Other mappings can be added to this menu using the menu additem function in S PLUS Gene Name connective tissue growth factor Volcano Plot Probe ld 36638_at a oO LocusLink R T 2 k D D 3 a 1 oO D o 4 Figure 9 1 A volcano plot which is the logarithm of p value versus fold change Points above the horizontal line are hyperlinked to annotation databases 339 Chapter 9 Annotation and Gene List Management 340 The heat map plot shown in Figure 9 2 shows a two way layout of the most differentially expressed genes along the vertical axis versus the experimental conditions on the horizontal axis This graph is also hyperlinked to access annotation information using the genes LocusLink ID or accession number Sample cg2b CEL Gene 33543_s_at Exp Value 1 26 Accession Number Locust ink X fe Summary Volcano Piot Heatmap Chromosome Variance Plat Figure 9 2 A heat map plot shows differentially expressed genes as a function of experimental conditions The map is hyperlinked to annotation databases Clicking one of the hyperlinked points in the Top 15 Genes Summary plot the volcano plot or the heat map pops up a menu for selecting the database to query for annotation information Selecting an entry in this menu opens an HTML page in your default Web browser displaying a brief description of the gene with hyperlinks to more detai
125. ading Layout Information From the Command Line We ll step through an example using each of the functions to give you a flavor of their use They are listed in typical order of use Layout files describe the structure of the microarray They include information on the arrangement of the spots e g number of rows and columns on the array where each gene is located which spots are control spots etc A typical layout file is the file fish gal provided as an example for the swirl data It has 21 lines of header information and then starts a data table The file is located in the examples directory of the ArrayAnalyzer module You can find the location of the file by doing gt AApath lt file path getenv SHOME module ArrayAnalyzer examples gt AApath C PROGRAM FILES INSIGHTFUL splus62 module ArrayAnalyzer examples By scanning the file in Notepad or Wordpad you can see that the spots are arranged in 16 blocks a 4 x 4 grid of 22 rows and 24 columns Note that the ID column column 4 has indicates which spots are controls We are now ready to read in the layout file gt swirl layout lt read marrayLayout file path AApath fish gal ngr 4 nde 4 ner 22 se 24 skip 21 tl col 4 gt swirl layout Array layout Object of class marrayLayout Total number of spots 8448 Dimensions of grid matrix 4 rows by 4 cols Dimensions of spot matrices 22 rows by 24 cols Currently working with a
126. age must be completed in order to create a data object for continued analysis MIAME Completing this page is optional but highly recommended because information on the MIAME tab is used for labeling tables and graphs MAS Variables amp Filtering This page has default settings depending on the type of data files e g MAS4 or MASS you select It also allows the selection of other variables which can be used for more general filtering by using the Filtering dialog CEL Filtering This page allows a couple of options for filtering out spots not to be used in subsequent analyses when importing probe level data Options This page provides options for specifying the number of header lines to skip and the delimiter used in the data file Before we can begin to associate data files with experimental conditions we need to set up the experimental conditions in S ARRAYANALYZER The easiest way to do this is through the Create Modify Design dialog Open the Create Modify Design dialog by clicking on the Create Modify Design button on the File Selection page of the Import Data From Affymetrix dialog create y modify Design Number of Arrays p Number of Factors fi of Levels Level Values 2 Al A2 Cancel Hep Figure 3 26 The default Create Modify Design dialog 97 Chapter 3 Examples Affymetrix Probe Level Data Step 2 Associating Files With Design Points 98 The Create Modify Design dialog allows you to spec
127. ain menu by selecting Help gt Available Help gt arrayanalyzer The HTML Help system includes a table of contents organized by library an index and a Search button You can also get help on any S PLUS function from either the command line or from Help gt Available Help gt Language Reference If you need help on the S PLUs GUI click the Help button at the bottom of any dialog or navigate to Help gt Available Help gt S PLUS Help In addition to the online help you can access a pdf of the User s Guide by going to Help gt Online Manuals gt ArrayAnalyzer User s Guide The S ARRAYANALYZER User s Guide is particularly helpful for those new to S PLUS and microarray analysis You can also access versions of the Bioconductor library pdfs in S tARRAYANALYZER The individual library pdfs are located at the top level of each library for example the Biobase library pdf is available at splus62 library Biobase Biobase pdf Just double click the file to launch the pdf Note these pdfs are current only up to this release For updated information please visit the Bioconductor Web site North Central and South America Contact Technical Support at Insightful Corporation Telephone 206 283 8802 or 1 800 569 0123 ext 235 Monday Friday 6 00 a m PST 9 00 a m EST to 5 00 p m PST 8 00 p m EST Fax 206 283 8691 E mail support insightful com Web http www insightful com support Chapter 1 Introduction To Microar
128. ality Control Diagnostics and Filtering For two channel data with multiple print tip groups the image plot also shows the print tip grouping Figure 5 2 displays an image plot for a two channel custom cDNA microarray with 16 print tip groups Red Foreground for Ref LC 1 1 63000 56000 49000 42000 35000 28000 21000 14000 7000 Figure 5 2 Image plot example for two channel data with 16 print tip groups M vs A Plot The MvA or M vs A plot is a scatter plot of M versus A For Affymetrix data M and A are defined as M log Ek Ek Pe log EXE k T where is the expression intensity for replication 4 147 i j 1 of replications of the kth treatment For two channel data M and A are defined as 194 Quality Control Diagnostics M log 4 and A logo RG 9G where R and G represent the expression intensities of the Red and Green channels The MvA plot is a plot of the fold change versus average logy intensity for each pair of replications within each experimental condition Figure 5 3 displays a typical MvAplot Swim3wks MvA Plots Swim3wks1 o o 3 2 7 6 8 10 12 14 16 1 38 Swim3wks2 2 4 a 8 10 12 14 16 1 42 0 578 Swim3wks3 Figure 5 3 An MvA plot for Affy MAS data The points of MvA plots should hover around zero because no differential expression is expected for experimental replicates The line going through the cent
129. an other data blocks valuenames There are no valuenames Instead each non empty line is interpreted as a factor Currently only two factors are allowed in S ARRAYANALYZER so only the first two non empty lines in this block are used Each line that defines a factor has a specific format factor code factor name levels where factor code is A B etc factor name is the name you wish to use for the factor and levels is a space or comma delimited list of level names to use for each factor Each element in the line is separated by a vertical bar character also known as the pipe symbol As an example consider the line 377 Appendix A Creating a Design File Design valuenames 378 A CondTime Swim3wks Swim4wks 1wk NoSwim4wks NoSwim4wks 1wk The factor code is A The factor name is CondTime The levels for this factor are Swim3wks Swim4wks 1wk NoSwim4wks and NoSwim4wks 1wk The lines in a design block are different than other data blocks There is only one valuename allowable in this block listed in a table below The other non empty lines in this block are interpreted as files to import and factors and levels to use for each of these files Each filename line is comma delimited and has a specific format file name factor A level factor B level where file is either the full pathname to the file to import or is just the filename without path If that is the case all the ot
130. and the LPE Test dialog produces a Variance plot displaying a graph of the baseline variance estimates as a function of the average expression intensity for each experimental condition The ANOVA Test produces Principal Components Plots Volcano Plots and Summary Table for each contrast selected as well as a heat map Each of these types of plots is discussed in the following sections Volcano Plot Differential Expression Analysis Plots A volcano plot displays the logarithm of adjusted p value versus average fold change The vertical lines indicate average fold change values of plus or minus two and the horizontal line indicates a significant adjusted p value Points located in the lower outer sextants using this graphing orientation are those with large absolute fold change and small significant p value Each of those points is active so you can click an individual point to access annotation from LocusLink or UniGene databases Volcano Plots may now be plotted in positive see below or negative classic volcano style orientation Gene Name 38428_at T 2 d L a 3 T 5 g o gt S a Mean Log2 Fald Change a Summary Volcano Plot Heatmap Variance Plat J Gene List Significant Genes Figure 7 13 A volcano plot shows the logarithm of the adjusted p value vs average fold change This is displayed using an S PLUS Java graphics device It may also be displayed in the browser for all the fo
131. appings mouseCHRLOC mouseLLMappings ratCHRLOC ratLLMappings The libraries in bold above are included in the S ARRAYANALYZER installation CD All of the Annotation libraries are available from the S ARRAYANALYZER data libraries Web site referenced on the previous page The Annotation libraries are automatically attached when you attach S ARRAYANALYZER To manually attach the Affymetrix chip specific libraries and use the libraries in scripting at the command line enter gt library lt chipname gt AnnoData Annotation Libraries Attaching an annotation library makes all the S PLUS annotation objects for that library available in the current S PLUS session For example the S PLUS annotation objects available in the hgu95av2 library are shown in Table 9 1 Table 9 1 S PLUS Affymetrix chip specific library objects S PLUs Annotation object Description hgu95av2ACCNUM Maps probe ids to GenBank accession number or user specified ids hgu95av2CHRLENGTHS A vector containing the lengths in base pairs of chromosomes hgu95av2CHRLOC Maps probe ids to chromosomal locations hgu95av2CHR Maps probe ids to chromosome numbers hgu95av2ENZYME2PROBE Maps enzyme EC numbers to probe ids hgu95av2ENZYME Maps probe ids to enzyme EC numbers hgu95av2GENENAME Maps probe ids to gene names hgu95av2GO2ALLPROBES An annotation data file for GO2ALLPROBES hgu95av2G02PROBE Maps GO ids to probe ids with
132. are those with large absolute fold change and small significant p value Each of those points is active so you can click an individual point to access annotation information from LocusLink or GenBank Gene Name 19 412 Volcano Plot T 2 g a D o 2 D 3 o m 2 D fa a Mean Log2 Fold Change Figure 4 15 A volcano plot which is the logarithm of the adjusted p value versus fold change 145 Chapter 4 Examples Two Color Data Heat Map A heat map plot shows a two way layout of the most differentially expressed genes along the vertical axis versus the experimental conditions on the horizontal axis This graph is also hyper linked to the annotation information if it exists Sample Swirl WildType 2 Cy3 Gene 27 N19 Exp Value 14 77 Figure 4 16 A heat map plot shows differentially expressed genes as a function of experimental conditions 146 Two Way Reference Design TWO WAY REFERENCE DESIGN Malaria Parasite Data In this section we examine two color microarray data from another developmental biology experiment The data come from the study of one stage of the life cycle of Plasmodium falciparum one of four parasitic protozoa which cause human malaria which annually affects 200 300 million people worldwide By understanding the life cycle through gene expression insight into the biochemical function and regulation of thes
133. ased probability of obtaining a test statistic at least as large as ti from simulated distributions of the test statistics generated by the decreasing sets fti tnp ftia tN fti tay The minP and maxT procedures are only available for the permute versions of the test procedures t permute t equalvar permute wilcoxon permute Furthermore the permute versions of the test statistics only have access to these two procedures for p value adjustment The other adjustment procedures are implemented for all the non permutation testing procedures described in the section Statistical Tests The results of the procedures are summarized using adjusted p values which reflect for each gene the overall experiment Type I error rate when genes with a smaller p value are declared differentially expressed Adjusted p values may be obtained either from the nominal distribution of the test statistics or by permutation The false discovery rate FDR is defined as the proportion of genes expected to be identified by chance relative to the total number of genes with significant tests of difference That is FDR FP IS FP in Table 7 1 Controlling the FDR has the advantage of maintaining a small number of false positives amongst only those tests which are significant BH The Benjamini and Hochberg procedure computes Pqy ming n min N k pqy 1 Any Pa lt ais significant with an overall FDR for the experiment not greater than a This proc
134. ation swirl 1 spot Figure 6 2 MvA plot for chip 81 of swirl dataset before normalization with loess curves for each print tip group S ARRAYANALYZER provides a wide variety of normalization methods depending on the form of the data prior to normalization For two channel chips the scanning software typically accounts for background noise and adjusts for control information For Affymetrix probe level data CEL files the analyst needs to account for background noise and may make use of controls often the mismatch probes MM are used for this to correct for random non specific binding The analyst must also summarize the probe level data into a single value per gene transcript Affymetrix summarized data e g chp files output from MAS 4 5 software has already been Normalization background adjusted and perhaps mildly normalized and the probe level data have been summarized into a single intensity value per gene transcript Because of the inherent differences in two channel data versus Affymetrix data the specifics of the normalization methods differ between data types However normalization in general can be thought of as either normalization to a point location normalization or scaling of the variability of the data scale normalization Having a visual representation of the data is very useful in the normalization process S ARRAYANALYZER includes a variety of diagnostic plots and the sections that follow di
135. ation and summarization steps can be done in one step using the expresso functions Details on expresso can be found in the help file invoked by entering gt expresso Summarization Examples In these examples we correct the data for background signals and noise normalize the data at the probe level and summarize the probe level data into one value per gene transcript We do this all using the expresso function Affymetrix MAS 5 0 To obtain a summary similar to MAS 5 0 use gt eset lt expresso affybatch example normalize FALSE bgcorrect method mas pmcorrect method mas 247 Chapter 6 Pre Processing and Normalization 248 summary method mas gt eset lt affy scalevalue exprSet eset Notice that in this case we normalize after we obtain summarized expression measures The function affy scalevalue exprSet performs a normalization similar to that described in the MAS 5 0 manual see the section on affy scalevalue exprSet on page 253 This is a simple global scaling in which the user enters a target value TGT value The average signal across all probes on each chip is calculated for each chip and a scale factor SF is determined for each chip such that chip mean SF TGT Thus the signals on each chip are scaled by a single number for each chip a crude form of normalization Li and Wong 2001 MBEI To obtain a probe level normalized summary similar to Li and Wong s MBEI one can use This is computat
136. aw created during the import step 139 Chapter 4 Examples Two Color Data 140 Save As In the Save As field the name SwirlMarrayRaw norm will be generated by default You can edit the name in this field if you wish Our example uses the default object name for the normalized expression data Normalization Now set the other options on the right side of the Normalization dialog The normalization methods are listed in the Normalization drop down list median loess twoD e printTipLoess e scalePrintTipMAD e global MAD e printTipMAD In this example select scalePrintTipMAD as the normalization method Select the Box Plot check box the MvA check box and the Before amp After radio button for pre and post normalization plots Click OK or Apply to produce the normalized data and create the pre and post normalization plots shown in Figures 4 12 and 4 13 Two Sample Design After scalePrintTipMAD Normalization 0 5 10 15 Figure 4 12 MvA plot of normalized data 141 Chapter 4 Examples Two Color Data Before scalePrintTipMAD Normalization After scalePrintTipMAD Normalization ot _ J Figure 4 13 Boxplots of before and after scalePrintTipMAD normalized swirl data Differential Open the Differential Expression Analysis Two Sample Tests Expression dialog by clicking ArrayAnalyzer gt Differential Expression Analysis gt Two Sample
137. ayInfo are info id gene ID numbers and associated labels labels the column number in fname which contains the names that the user would like to use to label spots or arrays e g for default titles in maPlot skip the number of lines of the data file to skip before beginning to read data We can now read in the data The four swirl files are located in the examples directory of the ArrayAnalyzer module like the other file read in the previous section The function read marrayRaw takes the following arguments fnames a vector of character strings containing the file names of each spot quantification data file These typically end in spot for the software Spot or gpr for the software GenePix path acharacter string representing the data directory By default this is set to the current working directory In the case where fnames contains the full path name path should be set to NULL name Gf character string for the column header for green foreground intensities name Gb character string for the column header for green background intensities name Rf character string for the column header for red foreground intensities name Rb character string for the column header for red background intensities name W character string for the column header for spot quality weights layout object of class marrayLayout containing microarray layout parameters From the Command Line gnames object of class marrayInfo containing probe
138. bases like LocusLink or UniGene This workflow shows the steps incorporated into the workflow of S ARRAYANALYZER when doing differential expression analysis Microarray technology is complex and experiments using microarrays are resource intensive As such there is an urgent need for rigorous statistical design and analysis of microarray experiments Chapter 1 Introduction To Microarray Data 10 Statistical issues in microarray experiments include Experimental design Pre processing e g normalization Differential expression testing Clustering and prediction Annotation All of these issues may be addressed with the use of modern statistical methods Care is required however to perform the analysis correctly and detailed collaborations between biologists and statisticians are a sound recipe for successful use of microarrays S ARRAYANALYZER provides off the shelf functionality for microarray data analysis as well as a toolkit and development environment for custom microarray analysis solutions Key packages from the Bioconductor project are included in S ARRAYANALYZER http www bioconductor org Reports from microarray analysis such as summary gene lists and sae volcano plots are presented using S PLUS Graphlets which facilitate interactive annotation of result summaries and allow you to share results via the Web Microarray Data MICROARRAY DATA DNA microarrays are now widely used as a key experimental p
139. be pair consists of a spot for the probe called a perfect match PM and a spot for a slight alteration of the probe called a mismatch MM Non specific binding may be accounted for by adjusting PM intensities to account for MM expression intensities Biotin labeled Total RNA cDNA cRNA Reverse in Vitro Za N AAAA Transcription S Transcription Ann AAAA Ne gt a B NANA AAAA Q cy gt i ee GeneChip E lon xpression N NS B M maea B EN Bloat G S lt p Q8 eos asc Figure 1 4 Affymetrix s GeneChip is a one color oligonucleotide array Mass produced reliable standardized microarrays like the GeneChip have help fuel the bioinformatics revolution i Affymetrix has revolutionized bioinformatics with its GeneChip technology To analyze Affymetrix expression data all the expression values for each probe pair are first summarized by a single value There are numerous ways to do this and Affymetrix provides methods for such probe level summarization in their MAS4 and MASS software S ARRAYANALYZER provides other methods for probe level analysis which are discussed in depth in Chapter 3 Our example in Chapter 2 uses Affymetrix MAS 5 summary data 13 Chapter 1 Introduction To Microarray Data Two Color Arrays 14 cDNA or two color microarrays are designed to compare two different samples the experimental conditions on each slide Each sample is treated with a different color before it is added
140. bles for filtering later All Variables Keep Detection Detection p value Probe Set Name Signal Stat Pairs lt lt Stat Pairs Used Cancel k afe entries Help Figure 2 9 The Variable Selection amp Filtering page of the Import Data From Affymetrix dialog Note the Apply Log2 Transformation check box which by default takes logo of the expression intensities before saving them in the resulting object The actual computation is log E if E gt 1 and 0 if E lt 1 29 Chapter 2 Examples Affymetrix MAS Data Fields you may be interested in changing are in the in the Remove Probe Set group There are four options 1 Remove probes that weren t detected in any sample 2 Remove a probe if detection p value is less then a specified value in all samples Default p value is 0 01 3 Remove a probe if the number of pairs used in computing the summarized expression intensities is less than a specified value in all samples The default value is seven The maximum is the total number of probe pairs in a set typically 11 16 or 20 4 Remove control probes that have a specified prefix Default prefix is AFFX Keeping Extra The last group on the MAS Variables amp Filtering tab allows you to Variables select Extra Variables for filtering later Use these extra variables for probe set selection or removal before doing differential expression analysis Detection variable is required for a
141. button on the File Selection page of the Import Data From Affymetrix dialog 71 Chapter 3 Examples Affymetrix Probe Level Data Number of Arrays fa Number of Factors fi Figure 3 3 The default Create Modify Design dialog The Create Modify Design dialog allows you to specify 1 The number of arrays to be read 2 The number of factor in the experiment Currently one or two are allowed 3 The name number of levels and level values for each factor To modify the default factor Name of Levels and Level Values type them into the appropriate field For the two sample Melanoma analysis there are 4 arrays and one factor so the first two fields do not change from default Type in the Name of the factor as Time and the Level Values as 2hr and 24hr The resulting dialog is displayed in Figure 3 4 Create Modify Design Number of Arrays fa 4 Number of Factors fi a of Levels Level Values 2 2hr 24hr Figure 3 4 The Create Modify Design dialog with two sample experiment setup 72 Step 2 Associating Files With Design Points Two Sample Design Once the design is complete click OK to copy it into the File Selection tab of the Import Data From Affymetrix dialog Notice that the values for the factor levels have been written into the Factor column to facilitate associating files with design points However the level values can be reset when the experiment is unbalanced or if you prefer an order differ
142. c list maNormLoess x maA y maM z NULL span 5 maNormLoess x maA y maM z maPrintTip a 1loc maCompNormA Simple wrapper functions to marrayNormMain are provided by maNorm and maNormScale These wrappers send default accessor methods and settings to marrayNormMain as outlined in Table 6 3 and Table 6 4 Two channel normalization from the S ARRAYANALYZER normalization dialog uses these functions and associated method names Normalization Methods for Two Channel Data Table 6 3 The norm parameter of maNorm results in the following normalization methods and settings being passed to maNormMain maSpotRow y maSpotCol z maM g maPrintTip w NULL subset subset span span Normalization Method floc Value Summary median f loc list maNormMed x Median normalization by chip NULL y maM subset subset loess f loc list maNormLoess x Normalization to loess curve of maA y maM z chip s M vs A NULL w NULL subset subset span span twoD f loc list maNorm2D x 2D spatial location normalization Normalizes to the smoothed intensity surface loess surface by print tip group at each x y coordinate printTipLoess f loc list maNormLoess x maA y maM z maPrintTip w NULL subset subset span span Normalizes to the loess curve of M vs A within each print tip group on each chip in the object scalePrint
143. cal_process physiological process h 632 cellular_component cell h 679 A sie SS ea Figure 9 15 DAVID EASE site Fisher exact test for gene enrichment for list of Melanoma data Affy IDs uploaded by S ARRAYANALYZER 355 Chapter 9 Annotation and Gene List Management Annotation Using LocusLink IDs 356 Figure 9 16 shows the General Options page of the Annotation dialog with options chosen from Use LocusLink IDs group This writes out a file of LocusLink IDs LocusLinkList txt by default corresponding to the genes selected according to the Annotation dialog options This LocusLinkList txt file can be uploaded One option that is made easily accessible with S ArrayAnalyzer is the Stanford Source site This site may be opened from the Annotation dialog so that all the user needs to do is browse to the file LocusLinkList txt and upload it Figure 9 17 shows a screen shot from the Stanford Source site as launched and uploaded from S ARRAYANALYZER Information from Stanford Source is used in constructing the annotation data in Bioconductor ox General Options Filtering Options m Data r Use LocusLink IDs Show Data of Type I Save LocusLink IDs to File DitexerTest 1 LocusLink Fie LocusLinkList ot Data eal PEBon he Neen Array Name hou95av2 p rM General Annotation Open Ont I LocusLink peen onea I OntoExpress I Unigene i Binomial 7 J Pubmed jonfer
144. cation of your files Two Way Design Import Data From Affymetrix E xj File Selection MIAME MAS Variables amp Filtering CEL Filtering Options m Step 1 Specify Design Read Existing Design Create Modify Design D Devel splus_modules 4rrayAnalyzerexamples SurgeryDesign txtllll Save Design File r Step 2 Associate Files with Design Points Factor Tim File Name Factor Age D Devel splus_modules Arr 4Old0hr1 CEL OldOhr1 Old D Devel splus_modules Arr Old0hr2 cEL OldOhr2 Old oh D Devel splus_modules Arr Old0hr3 CEL OldOhr3 Old foh D Devellsplus_modules arr Old1hr1 CEL Oldthr1 old lih D Devel splus_modules Arr Oldthr2 CEL Oldihr2 Old fih D Devel splus_modules arr Oldihr3 CEL Oldihr3 Old aliw Type filename or right click to brovical Aldabe Old xl 4hr File Type Probe Level CEL Type mgu74av2 7 CDF El r Step 3 Save Output ee EENE Save Data Set As SurgeryAffyB atch Display Report Cancel x afi entries Help Figure 3 28 Browsing for data files You can find the surgery example data by navigating to your splus62 module ArrayAnalyzer examples directory and selecting the Old0hr1 txt file Repeat for the other 17txt files entering one file per field Alternatively you can read the design file named SurgeryDesign txt in the examples directory to load the design and create file assoc
145. cel Apa current Help Figure 7 7 The Local Pooled Error Test or LPE test dialog 276 Options Variance Estimation GUI for LPE Testing Once a data object is selected the chip name is filled in the Chip Name field For custom 2 color cDNA or non Affymetrix oligonucleotide chips the chip name may be lt undetermined gt The Options group contains the procedures for controlling the FWER and FDR as shown in the drop down list in Figure 7 8 The procedures correspond to those described in section Controlling The False Positive Rate and section FDR Procedures Both FWER and FDR procedures are included in the drop down list Select one and specify the family wise error rate for either an FWER or FDR procedure in the FWER editable field Differential Expression Analysis LPE Test g 5 xj gt Data Variance Estimation Show Data of Type Smoother D F fo a Affymetrix Z Number of Bins ho y Data DAYSdetense Trim Ef Eactor A E r Output Options Compare Level 1 Joyo z IV Volcano Plot Compare Level 2 foy1 zl Y Axis Orientation Array Name Jmgu74av2 negative zi Options Eold Change Line po g FWER FDR IV Heat Map Adjustment J Chromosome Plot Alt Hypothesi IV Variance Plots f IV Top15 Genes r Output IV Display Output in S PLUS Tl Save Output as HTML Save HTML A myLPE Test html F Display HTML Gutput Save As mpLPETest Cancel Apply CECEN Help Figu
146. cessor methods are then used to extract the desired information from the data object for use in the normalization computation or plotting function Useful methods are maM and maA for obtaining the intensity log ratios and average log intensities respectively and the maPrintTip method which computes the print tip grid coordinates for the spots maPrintTip and maPlate are used to stratify the data by print tip groups or by chips Please refer to the marrayClasses documentation splus62 library marrayClasses marrayClasses pdf or the marrayRaw and marrayNorm help files for additional options Normalization usually begins with exploratory data analysis and diagnostic plots Two channel data typically includes two treatment conditions on one chip Dudoit et al 2002 and Yang et al 2001 suggest that the most useful way to view such data in order to identify spot artifacts and for normalization purposes is via an M vs A plot of 221 Chapter 6 Pre Processing and Normalization 222 the intensity log ratio M log 2 vs the mean log intensity 2 A log 4RG This amounts to a 45 degree counter clockwise rotation of the log G log R coordinate system followed by a scaling of the coordinates This plot highlights the difference between the red and green channels as a function of average intensity across the two channels Figure 6 2 on page 212 shows an example of an MvVA plot for two channel data Box plots are also available from t
147. cgLPEBon pea Aray Name hou35av2 r General Annotation M LocusLink LocusLinkList txt J Open Stanford Source m Open OntoExpress I OntoExpress IV Unigene Distribution Binomial T I Pubmed Correction Bonferroni id MV GO Website Username Use Affymetrix Ds Password I Save Affy IDs to File Affymetiy ID File ProbeList txt Browse F Gpen Affpmetiy GO Browser J Open DAVID EASE Browser Cancel Apply Iff ement Figure 9 4 The General Options page of the Annotation dialog Options are highlighted from the General Annotation group Annotation Libraries For part 2 we filter based on the p value of the LPE differential expression test and the fold change Figure 9 5 below shows the Filtering Options page of the Annotation dialog with the Contrast Filtering group activated and options fold change set to greater than 2 and p value set to significant adjusted p lt 0 05 in this case General Options Filtering Options r Contrast Filtering Gene List Filtering Data on which to Filter Data on which to Filter eoLPe Bon E Contrast cg0 cg24 x Filter on Gene List IV Genes with fold change m Cluster Filtering greater than et Data on which to Filter IV Significant genes J Filter on Cluster Summary m Expression Filtering Data on which to Filter Elton ideas z cg xprSet ma x m Gene Sort Order Optio
148. ch other within treatment condition or versus the median expression computed for all replicates within each treatment condition Options for the MvA plot are a a scatter plot produced by drawing a random sample of genes and b a hexbin plot of all genes where the hexagonal points are colored to give a sense of the density of points at each location 3 Genes Present Plot A simple barplot of percent of genes present on each array Available only for MAS 4 5 data when the Detection column has been selected for filtering during the data import stage 4 Intensity Boxplot Boxplots of expression intensities for each array 5 RNA Degrad Plot Plot of RNA degradation Only available for probe level CEL data 6 Principal Components Plot Plot of the first two principal components using treatment combinations i e expression intensities for the entire array as variables in the principal components analysis For a thorough analysis we would typically create most if not all of the diagnostic plots and check them for problem arrays and spots For this example we will do a few CEL specific plots as examples and refer you to the other chapters Chapter 2 Examples Affymetrix MAS Data and Chapter 4 Examples Two Color Data for more examples of other plots The following set of graphs are examples of the QC Diagnostic plots for the SurgeryAffyBatch dataset Figure 3 31 displays the image plot for one of the arrays in the study Each
149. clustl cex 1 rotate me T 1ty 1 The cluster analysis can also be done from the S PLUS menu system using Statistics gt Cluster Analysis Agglomerative Hierarchical although this does not produce the heat map with overlaid dendrograms as visual output like the above code snippet Note that the above code snippet can be easily saved as an S PLUS function for repeated use as follows gt cluster heat lt function cluster data sample colors rep 1 dim cluster data 1 stand norm lt function x x mean x na rm T sqrt var x na method available cmat lt apply cluster data 1 stand norm cluster rows distl lt dist t as matrix cmat hclustl lt hclust dist distl method average cluster cols dist2 lt dist as matrix cmat Examples from the Command Line hclust2 lt hclust dist dist2 method average plot heat map and dendrograms par mai c 0 0 0 0 omi c 0 2 7 1 4 1 1 image cmat hclust2 order hclustl order axes F bty n par new T omi c 6 55 2 75 0 1 15 plclust2 fn hclust2 cex 1 rotate me F 1ty 1 colors sample colors hclust2 order par new T omi c 0 02 0 95 1 42 7 75 plclust2 fn aliz hclustl cex 0 1 rotate me T 1ty 1 This function could then be called to analyze the Alizadeh data as follows Of course the function could do with some error checking if it were planned to be used by others gt cluster heat mat3a sample colors c rep 6 16 rep 1 2 rep 6 6
150. corresponds to a different vertical line in the plot Figure 2 14 displays MvA plots for four replicate pairs The plotting format uses hexbinning which shows the density of the data better than a simple scatter plot Note the legend to the right describing how the colored hexagons relate to the density of points at each location Figure 2 15 displays the Genes Present Plot a barplot representing the percent of genes detected on each array Figure 2 16 displays boxplots of expression intensities for all arrays in the study 33 Chapter 2 Examples Affymetrix MAS Data Figure 2 17 displays a plot of the first two principal components for all the arrays in the study Each point in the plot corresponds to a different array Different symbols correspond to different experimental conditions NoSwim4w1 NoSwim4w2 06 08 1 0 1 2 1 4 o6 08 1 0 1 2 1 4 2000 4000 6000 NoSwim4w3 o 2000 4000 6000 NoSwim4w1w1 06 08 1 0 1 2 1 4 0 6 0 8 1 0 1 2 14 0 2000 4000 6000 o 2000 4000 6000 Figure 2 13 Image Plot diagnostics for four of the arrays in the MouseSwimExprSet data set Swim3wks MvA Plots Swim3w1 Swim3w1 9 f 7 2 Swim3w2 Swim3w2 9 f 4 2 Swim3w3 Swim3w3 16 0 12 0 5 3 8 9 12 0 16 0 Hundreds 13579 Tens 13579 Ones 13579 Figure 2 14 A set of MvA hexbin plots for
151. cted to be small Figure 3 14 displays the MvA plot for the 24 hour samples The interpretation is the same as that for Figure 3 13 Figure 3 15 displays boxplots of logged expression summaries for each sample chip Visual inspection shows the distributions are well aligned at their centers and quartiles Although normalization may be repeated sequentially to summarized expression intensities there is little need to apply more normalization to cgExprSet rma Note The values displayed in the MvA plots in Figures 3 13 and 3 14 depend on the values used for Random seed You may see slightly different plots as a result 81 Chapter 3 Examples Affymetrix Probe Level Data Log of After applying normalization and summarization procedures to the Expression raw expression intensities a logy transformation is applied Intensities Consequently the returned summarized object contains expression intensities on a log scale The log transformation is computed as logo E ifE gt 1 0 if E less than or equal to 1 2hr MvA Plots 02 04 06 cg2a 0 2 0 176 cg2b Figure 3 13 MvA plot for the two replicate samples measured at two hours The value in the lower left panel of the plot is the interquartile range of M 82 Two Sample Design 24hr MvA Plots 0 5 0 5 0 0 cg24a 1 0 0 231 cg24b Figure 3 14 MvA plot for the two replicate samples measured at 24
152. d Workflow Ideas in Normalization for the set of chips provided i e all those read into one object during the data import phase Thus if the whole experiment is supplied normalization will be done across treatment groups From the command line the user has more control For example the user may choose to normalize within experimental conditions and merge the resulting normalized data When chips are normalized to an average reference it is assumed that there is a common underlying intensity distribution on each chip For this reason pairwise normalization where one chip is a target chip may be preferable when there are just 2 chips But pairwise normalization when there are more than two chips has been shown to give variable results depending on which chip is chosen to be the reference chip Bolstad et al 2002 Normalization of microarray data is currently an active research topic We leave it to the researcher to decide the best approach for their data Examples shown in this chapter are for demonstration purposes only Data corrections and normalizations can be done in series The suggested work flow is as follows For Affymetrix probe level data e Background correcting Probe level summary Summarizing the 11 20 probe pair sets into a single value for each transcript e Location scale normalization Affymetrix summary data e g CHP file data from Affymetrix 4 5 and two channel summary data GPR file
153. d tO was chosen as baseline level the contrasts would be t1 t0 t2 t0 t3 t0 i e the differences between baseline and each subsequent time The contrast analysis identifies genes that are differentially expressed at each particular time point compared to baseline These differentially expressed genes for each contrast provide visibility into the biological sequence of events across the time course The Sequential setting compares levels of the chosen factor in sequential ordered pairs For example if a factor time had levels t0 tl t2 t3 the contrasts would be t1 t0 t2 t1 t3 t2 These contrasts also provide systematic view of differential expression over the time course or factor levels The choice of Sequential vs Baseline may depend on whether a true baseline condition was run and or the interpretive goals of the experiment The Linear Quadratic setting fits linear and quadratic contrasts across the ordered levels of the chosen factor This analysis identifies genes that exhibit a a significant linear trend in either the positive up or negative down direction and b a significant quadratic trend in either the positive down in the middle or negative up in the middle direction Contrasts may be applied either Within or Across another factor For example if there are two sexes M F tested at each of the 4 time points mentioned above i e t0 t1 t2 t3 testing the time contrasts vs baseline Within the Sex factor woul
154. d this forms the RMA probe level analysis method of Irizarry et al 2002 2003b The playerout method computes a weighted mean of the PM values based on the method described by Lazaridris et al 2002 Our tests show this method gives unstable results This method is not recommended Pre Processing and Normalization for Affymetrix Probe Level Data Summarization in Summarization in S ARRAYANALYZER can be done through the S ARRAYANALYZER Affymetrix Expression Summary dialog as demonstrated in Command Line Chapter 3 Examples Affymetrix Probe Level Data From the command line summarization can be done as a separate step using the wrapper function computeExprSet on an AffyBatch object or through the functions generateExprVal method method These functions require a matrix of probe intensities with rows representing probes and columns representing samples Details of using these functions can be found in the help files for example type gt generateExprVal method avgdiff computeExprSet can be used to summarize probe data gt affybatch example gt ids lt c A28102_at AB000114 at gt eset lt computeExprSet affybatch example pmcorrect method pmonly summary method avgdiff ids ids warnings FALSE generateExprVal method method requires an d intensity matrix gt probes lt pm SpikeIn gt medianpolish lt generateExprVal method medianpolish probes The correcting normaliz
155. d design points MIAME is an acronym for Minimal Information About a Microarray Experiment This information can be entered on the second page of the Import Data From Two Channel dialog This information is not required but it is stored on the resulting object to identify the source of the data Once you ve entered MIAME information for any experiment the first three fields are saved and are filled automatically the next time you open this dialog The MIAME tab of the Import Data From Two Channel dialog is displayed in Figure 4 8 135 Chapter 4 Examples Two Color Data Import Data From Two Channel Bob Bryant bob igri com Zebrafish swirl mutant Zebrafish embryos from two genetic straings a swirl mutant and a normal wild type D Microarrays ArrayAnalpzersswirlNotes txt Figure 4 8 Entering chip information in the MIAME page 136 Two Sample Design Variable Selection The third page in the Import cDNA Data dialog is for variable and amp Filtering Page row selection There are two required fields on this page Green Foreground and Red Foreground where you must select the columns containing green and red foreground intensities Select the columns from the drop down list of variable names which is populated using the column headers in the imported data files Import Data From Two Channel a x File Selection MIAME Variable Selection amp Filtering Options m
156. d false discovery rate FDR to control the overall Type I error false positive rate based on adjusting individual test p values to account for multiple tests In our surgery study there are 12 488 genes so the Type I error is substantial without adjusting the p values 111 Chapter 3 Examples Affymetrix Probe Level Data There are many options for adjusting the p values to achieve the FWER or FDR you want Here we set the adjustment procedure to Benjamini Hochberg BH which has the nice property of limiting the Type I error i e false positive rate to a small percentage of the significant genes There are four options in the Output Options group 1 2 3 4 Volcano plot Heat map Parallel Coords Top 15 Genes Chromosome Plot Not Available for All Chips Note that chromosome plots are not available for arrays other than hgu95a Graphical Output Two examples are shown the Volcano plot and the Parallel coords plots Figure 3 41 displays the Volcano plot for the Young versus Old contrast at 1 hour There are 46 significant genes in the plot resulting from the BH correction Even with an 8 fold increase in significant genes the BH correction maintains a low false positive rate of 5 112 Two Way Design amongst the significant genes This translates to on average about 2 3 genes not really differentially expressed amongst those genes tagged as significant by the correction procedure Gene Name ubiqu
157. d for differential expression Please refer to Chapter 3 Examples Affymetrix Probe Level Data for details on this process There are three parts to this analysis 1 Choose the summary data analysis results for the melanoma experiment and the desired annotation sources 2 Filter the genes to identify interesting genes for annotation 3 Annotate these genes which uses functions from the Annotate library For part 1 we use data that have been read in at the GUI and analyzed for differential expression The object cgLPEBon is created by reading in the cel files summarizing with RMA and performing 341 Chapter 9 Annotation and Gene List Management 342 differential expression testing using LPEtest with Bonferroni FWER control Please refer to Chapter 3 Examples Affymetrix Probe Level Data to see how this object is created Figure 9 4 shows the General Options page of the Annotation dialog in S ARRAYANALYZER with the differential expression test result object cgLPEBon selected and the metadata sources LocusLink Unigene PubMed and GO Web site selected Note that data types can be annotated through the Annotation dialog via Affymetrix summary data MAS5 summarized CEL two channel data differential expression test results Dif fExprTest and the Gene Lists Tixi General Options Filtering Options Data Use LocusLink IDs Show Data of Type I Save LocusLink IDs to File DiffE xprT est E LocusLink File Data
158. d on the y axis affy scalevalue exprSet vsn To continue working with an exprSet object we can create a new exprSet object which has the normalized intensity information d normalize the data without subsetting gt DilutionEsetNorm matrix lt medianIQR norm Dilution exprSet exprs create new exprSet object with normalized intensities gt DilutionEsetNorm lt Dilution exprSet gt DilutionEsetNorm exprs lt DilutionEsetNorm matrix affy scalevalue exprSet shifts the mean intensity value of the chips to the same specified point The default reference value is 500 The function accepts exprSet objects and returns an exprSet object Similar to medianIQR affy scalevalue exprSet can be used to normalize summarized data as follows Normalizing with affy scalevalue exprSet gt DilutionEsetScaleTmtl lt affy scalevalue exprSet Dilution exprSet 1 2 sc 100 gt DilutionEsetScale lt affy scalevalue exprSet Dilution exprSet sc 100 vsn variance stabilizing normalization is available from the GUI and the command line This function can operate on exprSet objects and returns an exprSet object 253 Chapter 6 Pre Processing and Normalization Normalizing summarized data with vsn gt vsn Dilution exprSet gt vsn Dilution exprSet 1 2 Diagnostic MvA and box plots for Affymetrix summarized data are available Plots for through the Normalization dialog of S ARRAYANALYZER GUI menu An example of an MvA pl
159. d result in the comparisons t1 t0 t2 t0 t3 t0 for each of the 2 sexes Testing the time contrasts vs baseline Across the Sex factor would result in just one set of comparisons t1 t0 t2 t0 t3 t0 averaged across the 2 sexes Similarly if Sex was chosen as the contrast factor testing the Sex contrast Within the Time factor would result in the comparisons M F at t0 M F at tl M F at t2 M F at t3 and testing the Sex contrast Across the Time factor would result in a single sex comparison M F The Options group contains the procedures for controlling the FWER and FDR as shown in the drop down list in Figure 9 11 The procedures correspond to those described in section Controlling The Output Options Output GUI for ANOVA Testing False Positive Rate and section FDR Procedures Both FWER and FDR procedures are included in the drop down list Select one and specify the family wise error rate for either an FWER or FDR procedure in the FWER editable field m Options FWER FDR 0 001 IV Protected Adjustment BH Output Options Bonferroni Holm IV Volcano Plot Hochberg SidakSS Y Axis Orientation SidakSD BH Figure 7 11 Setting the p value adjustment procedure for controlling the FWER The Output Options group is a list of check boxes and a drop down box for selecting which graphs you want as output The options are r Output Options IV Volcano Plot Y Axis Orientation Fold Change Line ko a
160. data from GenePix are typically normalized to location and possibly scale before differential expression analysis 217 Chapter 6 Pre Processing and Normalization DIAGNOSTIC PLOTS Box Plots MvA Scatter Plots 218 Diagnostic plots of intensity data can help identify printing hybridization scanning artifacts and other sources of unwanted variability which can removed before analysis of differential gene expression The S ARRAYANALYZER GUI provides a variety of diagnostic plots to help identify such unwanted variability and guide subsequent adjustments and modeling procedures Please refer to Chapter 5 Quality Control Diagnostics and Filtering for additional details on these diagnostic plots Additional plots such as Histograms are also available from the S ARRAYANALYZER command line Box plots show side by side graphical summaries of intensity information from each array The summary consists of the median and the upper and lower quartiles 75th and 25th percentiles respectively of the data The central box in the plot represents the inter quartile range IQR which is defined as the difference between the 75th percentile and the 25th percentile The median is represented by a line or a dot in the middle of the box By default the upper and lower whiskers on the box plots are placed at the most extreme observation not exceeding plus and minus 1 5 times the IQR from the quartiles Data outside the whiskers are plotted separatel
161. dely applied to the analysis of gene expression data Eisen et al 1998 Scherf et al 2000 In particular the method of visualizing gene expression data based on cluster order or cluster image map analysis using Aierarchical clustering has been found to be an efficient approach for summarizing thousands of gene expression values and assisting in the identification of interesting gene expression patterns Partitioning clustering methods such as K means are used to identify candidate subgroups in experiments involving multiple samples and or experimental conditions Both hierarchical and partitioning clustering have been used for example in the identification of novel sub types of cancers Introduction Additional gene information is also extremely useful for discovering meaningful clustering patterns Prior to clustering it is also recommended to identify genes based on their statistical significance in differential expressions and to confirm consistent expression patterns within replicates e g Ross et al 2000 S ARRAYANALYZER provides both types of algorithms via it s GUI Also S PLUS has additional algorithms belonging to both categories You can access them either through the S PLUS GUI see Statistics gt Cluster Analysis from the main S PLUS menu bar or through the command line interface A good general reference to clustering methods is Kaufman and Rousseeuw 1990 303 Chapter 8 Cluster Analysis HIERARCHICAL METHOD
162. dule gt swirl samples lt read marrayInfo file path AApath SwirlSample txt The resulting object show the information stored in a rectangular array gt swirl samples From the Command Line Object of class marraylInfo maLabels of slide Names experiment Cy3 1 81 81 swirl 1 spot swirl 2 82 82 swirl 2 spot wild type 3 93 93 swirl 3 spot swirl 4 94 94 swirl 4 spot wild type experiment Cy5 date comments 1 wild type 2001 9 20 NA 2 swirl 2001 9 20 NA 3 wild type 2001 11 8 NA 4 swirl 2001 11 8 NA Number of labels 4 Dimensions of maInfo matrix 4 rows by 6 columns Notes C PROGRAM FILES INSIGHTFUL splus62 module ArrayAnalyzer examples SwirlSample txt Reading Gene ID s We are now ready to read the gene ID s gt swirl gnames lt read marrayInfo file path AApath Tish gal 1nfo td 4 5 labels 5 skip 21 gt swirl gnames Object of class marraylInfo maLabels ID Name genol control genol geno2 control geno2 geno3 control geno3 3XSSC control 3XSSC SXxSSC control 3xS5C ESTI control EST1 genol control genol geno2 control geno2 geno3 control geno3 3XSSC control 3XSSC O oOo O So tr DBU fh a Number of labels 8448 177 Chapter 4 Examples Two Color Data Reading Expression Intensity Data 178 Dimensions of maInfo matrix 8448 rows by 2 columns Notes C PROGRAM FILES INSIGHTFUL splus62 module ArrayAnalyzer examples fish gal The additional arguments to read marr
163. e DAT files contain the raw images as processed by the scanner The CEL files contain expression measures for each individual probe on the chip analysis of these probe level data is described in Chapter 3 Examples Affymetrix Probe Level Data and in this chapter in section Pre Processing and Normalization for Affymetrix Probe Level Data The CHP files contain summaries of the individual probe level data for each gene transcript Analysis of these summarized data is described in this section These data have been background adjusted and summarized into a single expression value per gene transcript using the Affymetrix MAS software Affymetrix version 5 0 software has adjusted the probe level intensity values as follows e Global background signal and noise have been subtracted and thresholded as described in the Affymetrix Statistical Algorithms Description Document SADD available from Affymetrix This method is also described in section mas on page 238 The 11 20 mismatch MM and perfect match PM values have been summarized using a Tukey biweight procedure as described in the Affymetrix document Statistical Algorithms Description Document SADD and section mas on page 246 Ifrequested by the user the software scales the signal using a trimmed mean 2 of the data at either end is trimmed away before the mean is computed The output intensity for MAS 5 0 data is termed Signal Affymetrix version 4 0 software adj
164. e both channels are clustered simultaneously As an example use one of the datasets created in Chapter 4 Examples Two Color Data Proceed as follows 1 Select data of type Two Channel from the Show Data of Type drop down list and select one of the two channel datasets from the Data drop down list 2 Select one of M or red and green channels from the Response drop down list in the Response Variable group Clustering Significant Genes From An ANOVA Examples from the GUI Go to the Filtering Options tab and select the Genes with maximum fold change greater than checkbox and set the associated field to 8 We want to limit the genes to only those with extreme fold change Select a clustering method and click Apply or OK to run the analysis The resulting plots are similar to those for Affymetrix Summary data The other data type option from the Show Data of Type drop down list is ANOVA DiffExprTest This option allows you to cluster on only those genes that are significant in an analysis As an example use the ANOVA object created during the analysis of the Swimming Mouse data in Chapter 2 Examples Affymetrix MAS Data We named the ANOVA object MouseANOVANoSwim4wksBH Proceed as follows l Select ANOVA DiffExprTest from Show Data of Type drop down list Select MouseANOVANoSwim4wksBH from the Data drop down list On the Filtering Options tab in the Contrast Filtering group from the Contrast drop down list select
165. e 2 29 Venn Diagram resulting from merging two gene lists Also you can compare the results between the ANOVA and LPE Test results for the Swim3wks NoSwim4wks contrast by comparing their gene lists Select the two differential test objects in the Gene List Management dialog and select the above contrast The results are displayed in Figure 2 30 49 Chapter 2 Examples Affymetrix MAS Data Annotation 50 MouseSwimANOVABs4wkBH MouseSwimANOVABs4wkBH Swim3wks NoSwim4wks Swim4wks NoSwim4wks 40 MouseSwimLPETestBsNoSwim4wks Swim3wks NoSwim4wks Figure 2 30 Comparison of gene lists generated by ANOVA and LPE Test procedures In addition to the annotation described in section Graphical Annotation on page 46 there is a dialog with more general annotation capabilities To open the Annotation dialog click ArrayAnalyzer gt Annotation The resulting dialog has two tabs 1 General Options 2 Filtering Options There are a number of options available on the Annotation dialog but for this first example we ll focus on generating the annotation information for the significant genes in MouseSwimANOVABs4wkBH For a more thorough description of the Annotation dialog see Chapter 9 Annotation and Gene List Management To annotate the significant genes identified in MouseSwimANOVABs4wkBH select DiffExprTest from the Show Data of Type list and select MouseSwimANOVABs4wkBH from the Data list Note the public annotation databases a
166. e Level 1 and 1 as Compare Level 2 Name the Save As object TPLPET31m1 The resulting dialog is displayed in Figure 4 33 Repeat the analyses for time points 31 and 7 and 31 and 27 naming the result objects TPLPET31m7 and TPLPET31m27 respectively For time points with a large spread we would expect many differentially expressed genes given the results of the clustering For closer time points we would expect fewer We will examine these results with the Gene List Management dialog Two Way Reference Design Differential Expression Analysis LPE Test Im r r r moo a hromosome Flat Alt Hypothesis L Output TPLPE31mi Figure 4 33 LPE Test dialog set up for comparing time points 37 and 1 165 Chapter 4 Examples Two Color Data Gene List Management 166 Volcano Plot Negative LogiD0 Adjusted p Value Mean Log2 Fold Change Figure 4 34 Volcano plot for comparing time points 31 and 1 The Gene List Management dialog allows you to merge and compare gene lists from testing different contrasts in an analysis Open the dialog by clicking Array Analyzer gt Gene List Management Now in each of the three data groups select DiffExprTest for the Data Type and each of the contrast objects you created for comparing time points The resulting dialog is displayed in Figure 4 35 For more information on the Gene List Management dialog see Chapter 9 Annotation and Gene List Management Clic
167. e P keys define the Insightful splus62 module Array Analyzer examples swirl Lspot swirl 1 Swirl WildType swirl 2 spot swirl 2 Swirl WildType swirl 3 spot swirl 3 WildType Swirl swirl 4 spot swirl 4 WildType Swirl Design END 373 Appendix A Creating a Design File Table J 3 Table of special keys Continued Line Description Rules Example DesignType START DesignType key OPTIONAL DesignType START Identifies a block Only used by Type 0 used for Two Two Channel DesignType END Channel imports a 4 e TWO_SAMPLE_ only which defines TERAS DyeSwap 1 more design The lines in LOOP Factor parameters as between these ZebraFish displayed in the keys define Create Modify additional LOOP_DyeSwap 1 Design dialog for design REFERENCE Factor Two Channel parameters A ZebraFish designs See the DesignType REFERENCE RefLevel valuenames Swirl table below for REFERENCE Reflype more 2 information DesignType END 374 Format Specification Importinfo valuenames Table J 4 Table of ImportInfo valuenames Value Name Rules Example FileType OPTIONAL FileType MAS 5 Summary Data Used only by Affymetrix import If not specified file type is auto determined by first file read Specifies the file type for the files in the Design data block see Design Lines table below The file type must be one of the following recognized types MAS 5 Summary Data MAS
168. e chip set to be normalized to create average reference values for the chip set We can think of normalizing groups of data to a reference point that is bringing the median of a data group to a fixed reference point through a shift of the values This reference point can be as simple as a given constant intensity value constant or median normalization or as complicated as fitting a locally weighted least squares regression loess normalization through the data We call this type of normalization location normalization Location normalization is necessary to correct for spatial variation e g such as when the slide is slightly tilted during hybridization which results in more mRNA available for binding at different locations on the slide One of the most common location normalization methods for microarray data is to normalize the data to a loess curve fit through the MvA plot The loess method fits a curve to the data using robust locally weighted regression as discussed in Cleveland 1979 Yang et al 2001 2002 and the Guide to Statistics Local regression is a Normalizing to Many Points Normalization Using Quantiles Ideas in Normalization smoothing method for summarizing multivariate data using general curves and surfaces The smoothing is achieved by fitting a linear or quadratic function of the predictor variables locally to the response data M in this case The loess procedure fits polynomials over contiguous subsets i
169. e exactly as needed However the level values can be reset when the experiment is unbalanced or if you prefer an order different from the default To associate data files with the design points in the uppermost File field type in the path to the file or right click and select Browse for file Repeat this step until you ve selected all your files being careful to match them with the design points Two Sample Design You can find the swirl example data by navigating to your splus62 module ArrayAnalyzer examples directory selecting SPOT spot as the Files of type and selecting swirl 1 spot Repeat for the other three spot files entering one file per cell as shown in Figure 4 4 Selecting Multiple Files You can select multiple files when you re in the browser The order you select them will be the order they fill in the file selection grid Import Data From Two Channel xi File Selection MIAME Variable Selection amp Filtering Options r Step 1 Specify Design Read Existing Design Create Modify Design Save Design File r Step 2 Associate Files with Design Points Factor Zebrafish File Cy3 C Program Files Insightf swirl 1 spot Swirl WildType 1 Swirl v WildType _ C Program Files Insightf swirl 3 spot Swirl WildType 2 Swirl xl WildType xl C Program Files Insightf swirl 2 spot WildType Swirl 3 WildType xl Swirl xl Type filename or right c
170. e genes will provide the foundation for future drug and vaccine development efforts toward eradication of malaria Bozdech et al 2003 The study was designed to track gene expression closely over a 48 hour period of time of a single stage of the life cycle The data consists of replicate dye swapped pairs of arrays taken at life cycle time points 1 7 11 27 and 31 hours Total RNA at each time point of the life cycle LC was compared to an arbitrary reference pool of RNA from all time points See Bozdech et al 2003 for more detail We acknowledge Dr Joseph DeRisi s generous offer to allow Insightful to include these data with S ARRAYANALYZER The complete data are also freely available for download from http dx doi org 10 1371 journal pbio 0000005 sd001 The experimental design is a two way reference design with observations at the design points displayed in Table 4 2 Table 4 2 The malaria parasite experimental design C532 C635 Time Array File Name Ref LC 1 1 TP_0la gpr LC Ref 1 2 TP_01b gpr Ref LC 7 3 TP_07a gpr LC Ref 7 4 TP_07b gpr Ref LC 11 5 TP_lla gpr 147 Chapter 4 Examples Two Color Data Importing Data 148 Table 4 2 The malaria parasite experimental design C532 C635 Time Array File Name LC Ref 11 6 TP_11b gpr Ref LC 27 7 TP_27a gpr LC Ref 27 8 TP_27b gpr Ref LC 31 9 TP_3la gpr LC Ref 31 10 TP_31b gpr
171. e inherent properties or changes of the tightness of transcriptional control in different conditions If necessary these need to be addressed by other methods For additional information on this method please refer to Huber 2002 and 2003 VSN works on the raw data from two channel affymetrix summary WARNING Cautions in Normalizing 216 Because VSN works on raw data be sure that when using VSN on Affymetrix summary data the log transform of the data is not done during import and affymetrix probe level sources For Affymetrix summary data VSN returns data on a generalized log scale For Affymetrix probe level data VSN works like other probe level normalization methods and returns data on an exponentiated generalized log scale this allows for subsequent logging that occurs in probe level summary methods For additional information on this method please refer to Huber 2002 and 2003 From the command line VSN normalization can easily be done on marrayRaw Class objects using the vsn function or by coercing an marrayNorm object to an exprSet object and passing that to vsn See the vsn help file Some care is required when normalizing across treatment groups to not wash out signal particularly for aggressive normalization approaches such as the quantile method This is not much of an issue with mild normalization approaches such as lining up medians and IQR s In the S ARRAYANALYZER GUI normalization is performe
172. ed graphs The name of the output files is generated by the name supplied in the Save Summary As field and then adding myANOVA html The default Graphlet name is myANOVA html Location of Output Files The location of these output files is determined by your S PLUS working directory To determine your working directory just type gt getenv S_ WORK D arrayanalyzer users lenk test The location of dumped files in general is the default S PLUS working directory If you specify no project folder when you start S PLUS your cmd directory is the default working directory gt getenv S_ WORK GUI for ANOVA Testing D Program Files Insightful splus62 cmd You should see two HTML files in your working directory when S PLUs has finished generating the output one for the summary table an 285 Chapter 7 Differential Expression Testing DIFFERENTIAL EXPRESSION ANALYSIS PLOTS Common Plots 286 The differential expression summary plots are designed to give you easy access to annotation data in public databases Two of the plots the volcano plot and the heat map have embedded hyperlinks so you can click on a point and bring up annotation from NCBI databases There are three plots common to both testing dialogs Volcano plot e Heat map e Chromosome plot Each of the dialogs optionally produces one additional plot The Two Sample Tests dialog produces a Q Q Normal Probability plot of the test statistics
173. ed structure There are at least a couple of ways that the value of such analyses can be assessed Hierarchical Methods The cophenetic distance between two observations 7 and jis defined to be the intergroup distance at which observations are first put into the same cluster The extent to which cophenetic distances reflect the true distances relates to the usefulness of the dendrogram as a tool for visualization This agreement can be assessed by the cophenetic correlation coefficient or the correlation between the true distances and the cophenetic distances The silhouette distance measures how well individual samples are classified into a discrete set of classes This is a particularly relevant measure in assessing the value of a partitioning cluster analysis but can be applied to a hierarchical analysis by cutting the tree at some point and classifying samples into the groups defined by the cut This is described further below 307 Chapter 8 Cluster Analysis PARTITIONING METHODS K Means Clustering Partitioning Around Medoids Silhouette Plots 308 The partitioning methods available through the S ARRAYANALYZER GUI are K Means and Partitioning Around Medoids One of the most well known partitioning methods is k means In the k means algorithm the observations are classified as belonging to one of k groups Group membership is determined by calculating the centroid for each group the multidimensional version of the mean
174. edure Scand J Statist Vol 6 65 70 Hsu J C 1996 Multiple Comparisons Theory and Methods London Chapman and Hall Lee J K and O Connell M 2003 An S PLUS library for the analysis of differential expression In The Analysis of Gene Expression Data Methods and Software Edited by G Parmigiani E S Garrett R A Irizarry and S L Zeger Springer New York Moore D S and McCabe G P 1999 Introduction to the Practice of Statistics 3rd ed New York W H Freeman and Company Snedecor G W and Cochran W G 1980 Statistical Methods 7th ed Ames Iowa Iowa State University Press Storey J D 2002 A direct approach to false discovery rates Journal of the Royal Statistical Society Series B 64 479 498 References Westfall P H and Young S S Resampling based multiple testing Examples and methods for p value adjustment John Wiley amp Sons 1993 299 Chapter 7 Differential Expression Testing 300 CLUSTER ANALYSIS Introduction Hierarchical Methods Agglomerative Nesting Model Based Hierarchical Clustering Understanding The Results of Hierarchical Clustering Partitioning Methods K Means Clustering Partitioning Around Medoids Examples from the GUI Basic Dialog Description Clustering Affymetrix Summary Data Clustering Two Channel Expression Data Clustering Significant Genes From An ANOVA Examples from the Command Line References 302 304 304 305 306 308 308
175. edure provides a good balance between discovery of significant genes and protection against false positives since occurrence of the latter is held to a small proportion of the significant gene list BY The Benjamini and Yekutieli procedure computes pa min n min Nsum 1 j k P 1 Controlling Type I Error Rates Any Pa lt a is significant with an overall FDR for the experiment not greater than a 269 Chapter 7 Differential Expression Testing GUI FOR TWO SAMPLE TESTING Two Sample Dialog Input Data 270 The dialog for Two Sample Tests is displayed in Figure 7 1 Open the dialog from the main S PLUS menu by clicking Array Analyzer gt Differential Expression Analysis Two Sample Tests The dialog is arranged in four main groups e Data e Options e Graph Options e Output The Data group allows you to select the expression object for testing You start by selecting the data type in Show Data of Type as one of Affymetrix or cDNA and then selecting a data object an expression object created by importing expression summarization for Affy CEL and normalization from the Data drop down list box Differential Expression Analysis Two Sample Tests ioj xj Data r Output Options Show Data of Type IV Volcano Plot Affymetrix zi Y Axis Orientation Eactor fA x Eold Change Line po o fa Compare Level 1 Dao z I Heat Map Compare Level 2 Day 1 IT Chromosome Plo mou dav Array Name
176. een discussed in section Mutant Zebra Fish Data on page 127 Figure 6 1 and Figure 6 2 shows a box plot for each chip in the experiment and an M vs A plot of one of the chips Figure 6 1 shows that there are significant differences in the median log intensity differences If the probes are placed randomly on the slide and the experimental conditions are well controlled we would expect the medians of each print tip group to be similar However the experimental conditions are not perfectly controlled as shown by the negative values for all of the print tip groups in Figure 6 1 The log ratio of intensities is a measure of the difference between the red and green fluorescence The fact that this quantity is always negative suggests an imbalance in intensities of the two dyes Cy3 and Cy5 qj i 1 gt 3 W swirl 1 spot swirl 2 spot swirl 3 spot swirl 4 spot Figure 6 1 Box plot for swirl experiment before normalization 211 Chapter 6 Pre Processing and Normalization Normalizing in ArrayAnalyzer 212 Figure 6 2 is an M vs A plot for one chip in the swirl set The loess curves for each print tip are superimposed on the scatter plot The plot shows a non linear dependence of the log ratio of red to green intensity M on the average log intensity A In section Normalization Methods for Two Channel Data on page 220 we examine a number of normalization methods for this dataset to correct this systematic vari
177. ef description of each Euclidean If ng is the number of arrays in which no missing values occur for the given genes then the distance returned is sqrt ncol x ng times the Euclidean distance between the two vectors of length ng shortened to exclude missing values Maximum No special handling for missing values is necessary Manhattan The rule is similar to the Euclidean metric except that the coefficient is ncol x ng Binary The rule excludes columns in which either row has a missing value If all values for a particular distance are excluded the distance is labeled as missing Model Based Another approach to hierarchical clustering is model based Hierarchical clustering This method is based on the assumption that the data are Clust ering generated by a mixture of underlying normal or Gaussian 305 Chapter 8 Cluster Analysis Understanding The Results of Hierarchical Clustering 306 probability distributions It provides insight into the number of clusters a quantity that is derived from a model selection process in its probability framework S PLUS provides a model based clustering algorithm for use at the command line or through the GUI However updated methods are now available for free download Hierarchical methods have been widely used for the cluster analysis of microarray data Yeung et al 2001 discuss the benefits of model based clustering for microarray analysis Results from the hierarchical m
178. el of the test procedure is A and the number of genes being tested is N A procedure is said to control the family wise error rate FWER if it adjusts the significance level so that the overall error rate is at most A Without adjusting the significance level there may be as many as N false positives For arrays with many genes the number of false positives without correcting for multiple tests can be quite large Consequently a number of procedures have been implemented in S ARRAYANALYZER for controlling FWER and FDR The results of the procedures are summarized using adjusted p values which reflect for each gene the overall Type I error rate when genes with a smaller p value are declared differentially expressed Define Papi Lagu where N number of comparisons genes tested as the ordered p values from smallest to largest resulting from the statistical tests t ordered test statistics from largest to smallest i 1 N Ho Null hypothesis no differential expression Then the p value adjustment procedures are defined below Bonferroni The Bonferroni correction is p min py N 1 for each i All genes with adjusted p values p less than Lia are significant with an overall FWER of at most Ua Note that the raw p values have simply been multiplied by the number of comparisons Hochberg The Hochberg 1988 step down correction is Pq min _ n min N k 1 pqq 1 The procedure sequentially comp
179. ells at the end of the lines The first few lines will include the delimiters at the end of the lines but then Excel will stop outputting the delimiters in subsequent lines For either Affymetrix Summary MAS or two channel Excel data the files must meet these requirements 1 The data for each array is in a separate file For Two Channel data data for both channels must be in the same file 2 The first line of the file must a column name row with unique values 3 There cannot be any tail section 387 Appendix B Importing data Layout Information for Two Channel Data 388 To read two channel data information about the layout must be read first This includes information about the spot grid rows and columns gene names and control spots Sometimes this information is included in the data file and other times it is available in a separate file e g a GAL file for GenePix data The GAL file versions are currently all 1 0 even if they are created by different GenePix versions In S ARRAYANALYZER 2 0 the layout information must be in an ASCII file it cannot be read from an Excel file If you have your layout information in an Excel file you can export it as a delimited ASCII file from Excel INDEX A annotate library 341 B BH 144 Bonferroni 84 88 144 145 188 Box Plot 140 C cDNA normalization median 140 183 printTipMAD 140 scalePrintTipMAD 142 Chromosome Plot 90 Chromosome plot 84 F
180. ensities are plotted on the log scale Once we ve imported the data files we need to convert the raw probe level expression intensities to expression summaries before testing for differential expression This is usually done in a series of steps including some combination of the following Background correction e Normalization e Probe specific background correction e g subtracting mismatch MM expression intensities e Summarizing the probe set values into one expression measure and sometimes a standard error for this measure 79 Chapter 3 Examples Affymetrix Probe Level Data RMA Summary 80 An assortment of procedures are available for completing these steps You can find much more detail in Chapter 6 Pre Processing and Normalization In addition to normalization in the context of summarizing raw intensities you can also normalize without the summarization step In this section we focus on one sequence of steps referred to as robust multichip analysis or RMA for short This procedure completes the following steps 1 Probe specific correction of the perfect match PM probes using a model based on observed intensity being the sum of signal and background noise Irizarry et al 2002 Irizary et al 2003 2 Normalization of corrected PM probes using quantile normalization Bolstad et al 2002 3 Calculation of expression measures using median polish This sequence of steps is available by simp
181. ensity columns for the given set of chips and pass each intensity column or matrix depending on whether the normalization is over a chip or across a group of chips to the functions normalize method The normalize method functions can be called directly by the user by passing in the appropriate intensity vector or matrix You can obtain a list of normalization methods for an object by typing gt normalize methods Dilution 241 Chapter 6 Pre Processing and Normalization This list of methods for AffyBatch objects can also be obtained by typing gt normalize AffyBatch methods 1 constant 4 loess contrasts qspline 7 quantiles robust vsn invariantset quantiles The normalization methods available in S ARRAYANALYZER for Affymetrix CEL files AffyBatch objects are shown in Table 6 6 Table 6 6 Normalization methods available through the normalize function T subset size 5000 verbose T family symmetric Normalization Methods Default Function Values Description Location Normalization Methods constant ref 1 FUN mean na rm T Normalizes one chip toahavea given mean or median value or normalizes a set of chips if the object is an AffyBatch object to have the same mean or median as a given reference chip contrasts span 2 3 choose subset Performs a modified loess normalization using contrasts to create a linear combination of all pai
182. ent from the default The next step is to associate files with each design point To do so right click in one of the file fields and browse to the location of your files When you ve entered the file names the dialog should look like Figure 3 5 Import Data From Affymetrix a x File Selection MIAME MAS Variables amp Filtering CEL Filtering Options r Step 1 Specify Design Read Existing Design Create Modify Design Save Design File rm Step 2 Associate Files with Design Points Factor Tim C Program Files InsightFul spl cg24a CEL cg24a C Program Files InsightFul spl cg24b CEL cg24b C Program Files InsightFul splu cg2a CEL cg2a C Program Files InsightFul splus jp a E File Type Probe Level CEL Array Type HG_U954v2 z CDF O l m Step 3 Save Output Save Data Set As myAffyBatch IV Display Report ma Figure 3 5 The Import Data From Affymetrix dialog after the data files have been entered T see the full path of a file hover your mouse over the filename You can find the Melanoma example data by navigating to your splus62 module ArrayAnalyzer examples directory and selecting the cg2a CEL file Press the CTRL key and select in order the cg2b CEL cg24a CEL and cg24b CEL Alternatively you can read the design file named cgDesignFile txt in the examples directory to load the design and create file associations 73 Chapte
183. ent over the chip Background correction aims to quantify and subtract this background signal from the expression intensities S ARRAYANALYZER provides four methods for correcting Affymetrix probe level chips for background signal and inconsistencies Three of these methods are available through the function bg correct The fourth method GC RMA is a separate function gcrma 237 Chapter 6 Pre Processing and Normalization Methods via bg correct The background correction methods available with bg correct can be obtained by typing gt bgcorrect methods 1 mas none rma rma2 rma and rma2 The rma background adjustments assume the PMs are a convolution of the normal and exponential distributions According to Bolstad 2002a we can write this as O S N where N is the background and S is the signal It is assumed that S is distributed exp a and N is distributed N 0 The background corrected PM values returned for each chip in the object are then E s O o This expectation is equal to i f ERCI Larz 2 where a s u 0o a b o and p and are the normal density and cumulative distribution function respectively oa Caution 238 The rma methods adjust the PM values but leaves the MM values intact This is problematic if a PM correction is done after the background correction using MM values which have not been background corrected rma and rma2 differ only in how the
184. er experiment with the same design by modifying the file locations and names and factor levels as needed In fact if you have many arrays in your experiment you can create a file with all the design content and read it with the Read Design button which will fill the file name fields and their associated factor levels MIAME is an acronym for Minimal Information About a Microarray Experiment and this information can be entered on the second page of the Import Data from Affymetrix dialog This information is not required but it is used in table output and graphics and thus it is to your advantage to complete the information in this page Once you ve entered MIAME information for any experiment the first three fields are saved and are filled automatically the next time you open this dialog The third page of the Import Data From Affymetrix dialog is for variable and row selection for Affymetrix MAS 4 5 data It is not used for CEL data Two Way Design CEL Filtering The fourth page of the Import Data From Affymetrix dialog is for spot filtering probe level data The options are 1 A checkbox to convert spots labeled as MASKS to missing so they aren t used in subsequent analyses 2 A checkbox to convert spots marked as OUTLIERS to missing so they aren t used in subsequent analyses Other Options The last page on the Import Data From Affymetrix dialog is the Options tab The tab provides two options used during data import
185. er of the points estimates the relationship between fold change and average log intensity between a pair of replicates Deviations from zero suggest systematic bias indicating the need for normalization or possible removal of an array from subsequent analysis When MvA plots are plotted as scatter plots sampling is done to keep the plotting overhead to a reasonable level The default sample percentage is 20 For an array with 10 000 genes 2 000 points are 195 Chapter 5 Quality Control Diagnostics and Filtering plotted by default The percentage of points plotted referred to as the sample size in the GUI is editable by the end user Increasing the sample size to 100 is possible but at the expense of increased overhead with little to no increase in useful information An alternative to the MvA scatter plot is the MvA hexbin plot which shows plotted points as colored hexagonal bins The hexagonal bins represent the number of points at each bin location by the intensity and color of shading of the bin A legend displays a set of colored bins and how they correspond to the density of points in each bin Swim3wks MVA Plots oy 4 2 2 Q Q Swim3wks13 2 Swim3wks13 2 Hundreds 60005 13579 Tens 6 4 6 4 13579 Ones 6 13579 9 2 4 Swim3wks 14 2 g 5 10 Swim3wks15 Figure 5 4 An MvA hexbin plot 196 Quality Control Diagnostics Genes Present
186. ernative to the Null Hypothesis for the statistical tests that there is no differential expression Significant differential expression for any given gene means that Compare Level 1 Swirl is greater than less than or not equal to Compare Level 2 WildType Leave the default setting of Not equal for the swirl example There are many options for adjusting the p values to achieve the FWER We describe them in more detail in the Differential Expression Testing chapter Typically we start with the default Bonferroni procedure but instead we select the BH Benjamini Hochberg procedure which is less conservative than Bonferroni and controls the false discovery rate rather than the family wise error rate See Chapter 7 Differential Expression Testing for more detail There is little expression activity in this study so set the FWER FDR to 0 15 to be less conservative There are three options in the Output Options group to display any of the following a volcano plot a heat map or the top 15 gene list Clicking OK or Apply produces the output plots which are discussed in the following section Two Sample Design Volcano Plot A volcano plot displays the logarithm of p value versus fold change as shown in Figure 4 15 The vertical lines indicate fold change values of plus or minus two and the horizontal line indicates a significant LPE Test p value after doing the Bonferroni correction Points located in the lower outer sextants
187. erpreted as settings for the imported data See ImportInfo valuenames table below for more information about these settings Line Description Rules Example 3V2 01 Version key REQUIRED 3V2 01 eae the version This line must z be the first non S ARRAYANALYZER emptylin of used when the file pty the file was written sImportInfo START Import key OPTIONAL sImportInfo START eee All lines FileType MAS 5 sImportInfo END information about Pebvecn ese Dumunary gala two keys are ChipName mgu74av2 CDFPath SaveAs MouseSwimExprSet PrintOutput 1 sImportInfo END 372 Table J 3 Table of special keys Continued Format Specification the factors and levels to use files to import and the factors and levels to use for each file See the Design valuenames table below for more information Line Description Rules Example 3FactorInfo START Factor key Identifies OPTIONAL 3FactorInfo START a pee vee The lines in A ZebraFish No Swirl contains information between these WildType 3FactorInfo END about the factors and kevys define th y levels used eys cenne te FactorInfo END factors and levels to be used See the FactorInfo value names table below for more information Design START Design key OPTIONAL Design START rire Bes apices The lines in relativepath which defines the Design END files to import and between these C Program Files tae o
188. es are believed to be relevant to tumor invasion and metastasis 69 Chapter 3 Examples Affymetrix Probe Level Data To answer this question we need to compare gene expression at the two time points a two sample problem Impo rtin g Start by reading in all the arrays To import Affymetrix data from the Data main S PLUS menu select Array Analyzer gt Import Data gt From Affymetrix ArrayAnalyzer Import Dats rong Quality Control Diagnostics gt Fromtwo Channel Filtering Affymetrix Expression Summary Normalization Differential Expression Analysis gt Cluster Analysis Annotation Gene List Management Figure 3 1 Menu selection to import Affymetrix data Import Figure 3 25 shows the Import Data From Affymetrix dialog with Affymetrix Data the File Selection page displayed The primary tasks of the import Dialog process are 1 create an experimental design 2 associate data files with the experimental conditions and 3 save the resulting S PLUS data object for later use Secondary tasks include inputting meta MIAME data describing the experiment and specifying options for handling data marked as MASKS or OUTLIERS by the Affymetrix software The Import Data from Affymetrix dialog has five pages 1 File Selection This page must be completed in order to create a data object for continued analysis 2 MIAME Completing this page is optional but highly recommended because information
189. es maNormCal1 marrayInfo marrayInfo character eall Given that we have a dye swapped cDNA experiment we ll use a traditional paired t test to test for differences in expression We first have to create a couple of objects that are arguments to the aa teststat function First extract M the log intensity ratios JHHF Extract M s log2 R G for each chip M lt maM swirl norm Now compute fold change in preparation to doing a paired t test See code for aa teststat for details on how this is done in general JHH Compute fold change prep to a paired t test foldChange lt rowMeans M c 1 3 M c 2 4 Get gene names and label the rows of the M matrix dHHE extract gene names Sw probes lt maLabels maGnames swirl norm dimnames M 1 lt sw probes Set up a factor which indicates which cell type is colored red JHH Set the factors factor indicates which cell type is colored Red gfac lt factor c swirl wild type swirl wild type JHHF Compute test statistics testStat lt aa teststat M gfac test pairt 187 Chapter 4 Examples Two Color Data Adjusted p values Compute adjusted p values for both Bonferroni and Benjamini and Graphics 188 Hochberg methods JHHF Compute adjusted p values rawp lt testStat pValue testObj lt mt rawp2adjp rawp proc c Bonferroni BH We can now print the top 10 genes gt testObj adjp 1 10 rawp Bonferroni BH 1
190. es per page Also note that for the MvA plot you can produce either a scatter plot with the degree of sampling specified in the Sample Size field or a hexbin plot with the number of bins specified by the Number of bins field Clicking Apply or OK generates the plots you selected Be aware that RNA Degradation and Principal Components plots may take some time to complete 201 Chapter 5 Quality Control Diagnostics and Filtering Diagnostics for Two Channel Data 202 The use of the Quality Control Diagnostics dialog for two channel data is identical to that for Affymetrix However there are a couple of options missing There is not a Genes Present plot nor an RNA Degradation plot Figure 5 11 displays the dialog with some options selected Quality Control Diagnostics Two Channel oj x r Data r MvA Plot Show Data of Type IV MvA Plot Two Channel E I Smooth Curves Data TPManay Raw zi Plot Type Scatter plot Array Name D Microarays Mi C Hex bin m Image Plot 3 Sample Size 20 I Image Plot ea Number of bins 420 Channel both hd Channel T jr d le annel Type X Me bale Lad T Intensity Boxplot Signal Ratio M z Boy Plot Type C By Array Color Map blue yellow x By Prnttip T Prin Comp Plot OK Cancel Apply K current Help Figure 5 11 Example dialog for Quality Control Diagnostics for two channel data FILTERING Array Filterin
191. esponse 3 11 5 of 78 ae 42830 defense response to pathogenic 6954 inflammatory response 9 bacteria 1 11 8 af 76 20 0 of 5 Figure 9 24 Defense response hierarchy in GO Biological Process ontology All that S ARRAYANALYZER needs to do in the filtering is the GO term The GO term can be selected from the GOAnnoData library Step 1 below This GO term is then used to pick off the corresponding Affy IDs on the mgu74av2 chip Step 2 below These Affy IDs are then used to subset the S ARRAYANALYZER data set prior to differential expression testing Step 3 below 1 Select GO terms of interest JHF Load the GO AnnoData library if not already loaded goterms lt unlist GOBPID2TERM position lt match defense response goterms def resp lt names GOBPID2TERM position def resp 1 GO 0006952 dH 2 Obtain identifiers Affy IDs for genes on the chip JHF with the GO annotation of interest gt library mgu74av2AnnoData gt def IDs lt mgu74av2GO02ALLPROBES names mgu74av2GO2ALLPROBES def resp JHF 3 Subset the expression dataset to genes JHF with the GO annotation of interest 365 Chapter 9 Annotation and Gene List Management 366 gt DAYSdefense lt DAYSrmaExprSet gt exprs DAYSdefense lt exprs DAYSrmaExprSet summ def IDs JHF 4 Perform differential expression analysis on the JHF expression data subset The filtering operation in Step 3 above reduces the number of genes fr
192. estimating the baseline variance function for each of the compared experimental conditions say U and V For example when duplicated arrays U Uo are used for condition U the variance of M expressed as U U on each percentile range of A expressed as U U is evaluated When there are more than duplicates all pairwise comparisons are pooled together for such estimation A non parametric local regression curve is then fit to the variance estimates on the percentile subintervals refer to Figure 7 18 as an example The baseline variance function for condition V is similarly derived and the LPE test statistic for comparison of median log intensities between the two samples is where s rPE 1 57 s1 Med 2n s Medy ny n and n numbers of replicates for the samples compared s M ed i 1 2 is the error estimate from the th LPE baseline error distribution at each median Med For more details see Lee and O Connell 2003 Note that the LPE statistic based on medians is robust to outliers if there are three or more replicates Running any of the above statistical procedures produces raw p values the p values associated with the individual statistical tests To make confident statements about differential expression for the entire 263 Chapter 7 Differential Expression Testing experiment you need to compute adjusted p values which control the family wise error rate or false discovery rate See the sectio
193. ests For the Two Sample Tests dialog in addition to the volcano heat Specific Plots map and chromosome plots you may also generate a Q Q Normal Probability plot of the test statistics This plot provides a visual assessment of the distribution of the test statistics relative to the standard normal distribution as shown in Figure 7 17 Graph Window 11 File View Options 8 2 T a pa D fe Quantiles of Standard Normal E GQNorm Figure 7 17 Q Q Normal Probability plots of the test statistics generated by the Two Sample Tests dialog 291 Chapter 7 Differential Expression Testing LPE Specific Plots For the LPE Test dialog in addition to the volcano heat map and 292 chromosome plots you may also generate plots of the local pooled error variance versus the overall intensity within experimental conditions Two plots are produced one for each experimental condition as shown in Figure 7 18 Graph Window 9 File View Options Y o o o i amp amp G ta F gt gt Ww Ww a a ad al 6 8 10 12 14 A for Ohr A for 24hr a gt Summary Volcano Plot variance Plot Figure 7 18 Plots of the local pooled error variance within treatments versus the overall intensity within treatments Differential Expression Analysis Plots ANOVA Specific For the ANOVA Test dialog in addition to the volcano for each Plots contrast and heat map you may also generate parallel coordinates plots for each co
194. ethods are typically represented with a dendrogram showing the hierarchy from all samples to individual samples or from all genes to individual genes Genes with obviously non significant expression values should be omitted from the clustering analyses Genes included in the clustering analyses may be chosen using the statistical hypothesis tests for differential expression described in Chapter 7 Differential Expression Testing It is important to understand that hierarchical approaches do not directly provide any reliable measure of confidence for clustered expression patterns A hierarchical clustering method heuristically reorganizes the genes based on its predefined association distance and allocation algorithm which only aids in discerning co expression patterns visually Therefore a validation step is required for such hierarchical clustering discoveries before further inference can be drawn For example a bootstrapping method can be used for assessing reliability of clustering classifications of a fixed known number of groups Kerr and Churchill 2001 Hierarchical clustering results are typically summarized with a dendrogram in which samples or genes are joined in a tree structure where the leaves branches successively join samples or genes that are most similar We note that this needs to be interpreted with care since hierarchical clustering imposes structure whether it is there or not and dendrograms then reflect that impos
195. evel Values Format The Level Values strings you enter can contain no spaces In this example we use WildType not Wild Type 3 In the Design Type group ensure that the Two Sample radio button is selected and the Dye Swap checkbox is checked 129 Chapter 4 Examples Two Color Data Step 2 Associate Files with Design Points 130 4 Clicking OK on the Create Modify Design dialog generates the experimental design points that populate the file selection grid You are now ready for Step 2 Associate Files with Design Points x m Factors Number of Arrays p a Number of Factors 1 Factor Name of Levels Across Chip Level Values A ZebraFish 2 No Swirl WildType m Design Type Two Sample C Loop M Dye Swap Factor ZebraFish gt Z Dye Swap C Reference Factor Zebrafish z C Other Ref Level C RefGy3 C Rec5 Dye Swap Cancel Hep Figure 4 3 The Create Modify Design dialog set up for the Zebrafish study Once the design is complete click OK to copy it into the File Selection tab of the Import Data From Two Channel dialog Notice that the number of rows for the File Selection box matches the number of arrays specified on the Create Modify Design dialog Furthermore values for the factor levels have been written into the Factor columns to facilitate associating files with design points If the experiment is balanced the factor level settings will b
196. f genes to 500 I Significant genes Recalculate m Expression Filtering Data on which to Filter Number of genes selected by filtering i PMarrayRawF_ J Genes with maximum fold change Sort Order greater than 2 f I Genes with Expression values Move Up Move Down exceeding fio in at least experiments chips OK Cancel Apply kk gt current Help Aey Figure 4 31 Filtering Options tab of the Cluster Analysis dialog ready for a hierarchical cluster analysis of the TP data 163 Chapter 4 Examples Two Color Data Differential Expression Analysis 164 Figure 4 32 Hierarchical clustering of the 500 most expressive genes in the TP data Note the dramatic shift in expression values between late left four channels and early right six channels times points We are now ready to focus on differential expression analysis Recall when we set up the design we stated one of the critical questions was to find how gene expression changed over time We can test for that by looking a simple contrasts between time points in the study To set up the analysis click ArrayAnalyzer gt Differential Expression Analysis LPE Test Select Two Channel data type TPMarrayRawFiltered norm as the Data drop down list and BH for adjustment Run a contrast comparison by selecting Time as the Factor and 31 as Compar
197. ferentially expressed genes are highlighted with color orange to indicate their location on the chromosome Hovering the mouse over one of the colored active points displays the gene ID in the upper right hand corner of the 90 graph as shown in Figure 3 21 o a E f 0 f E 2 53 riais PHH H J E E rA miooo p p E op pte yen Mf Af Gl HA ot ot ARR it tf Tt HA C aA BEL LLL AA LDL LM ALRARORNE 1A A a L E e i L IL i it L L as 4 expressed genes Figure 3 21 The chromosome plot displays the entire chromosome with differential expression marked up for positive down for negative for each gene on the chip The orange color indicates the location of the top 15 most significant differentially Variance Plots Two Sample Design The variance plots display the variance estimates used for the LPE test as a function of differential expression for each treatment condition In this example the plot shows the variance decreasing dramatically as differential expression increases as shown in Figure 3 22 o o oO 0 i S amp z gt gt wW wW a a al 8 10 12 14 4 6 8 10 12 A for 2hr A for 24hr Figure 3 22 Variance plots for the 2 hour and 24 hour data 91 Chapter 3 Examples Affymetrix Probe Level Data Annotation Clicking one of the hyperlinked points in one of the Top 15 Genes Summary the volcano plot or the heat map pops up a menu
198. file can be reused for another experiment with the same design by modifying the file locations and names and factor levels as needed In fact if you have many arrays in your experiment you can create a file with all the design content and read it with the Read Design button located at the top of the dialog Reading the design in MIAME Page MAS Variables amp Filtering Page Two Sample Design this way will fill the file name fields and their associated factor levels For information about creating a design file see Appendix A Creating a Design File MIAME is an acronym for Minimal Information About a Microarray Experiment This information should be entered on the second page of the Import Data from Affymetrix dialog This information is not required but it is used to label the resulting S PLUS object for later identification Once you ve entered MIAME information for any experiment the first three fields are saved and are filled automatically the next time you open this dialog I x File Selection MIAME MAS Variables amp Filtering CEL Filtering Options Experimenter s Name Bob Bryant Laboratory E Contact Information bob igrl org Experiment Title Melanoma Gel Matrix Experiment Description of melanoma was added at 2 hours and 24 hours later Fox et al 2001 This simple experimental design involves one factor matrix condition at two levels 2 and 24 hours with expression being measured twice on dupl
199. following manner 1 Select Affymetrix Summary from the Show Data of Type drop down list 2 Select MouseSwimExprSet norm from the Data drop down list 311 Chapter 8 Cluster Analysis 3 Go to the Filtering Options tab and select the Genes with maximum fold change greater than checkbox and set the associated field to 8 We want to limit the genes to only those with extreme fold change 4 Click the Recalculate button to the right in the Gene Sort Order Options group You should get 51 genes resulting from the filtering which will be done prior to clustering The resulting dialog tabs are displayed in Figures 8 3 and 8 4 aici General Options Filtering Options m Data r Partitioning Methods Show Data of Type J Partitioning Around Medoids Affymetrix Sumr gt nne ere Data MouseSwimE xp Auto y Array Name mgu74av2 Cluster on Genes r Response Variable T KMeans Response Expression val x Genter R z Standardization Standard value x Hierarchical Methods IV Hierarchical I Model based Cluster or Genes r Output Dist metric euclidean gt IZ Display Output in S PLUS IV Names on Graph Weighting method PERG I Save Output as HTML Save HTMLA myCluster html F Display HTML Output Save As myCluster Cancel Appt K j current Help Figure 8 3 General Options ready for simple hiera
200. from a spreadsheet To see descriptions of other arguments check the help file gt help importData For the melanoma data we do a series of four importData commands to read the four files import From rie O T Data Specs Options Filter r General r Additional Col names row auto imi Worksheet number Row name col auto hed Auto z Start col ooo IV Strings as factors End col en gt gt i IV Sort factor levels SIANA DO o I Labels as numbers Endiow KEN gt Century cutoff fi 930 m ASCII Format string Delimiter Decimal Point Period J x 1000s Separator None amp lV Separate Delimiters Date format Time format Cancel Apply if j current Figure 2 35 Setting the Start row option of the Import From File dialog gt cga lt importData paste getenv SHOME module ArrayAnalyzer examples OhA csv sep type ASCII gt cgb lt importData paste getenv SHOME module ArrayAnalyzer examples OhB csv sep type ASCII gt cg24a lt importData paste getenv SHOME module ArrayAnalyzer examples 24hA csv sep From the Command Line type ASCTI gt cg24b lt importData paste getenv SHOME module ArrayAnalyzer examples 24hB csv sep type ASCII The resulting objects are data frames so you can do whatever you want to do in the way of data summaries and exploratory plots First we ll take care of some orga
201. g Filtering S ARRAYANALYZER provides simple mechanisms for removing arrays or genes or both before proceeding with an analysis Because of the way S PLUS saves data on hard disk you don t have to worry about losing your original data The default object name for the result of any filtering operation is different than the original data object so the original data stays intact as long as you preserve the original name To start a filtering operation click ArrayAnalyzer Filtering from the main S PLUS menu The resulting dialog has two pages One for samples or entire arrays and one for genes Filtering on arrays or samples is straight forward 1 Select the type of data from the Show Data of Type drop down list Note that only Affymetrix Summary and Two Channel data are selectable Note If you know based on QC Diagnostics or by other means that you want to delete an array from a probe level data set you can do that once it is summarized 2 Select the data object you want to filter from the Data drop down list 3 Select the samples you want to drop from the analysis by selecting them in the Samples to Keep list and moving them to the Samples to Drop list by clicking on the Drop button Note Once you ve selected a sample and moved it into the Samples To Drop list you can move it back into the Samples to Keep list by selecting it in the Samples to Drop list and clicking the Keep button Note
202. h entering layout information in the dialog your input should look like what is shown in Figure 4 7 Click OK to create and save the layout object This returns you to the Import cDNA Data dialog Figure 4 4 where you can finish the import dialog To save the data object type a name in the Save Data Set As field near the bottom of the dialog Step 3 Save Output Remember this name It is used in other analysis steps such as quality checks filtering and normalization For our example enter SwirlMarrayRaw as the Saving the Design Reading Designs MIAME Page Two Sample Design object name The Display Report checkbox indicates whether or not to print summary information resulting from reading the data into an S PLUS report window Once you ve entered all the information on this tab you can save it for later use by clicking the Save Design File button at the top of the dialog A txt file is written to the directory of your choice with number of factors number of levels repetitions and the full path file names and their associated design points You can reuse the design file for another experiment with the same design by modifying the file locations and names and factor levels as needed In fact if you have many arrays in your experiment you can create a file with all the design content and read it with the Read Existing Design button which will set the number of arrays and fill the file name fields along with their associate
203. he Normalization Dialog Two Sample Design Furthermore normalization allows for the use of control spots on the array or spiked into the mRNA samples In Chapter 6 Pre Processing and Normalization we provide more detail on different methods of normalization Here we list them briefly and work through a simple example To normalize cDNA data go to the main S PLUS menu and select ArrayAnalyzer gt Normalization 4rrayAnalyzer Import Data gt Quality Control Diagnostics gt Filtering Affymetrix Expression Summary Differential a wae gt Cluster Analysis Annotation Gene List Management Figure 4 10 Selecting Normalization from the main S PLUS menu Show data of type Select the type of data you are normalizing In this example select Two Channel in the Show Data of Type field for the swirl example as shown in Figure 4 11 aox m Data Normalization Show Data of Type Two Channel x Normalization median hd Between Array none hd Data NIEHS1Raw x Plot Sare Ae NIEHS 1Raw NIEHS1RawDropControl NIEHS1RawNokeep PramilaMarrayRaw SwilMarayRaw TPMarrayRaw TPMarrayRawFiltered2 Cancel Appt if j current Help Figure 4 11 Selecting cDNA before selecting for a list of cDNA data Probe Set Show Before amp After C Only After Data In the Data field select SwirlMarrayRaw from the drop down list as the object of class marrayR
204. he actual raw data but rather data as summarized by Cluster Eisen et al 1998 and prepared for viewing in TreeView Eisen et al 1998 We treat it as raw data to show the cluster methods in S PLUS but the resulting output should not be directly compared with fig3a of Alizadeh et al 2000 Partitioning Around Medoids Example Since Alizadeh et al 2000 were interested in identifying two specific subpopulations within the DLBCL samples they may have used a partitioning clustering method We use the partitioning around medoids method pam This analysis provides some evidence for the existence of two subpopulations rather than three four or five subpopulations based on average silhouette width The average silhouette width for two subpopulations is 0 19 compared to 0 15 and 0 08 for three and four subpopulations However absolute values of the average silhouette width are fairly small in all cases Examples from the Command Line The partitioning around medoids analysis and graphical summaries are presented in Figure 8 14 and Figure 8 15 Figure 8 14 shows the two clusters projected onto biplot of the first two principal components A silhouette plot for two subpopulations is provided in Figure 8 15 partitioning 2 classes compare to 3 and up mat3a 2 pam lt pam t mat3a 2 plot mat3a 2 pam Component 2 Component1 These two components explain 45 12 of the point variability Figure 8 14 Partitioning around med
205. he expression intensities for all replicates of a given treatment condition An optional second argument number bins sets the number of the bins for the partition used to compute local error estimates The default value for number bins is 100 implying that 1 of the intensity values will be in each bin We are now ready to compute the LPE test The function used is 1pe0LIG which takes three arguments the baseline variance objects for each treatment condition and the size of the sample i e number of arrays for each treatment condition For the Melanoma data the call is gt LPEObj lt IpeOLIG OLIGgrp0 OLIGgrp24 sample c 2 2 gt testStat lt LPEObj outputL z stats IpeOLIG computes raw p values but you can adjust them to control the family wise error rate FWER or False Discovery Rate FDR with the mt rawp2adjp function mt rawp2adjp takes the vector of raw p values plus a character string indicating the adjustment procedure The function call is gt adjpObj lt mt rawp2adjp LPEObj pvalue proc Bonferroni We can plot the results in a Graphlet similar to what is obtained from the GUI We compute fold change first and then call the Graphlet function gt foldChange lt rowMeans LCG N 1 2 rowMeans LCG N 3 4 gt procedure lt Bonferroni gt LPESumm lt lpetest graphlet LCG N adjpObj adjp 2 LPEObj pvalue adjpObj index testStat foldChange procedure procedure chip name hgu95a vo
206. he mice continued for 4 weeks For controls age matched mice were used that did not exercise For more details see http cardiogenomics med harvard edu groups proj1 pages swim_home html This experiment is a two factor design with one factor indicating whether or not the mice were conditioned by swimming and the other factor indicating the amount of conditioning which ranged from 10 minutes to daily exercise for four weeks Gene expression was measured in three replicate sets i e three mice for each time point and each treatment swim control no swim combination The main hypothesis of interest involves discovering genes showing differential expression between the time points and the treatment conditions It is believed that certain genes are involved in the enlargement of ventricular mass during chronic conditioning The arrays and data files are listed in Table 2 1 Table 2 1 Experimental design and file association for the melanoma cancer study Cond Time Rep Array label File Name Swim 10min 1 Swim 10min 1 Swim10min1 Swim 10min 2 Swim 10min2 Swim10min2 Swim 10min 3 Swim 10min3 Swim10min3 Swim 2 5days 1 Swim 2 5d1 Swim2 5d1 Swim 2 5days 2 Swim2 5d2 Swim2 5d2 Swim 2 5days 3 Swim2 5d3 Swim2 5d3 One Way Design Table 2 1 Experimental design and file association for the melanoma cancer study Co
207. he normalization dialog The y axis typically shows the intensity log ratio M The x axis shows the grouping chip or print tip Please see section Box Plots on page 218 for additional information on box plots From the command line we can create a box plot by print tip groups for chip 93 of the swirl dataset discussed in section Mutant Zebra Fish Data on page 127 as follows Information on the swirl data set is also available in the swirl help file type he1p swir1 gt maBoxplot swirl 3 main Swirl array 93 pre normalized Normalization Methods for Two Channel Data Swirl array 93 pre normalized e 8 af ET i i lm A i 3 hi x i l A i 7 ret 5 4 H m g A mo ri i t TE f rm Tt Obed H ri i i E E i oe a EEFT 4 ye ae ae a od TEREE o l e 1 1 4 2 1 3 1 4 2 1 2 2 2 3 2 4 3 1 3 2 3 3 3 4 4 1 4 2 4 3 4 4 PrintTip Figure 6 4 Box plot for chip 93 of swirl dataset before it is normalized Plots for two channel data are available from the command line using the functions in Table 6 1 Following are examples of how to use these functions Please refer to the marrayPlots library pdf splus62 library marrayPlots marrayPlots pdf and the marrayPlots library help files for detailed descriptions of the function arguments The section Notes For Command Line Users on page 221 discusses the meaning of the x y and z parameters u
208. he selection of the Affymetrix Quality Control Diagnostics dialog ArrayAnalyzer Import Data gt Quality Control Diagnostics Filtering Affymetrix Expression Summary Normalization Differential Expression Analysis gt Cluster Analysis Annotation Gene List Management Figure 5 9 Selecting Quality Control Diagnostics for Affymetrix data The resulting dialog has four groups of controls 1 Data Selecting Data Selecting Graphical Options Quality Control Diagnostics 2 Image plot 3 MvA plot Bland Altman 4 Other plots Quality Control Diagnostics Affymetrix oj x m Data Mv Plot Bland Altman Show Data of Type IV MvA Plot aty metriz CEL zi Formed From C Median Data ERNE T Replicate Array Name mgu74av2 J Smooth Curves rl Po moas Plot Type Scatter plot J Image Plot C Hex bin Page Layout 1 array per pag Y Color Map Sample Size 20 E Number of bin 20 r Other Plots Genes Present Plot I Intensity Boxplot I RNA Degrad Plot J Prin Comp Plot Cancel Apply k j current Help Figure 5 10 The Quality Control Diagnostics Affymetrix dialog Select data by choosing Affymetrix Summary or Affymetrix CEL from the Show Data of Type drop down list and then select the data of your choice Now choose the plots you wish to display Note the page layout options for the image Plot one or four imag
209. he sort order list 345 Chapter 9 Annotation and Gene List Management In this example there are 136 genes meeting the filtering criteria and we are choosing the top 10 of these based on p value followed by fold change The results of this filtering are shown in Figures 9 6 to 9 9 File Edit View Favorites Tools Help Back gt amp A A Gsearch Favorites media 3 awm d Address ja http www ncbi nih gov LocusLink LocRpt cqi7I 7465 2c83939 2c 10051 2c2364 x 60 gt gt NCBI PubMed Entrez BLAST OMIM Map Taxonomy Structure Search LocusLink Display Brief Organism All Query Go Clear View Hs WEE1 One of 10 Locil_SeveAllLoci_ MNOPQRSTUVWXYZ Top of P ic Alignments spanning 16958 bps UNIGENE MAP VAR HOMOL WEEL WEEL homolog S pombe LocusID 7465 Overview RefSeq Summary This gene encodes a nuclear protein which is a tyrosine kinase belonging to the Ser Thr family of protein kinases This protein catalyzes the inhibitory tyrosine phosphorylation of CDC2 cyclin B kinase and appears to coordinate the transition between DNA replication and mitosis by protecting the nucleus from cytoplasmically activated CDC2 kinase Collabo Download Locus Type gene with protein product function known or X rf ee Pe Figure 9 6 LocusLink results for Melanoma data with fold change set to greater than 2 and p value set to significant adjusted p lt 0 05 for LPE test Bonferroni FWE
210. help file for more details on the data gt tmp lt bg correct Dilution method mas gt tmp lt bg correct rma Dilution gt tmp lt bg correct mas Dilution GC RMA is a robust multi array expression measure using sequence information It provides background estimates based on a model using GC content GC RMA is a modified version of RMA that models intensity of probe level data as a function of GC content The theory being that you would expect to see higher intensity values for probes that are GC rich due to increased binding This is seen as an improvement over RMA which does not consider the physical processes gcrma calls the rma function after background correction Therefore expression intensities returned from gcrma are normalized and summarized using quantile normalization and median polish summarization The function accepts an AffyBatch class object and returns and exprSet class object One may wish to correct the PM intensities in a ProbeSet for non specific binding hybridization that occurred at random Affymetrix chips provide a mechanism for measuring non specific binding 1 A ProbeSet object is the collection of cloned PM and MM spot intensities for one gene 239 Chapter 6 Pre Processing and Normalization An Example with pmcorrect 240 through the mismatch probes MM The amount of binding that occurs at these spots is a measure of the amount of random binding that is occurring in the experi
211. her files in this block must be from the same path and the valuename RelativePath should be defined see table below If RelativePath is not defined then the current working directory is used to find the files name is the name to assign this array to import factor A level is the level to use for this file for factor A The level specified here should match one of the levels for this factor defined in the FactorInfo block see above and likewise for factor B level etc Currently only two factors are allowed in S ARRAYANALYZER so only the third and fourth delimited values in the lines are used to define levels for each factor As an example consider the lines FactorInfo START A CondTime Swim3wks Swim4wks Swim4wks 1wk NoSwim4wks NoSwim4wks 1lwk FactorInfo END Design START Swim3wl txt SSwim3wks1 Swim3wks Design END Format Specification The file to import is Swim3w1 txt Since there is no path specified to the file here and since the valuename RelativePath was not specified in the block the current working directory is used to find Swim3wl txt The name is Swim3wks1 The factor A level for this file is Swim3wks Notice that Swim3wks appears as an allowable level for factor A in the FactorInfo block Table J 5 Table of Design valuenames Value Name Rules Example RelativePath OPTIONAL If not specified and no full path is specified for a file the current working directory will
212. hey can be used to assess the quality of your microarray data Once quality assessments are complete you can apply filtering methods to remove unwanted arrays and genes from your analyses These methods are described subsequently in section Filtering on page 203 The diagnostic methods vary slightly depending on the type of microarray data you have For Affymetrix MAS data the following methods are available 1 Color image plot of the entire array M vs A plot as either a scatter plot or a hexbin plot Genes Present plot 2 3 4 Intensity boxplot 5 RNA degradation plot 6 Principal components plot Color Image Plot Quality Control Diagnostics The color image plot provides a whole array view of expression intensities Each pixel of the image represents the expression intensity of one spot on the array The intensities are color coded to provide a quick visual inspection of the entire array The pixels are arranged geographically just as they are on the array Figure 5 1 shows an example image plot for Affymetrix CEL data 11 0 10 0 3 2 8 3 74 6 6 5 7 4 8 3 9 3 0 iss a ach ai Ra att fan 5 1 Image plot example for Affymetrix CEL data 100 200 300 400 500 600 0 The image plot is ideal for discovering imperfections on the array surface such as scratches contamination from debris and uneven hybridization This is a useful plot for both Affymetrix and two channel microarray data 193 Chapter 5 Qu
213. hip Array Biological Process 7 HG U95Av2 Probe Set Count Threshold Probe Set Percentage Threshold 0 0 Node Color Scheme Node Shape Number of probe sets z Ellipses z Node Text Options Graph Window M G0 Ds W GO Titles New Window z Probe Set Count IV Probe Set Percent Upload Probe Set List C Program Files Insigh Browse Get Graph Ey 888 DNA CHIP 888 362 2447 44 0 1628 552550 feedback e mail support terms of use privacy policy L A eg internet Figure 9 11 Affymetrix NetAffx Web site launched by StARRAYANALYZER with list of IDs for Melanoma data ie 352 Annotation Libraries Ele Edt View Favorites Tools Help Ea Back O A A Asearch GFavorites GHmedia lt 4 Eh S fw S Address https jwaw affymetrix comfanalysisiquery go view resutisp SY G0 Links Sat A GONode Tak s probe sets in Interactive Query Annotation Table gt In This Window z Summary HG U95Av2 probe sets 12625 HG U95Av2 probe sets annotated biological process 7699 Uploaded probe sets 5 Uploaded probe sets annotated biological process 1 Nodes J Nodes with given thresholds 19 19538 protein metabolism 1 0 0 of 1648 7582 physiological process 1 8152 metabolism 1 E metabolism 1 0 0 of 6557 0 0 of 4502 0 1 of 619 biological process 9987 cellular process 1 8151 call growth and or maintenance 1 0 0 of 4523 0 0 of 2501 8283 call proliferation 1 0
214. hours The value in the lower left panel of the plot is the interquartile range of M Expression Summaries 7 i T i T i rey w o oT i i fi e e e e 9 _ N 7 r a f a l 4 cg2a CEL cg2b CEL cg24a CEL cg24b CEL Figure 3 15 Boxplot of log expression intensities for the four samples after applying the composite RMA procedure 83 Chapter 3 Examples Affymetrix Probe Level Data Differential After normalizing and summarizing the probe level data we are ready Expression to do differential expression analysis From the main menu open the Analysis Local Pooled Error Test dialog by clicking ArrayAnalyzer gt Differential Expression Analysis gt LPE Test ArrayAnalyzer Import Data gt Quality Control Diagnostics gt Filtering Affymetrix Expression Summary Normalization Differential Expression Analysis ee N Cluster Analysis Two Sample Tests Annotation ANOVA Gene List Management Figure 3 16 Selecting LPE Test to open the Local Pooled Error Test dialog Local Pooled The Local Pooled Error LPE Test dialog contains groups for Error Test specifying the data and levels of the factor you wish to compare how to adjust for multiple testing options for controlling variance functio
215. iations File Type Note that the File Type Probe Level CEL listed below the files is automatically detected once a file is selected The dialog is designed to prohibit mixing file types The Array Type The Array Type is automatically detected for probe level data For this experiment it is MG_U74Av2 or mgu74av2 as it is listed in the Array Type drop down list 99 Chapter 3 Examples Affymetrix Probe Level Data Step 3 Saving the Data Object MIAME Page MAS Variables amp Filtering Page 100 To save the data object type a name in the Save As field near the bottom of the dialog Step 3 Save Output Remember this name as it is used in the other analysis steps such as quality checks filtering and normalization For our example enter SurgeryAffyBatch as the object name The Display Report checkbox indicates whether or not to print summary information into an S PLUS report window Step 3 Save Output Save Data Set As SurgenAtfyB atch M Display Report Figure 3 29 Saving the imported data as SurgeryAffyBatch Saving the Design Once you ve entered all the information on this tab you can save it for later use by clicking the Save Design button at the top of the dialog A txt file is written to the directory of your choice with number of factors number of levels repetitions and the full path file names and their associated factor levels Reading Designs This design file can be reused for anoth
216. icated arrays for each time point Existing Notes D Microarrays D ata Affymetrix CG Figure 3 7 Entering experiment information on the MIAME page The third page of the Import Data From Affymetrix dialog is for variable and row selection for Affymetrix MAS 4 5 data It is not used for CEL data 75 Chapter 3 Examples Affymetrix Probe Level Data CEL Filtering The fourth page of the Import Data From Affymetrix dialog is for spot filtering probe level CEL data The options are 1 A checkbox to convert spots labeled as MASKS to missing so they aren t used in subsequent analyses 2 A checkbox to convert spots marked as OUTLIERS to missing so they aren t used in subsequent analyses Import Data From Affymetrix x File Selection MIAME MAS Variables amp Fittering CEL Filtering Options CEL File Filtering I Set spots marked as MASKS to missing NA I Set spots marked as OUTLIERS to missing NA Cancel x gt l j entries Figure 3 8 CEL Filtering page of the Import Data From Affymetrix dialog Options The last page on the Import Data From Affymetrix dialog is the Options tab The tab provides two options used during data import 1 The number of header lines to skip in each file before reading the data Normally this can be detected automatically but it is provided as an option for unusual cases where auto detection can not find the row with column names 2 The delimiter
217. id Column fields For our example they are automatically detected and filled in as Row and Column respectively 4 Control A column indicating which spots are control spots Enter ID for the Zebrafish example 5 Control Value Values which label the control spots In some cases this may be the single value control In others it may be the numeric values 1 and 1 indicating positive and negative controls 133 Chapter 4 Examples Two Color Data Step 3 Save Data Set As 134 6 Gene Name A column with unique gene or probe names Enter Name for the Zebrafish example 7 Save Array Layout As A name for the saved layout object You need to name the layout object which can be reused for all arrays with the same layout Enter SwirlLayout as the object name in the Save Array Layout As field which is used when control returns to the Import Data From Two Channel dialog zx File Selection Options Scanner Layout C Program Files Insightfulsplus62 modulesArray nalyzersex Browse m Outer Grid number of grid rows and columns Grid Rows 4 Grid Columns fa Inner Grid choose columns in data file representing Inner Grid Row Rows o Inner Grid Col Coum S Control ID Control Value eont tt ss SSSSCiS Gene Name Name YW m Output Save Layout As SwirlLayout Cancel K j current Help Figure 4 7 The completed layout for the swirl example When you finis
218. identifiers x Select the type of input identifier ClonelD x Figure 9 17 Stanford Source site launched from S ARRAYANALYZER with LocusLink gene list upload Figure 9 18 shows the General Options page of the Annotation dialog with options chosen from Open OntoExpress group This sends a log in message to the Onto Express site http vortex cs wayne edu projects htm Onto Express You will need an account on OntoExpress for this log in to be successful You can select options for a gene enrichment analysis on OntoExpress from S ARRAYANALYZER including the distribution of the test binomial hypergeometric and chi square and FWER FDR correction Bonferroni Sidak Holm BH FRD 357 Chapter 9 Annotation and Gene List Management 358 S ARRAYANALYZER sends the selected filtered gene list of IDs to OntoExpress for this analysis OntoExpress then opens itself in a java applet with the analyses done and additional analyses available In this case we send the Melanoma data with fold change set to greater than 2 and p value set to significant adjusted p lt 0 05 for LPE test Bonferroni FWER in this case top 10 genes based on P Value followed by fold change Results are displayed in the OntoExpress applet in Figure 9 19 ox General Options Filtering Options m Data m Use LocusLink Ds Show Data of Type I Save LocusLink IDs to File DitfExprTest zi LocusLink File Data eal PEBon ha Array Name hou95av
219. ify 1 The number of arrays to be read 2 The number of factors in the experiment Currently one or two are allowed 3 The name number of levels and level values for each factor To modify the default factor Name of Levels and Level Values type them into the appropriate field For the two way analysis of the Surgery data there are 18 arrays and two factors The Age factor has two levels Young and Old The Time factor has three levels Ohr lhr and 4hr The resulting dialog is displayed in Figure 3 27 x Number of Arrays fis Number of Factors B 4 ofLlevels Level Values Old Young Ohr thr 4hr Cancel Help Figure 3 27 The Create Modify Design dialog with two way experiment setup Once the design is complete click OK to copy it into the File Selection page of the Import Data From Affymetrix dialog Notice that the number of rows for the File Selection box is modified to match the number of arrays specified on the Create Modify Design dialog Furthermore values for the factor levels have been written into the Factor columns to facilitate associating files with design points If the experiment is balanced the factor level settings will be exactly as needed However the level values can be reset when the experiment is unbalanced or if you prefer an order different from the default The next step is to associate files with each design point To do so right click in one of the file fields and browse to the lo
220. ilin 2 Volcano Plot thr Old Young Probe Id 102002 Accession Number LocusLink T 2 2 a k D 2 a 3 a Ko D o al Figure 3 41 A volcano plot which is the logarithm of p value versus fold change The Benjamini Hochberg FDR correction method was used for the surgery data at an FDR of 0 05 The BH correction is less conservative than the Bonferroni procedure yet maintains a small proportion of false positives amongst those genes tagged as significant Figure 3 42 displays the parallel coords plots which show how the expression intensity varies across the treatment conditions for the significant genes 113 Chapter 3 Examples Affymetrix Probe Level Data LPE Test Dialog 114 Parallel Coords Plot 1hr Old Young gt 0 E o amp 2 0 0 L a x Ww o E o Oo r T T T T 1 Old Ohr Old 1hr Old 4hr Young Ohr Young thr Young 4hr Experimental Condition 4 Ling f Summary Thr Old Youna Summary 4hr Old Youna J Volcano Ohr Old Young J Volcano Thr Old Young J Voleano 4hr Old Youna Parallel Thr Old Young Heatmap Figure 3 42 Parallel coords plots of expression intensities for the significant genes Each line shows the expression intensity profile for a gene across the experimental conditions An alternative way to think about the analysis is a simple comparison between Young and Old ignoring the Time component This then becomes a two sample problem S ARRAYA
221. iltering CEL Fitering Options r Step 1 Specify Design Read Existing Design Create Modify Design Save Design File r Step 2 Associate Files with Design Points r Step 3 Save Output Save Data Set As mySet IV Display Report Figure 2 2 The Import Data from Affymetrix dialog 22 Step I Create The Experimental Design One Way Design The Import Data from Affymetrix dialog has five pages e File Selection This page must be completed in order to create a data object for continued analysis e MIAME Completing this page is optional but highly recommended because information on the MIAME tab is used for labeling tables and graphs e MAS Variables amp Filtering This page has default settings depending on the type of data files e g MAS4 or MASS you select It also allows the selection of other variables which can be used for more general filtering by using the Filtering dialog e CEL Filtering This page allows a couple of options for filtering out spots not to be used in subsequent analyses when importing probe level data e Options This page provides options for specifying the number of header lines to skip and the delimiter used in the data file Data import is accomplished in three steps 1 Create the experimental design 2 Associate files with design points 3 Specify a name for saving the resulting data object Before we can begin to associate data file
222. ing the loess function twoD 2D spatial location normalization using the loess function printTipLoess within print tip group intensity dependent location normalization using the loess function scalePrintTipMAD within print tip group intensity dependent location normalization followed by within print tip group scale normalization using the median absolute deviation MAD The default normalization method is printTipLoess To normalize swirl raw with the scalePrintTipMAD method the function call is wa gt swirl norm lt maNorm swirl raw norm s Note that only the first letter of the method s in this case is needed We can do a series of plots to compare before and after normalization First we do a pair of M vs A plots The maP1ot function handles all the details We pick off one of the arrays the third one for simplicity gt maPlot swirl rawL 3 main Pre normalization MvA Plot gt maPlot swirl norml 3 main Post normalization MvA 183 Chapter 4 Examples Two Color Data Pre normalization MvA Plot Figure 4 44 Pre normalized M vs A plot for the swirl data 184 From the Command Line After scalePrintTipMAD Normalization Figure 4 45 Post normalized M vs A plot for the swirl data We can also do boxplots as a function of print tip groups as follows gt par mfrow c 1 2 gt maBoxplot swirl rawL 3 main Pre normalization srt 9
223. inimum recommended system configuration is a Pentium ITI 1Ghz processor at least 1GB of RAM and an SVGA or better graphics card and monitor You must have at least 850MB of free disk space for the typical installation and even if not installing on drive C an additional 2MB of free disk space on drive C to unpack the distribution To install S tARRAYANALYZER insert the S ARRAYANALYZER CD double click the setup exe file in the CD ROM drive of your Windows Explorer and follow the step by step installation instructions In S PLUS load the S ARRAYANALYZER module from the command line by entering gt module ArrayAnalyzer You can also load S ARRAYANALYZER by choosing File gt Load Module and selecting ArrayAnalyzer from the menu To detach or unload StARRAYANALYZER type gt detach ArrayAnalyzer S ARRAYANALYZER also includes an online HTML Help system for all the available functions After you have loaded the S ARRAYANALYZER module you can get help for any command by using the or help function For example if want help on the maBoxP1ot function simply type gt module ArrayAnalyzer gt help maBoxPlot at the Command line HTML Help Online Reference Technical Support Supported Platforms and System Requirements HTML Help in S PLUS is based on Microsoft Internet Explorer and uses an HTML window to display the help files You can access help on any function or GUI dialog in St ARRAYANALYZER from the m
224. interest listed in Table 3 3 117 Chapter 3 Examples Affymetrix Probe Level Data 118 The results of the ANOVA are contrasts of Old and Young for each time point Use the Gene List Management dialog to combine the gene lists from each contrast Open the Gene List Management dialog by clicking ArrayAnalyzer gt Gene List Management In Data Group 1 select DiffExpr Test from the Show Data of Type drop down list Then select the Surgery ANOVABsYoungBH object from the Data drop down list and select one of the contrasts listed in Figure 3 46 Start with 1hr Old Young In Data Group 2 also select DiffExprTest as the data type and choose 4hr Old Young as the contrast Now check the Union radio button in the Output Options group confirm that the Venn Diagram checkbox is checked and type in SurgeryGeneList in the Save As field to save the resulting gene list Gene List Management d la xj Data Group 1 _ Data Group 3 Show Data of Type Show Data of Type Data SugewANOVA gt Data x Array Name mqu74av2 Array Name f undetermined gt Contrast ih OldYoung Contrast x Cluster Mem m y Cluster Mem x gt Data Group 2 Output Options Show Data of Type Choose gene list creation l DiffExpTest Data SurgerANOVA v C Intersection Array Name mgu74av2 Union Contrast ahr 01d Young EAI Wy Nenn Disgam Elite i Save As SurgeryGeneList OK Cancel Apply K j current Help Figure 3 46
225. ionaly intensive gt eset lt expresso Dilution df normalize method invariantset df bg correct FALSE pmcorrect method pmonly dF summary method 1iwong This gives the current PM only default The reduced model previous default can be obtained using pmcorrect method subtractmm RMA method of Irizarry et al 2002 The RMA method of Irizarry et al 2002 can be obtained using expresso as follows gt eset lt expresso affybatch example normalize method quantiles bgcorrect method rma pmcorrect method pmonly summary method medianpolish Equivalently the rma function can be used and is faster for this series of operations gt eset lt rma affybatch example Summarization in S ARRAYANALYZER GUI Dialog Pre Processing and Normalization for Affymetrix Probe Level Data Affymetrix Expression Summary of x Data Options CEL Data Summary Options GCRMA Save As ExprSet Bi abet aetna Mix amp Match Bkad Correction mas hd Perfect Match mas g Summary avadiff al Normalization none he r Summary Graphics M Mva Plot M Box Plot Cancel Appt current Help Figure 6 6 Affymetrix Expression Summary dialog The Affymetrix Expression Summary dialog provides a convenient way to transform raw probe level data into data ready to be tested for differential expression This dialog provides options for background correction
226. ip Non controls by Print Tip PrintTip PrintTip Figure 4 42 Controls versus noncontrols by print tip group 181 Chapter 4 Examples Two Color Data 182 We can also plot controls versus noncontrols with a boxplot for each chip as follows JHHF Boxplots of controls vs noncontrols gt graphsheet gt par mfrow c 1 2 gt maBoxplot swirl raw controls main Controls by Print Tip Group srt 90 gt maBoxplot swirl raw main Non controls by Print Tip Group srt 90 Figure 4 43 displays the resulting graph Controls Across Chips Non controls Across Chips 81 82 93 94 81 82 93 94 Figure 4 43 Controls versus noncontrols across chips Normalization From the Command Line The swirl raw object resulting from the call to read marrayRaw is an object of class marrayRaw Removing the controls does not effect the class of the object The normalization function we use for marrayRaw objects is maNorm Its arguments are listed here mbatch Object of class marrayRaw containing intensity data for the batch of arrays to be normalized An object of class marrayNorm may also be passed if normalization is performed in several steps norm Character string specifying the normalization procedures The options to the norm argument are none no normalization median global median location normalization loess global intensity or A dependent location normalization us
227. is They include the t test with or without assuming equal variance for the groups the Wilcoxon rank sum test and several permutations tests In addition the LPE test procedure which produces improved error estimates when there is little replication in the design is implemented for two sample problems All these procedures are suitable for doing simple comparisons between two groups treatment versus control tissue 1 versus tissue 2 etc For more details about two sample procedures see Chapter 7 Differential Expression Testing In this section we step through the analysis of an experiment using an MM5 melanoma cell line in which a gel matrix that simulates the in vivo cellular condition and progression of melanoma was added at 2 hours and 24 hours Fox et al 2001 This simple experimental design involves one factor matrix condition at two levels 2 and 24 hours with expression being measured twice on duplicated arrays for each time point Condition replication and file name for each chip are displayed in Table 3 2 Table 3 1 Experimental design and file association for the melanoma cancer study Experimental Condition Repetition chip label File Name 2 hours 1 cg2a cg2a CEL 2 hours 2 cg2b cg2b CEL 24 hours 1 cg24a cg24a CEL 24 hours 2 cg24b cg24b CEL The fundamental question of the study is Which genes are active at 24 hours that weren t active at 2 hours These differentially expressed gen
228. ive genes The two tabs should now look like Figures 4 30 and 4 31 Now click OK or Apply to run the analysis Figure 4 32 displays the results of the cluster analysis The dendrogram on the vertical axis corresponds to clustering on genes The dendrogram on the top horizontal axis corresponds to clustering on experimental conditions or arrays In this graph the early times 1 7 11 are on the right side the right six channels and the later times 27 31 are on the left left four channels Note the clear separation in expression between the two time groups Genes that are positively expressing green in color during the early times are negatively expressing red in color during the late times In fact if you look closely at the expression pattern in the right six channels you can see a shift in the expression values for genes in the top of the list as you scan from 1 hour to 11 hours The values shift from Two Way Reference Design negative red for 1 hour to positive green at 11 hours This provides a simple verification of the expression patterns that Bozdech et al 2003 discovered in their analysis FEE General Options Filtering Options m Contrast Filteing gt r Gene List Filtering Data on which to Filter Data on which to Filter MouseAN OVA x GeneListLPEAr x Contrast NoSwimdwks x I Filter on Gene List J Genes with fold change r Gene Sort Order Options greater then Limit number o
229. ke a copy of the Melanoma S PLUS object JHF from RMA and LPE with Bonferroni FWER correction summ0bj lt cgLPEBon allData 359 Chapter 9 Annotation and Gene List Management 360 JHF 2 Filter the data to genes with fold change JHF greater than 2 and LPE p value less than 0 001 fc2 p001 lt summObj summObj AdjPvalue lt 0 001 amp summObj foldChange gt 2 mel gnames lt fc2 p001 GeneName JHF 3 Get LocusLink IDs and make a call to locuslink mel linames lt as numeric unlist hgu95av2LOCUSID mel gnames locuslinkByID mel 11names JHF 4 Get accession numbers and call to Entrez uids lt unlist hgu95av2ACCNUM fc2 p001 GeneName genbank uids type accession disp browser JHF 5 Get Pubmed ids and call to Pubmed for articles pmedids lt hgu95av2PMID mel gnames pubmed pmedids disp browser In this example the filtering of the Melanoma data with fold change greater than 2 and p value set to significant adjusted p lt 0 05 for LPE test Bonferroni FWER in this case produces four genes The LocusLink information for these four genes is obtained using the S PLUs function locuslinkByID This returns the Web page shown in Figure 9 20 Note that the View field in LocusLink is populated with the four genes identified in the analysis Annotation Libraries Z LocusLink Report Microsoft Internet Explorer E i ioj x File Edit View Favorites Tools Help IBak e A search Favorites meda
230. king OK or Apply generates the Venn diagram displayed in Chapter 4 36 Examples Two Color Data Note the large number of differentially expressed genes for large time spans and relatively few for short time spans We expect short time spans to have similar gene lists and consequently few differentially expressed genes Also note little overlap between the three gene lists indicating changing gene expression over time Two Way Reference Design Gene List Management A xj Data Group 1 Data Group 3 Show Data of Type Show Data of Type Data TPLPESIm gt Data TPLPESIm27 gt Array Name fiplayout Array Name fiplayout Contrast 31 1 x Contrast fz Cluster Mem i Cluster Mem gt Data Group 2 Output Options Show Data of Type Choose gene list creation Dire xprT est 7 Data TPLPE31 m7 x gt Intersection Array Name iplayout Union Contrast 31 7 X JV Venn Diagram Cluster Mem Z Save As myCieneList OK Cancel Apply K f current Help Figure 4 35 The Gene List Management dialog set to compare gene lists from different contrast tests TPLPE31m1 TPLPE31m7 31 1 31 7 61 1072 44 TPLPE31m27 31 27 Number of Genes in Union 1526 Number of Genes in Intersection 6 Figure 4 36 Venn diagram contrasting lists generated by different contrast tests 167 Chapter 4 Examples Two Color Data Annotation Creating the Named List 168 For Affymetri
231. ks 3 reps 10 min 5 3 reps 3 wks 3 reps 4 wks 6 3 reps 4 wks 3 reps 4 wks 7 3 reps 4 wks 1 wk 3 reps 4 wks lwk For simplicity in this first example we will focus on a subset of the experimental data We will look only at the 3 week 4 week and 4 week 1 week data i e tests 5 6 and 7 The design is still unbalanced because of the reuse of the control data for the mice conditioned for three weeks Consequently the setup remains a one way ANOVA with five factor levels Swim3wks Swim4wks Swim4wks 1wk NoSwim4wks and NoSwim4wks 1wk Start the example by reading in the arrays To import Affymetrix data from the main S PLUS menu select ArrayAnalyzer gt Import Data gt From Affymetrix ArrayAnalyzer Import Data Quality Control Diagnostics gt Fromo Channel Filtering Affymetrix Expression Summary Normalization Differential Expression Analysis gt Cluster Analysis Annotation Gene List Management Figure 2 1 Menu selection to import Affymetrix data 21 Chapter 2 Examples Affymetrix MAS Data Import Figure 2 2 shows the Import Data from Affymetrix dialog with Affymetrix Data the File Selection page displayed The primary task of the import Dialog process associates data files with experimental conditions and selects variable columns that are used in subsequent analysis Import Data From Affymetrix Xx File Selection MIAME MAS Variables amp F
232. lan et al 1994 mice do not dive The water temperature was maintained at 30 32 degrees C Existing Notes lt Fie Path or URL gt Figure 2 8 Entering experiment information in the MIAME page MAS Variables amp Filtering Page One Way Design The third page of the Import Data From Affymetrix dialog is for variable and row selection When reading MAS 4 5 data this page is automatically filled The Probe Name and Expr Intensities drop down fields are for selecting the columns in the data files corresponding to the probe names and expression intensities respectively Although it is possible to change the variables in the Probe Name and Expr Intensities fields in this dialog it is not recommended These fields correspond to the columns read from the files and are used in subsequent analyses The dialogs that follow in the data analysis e g normalization and differential expression testing expect expression data without control rows Import Data From Affymetrix K xj File Selection MIAME MAS Variables amp Filtering CEL Filtering Options l m Import Variable Names Probe Name Probe Set Name z Expr Intensities Signal 7 IV Apply Log2 Transformation gt Remove Probe Set IV If Detection is Absent 4 in all samples I If DetectionPValue is less than footy in all samples IV If Pairs Used is less than FP x in all samples IV If Control where Control prefix is AFFX Extra Varia
233. latform in drug discovery e g functional genomics and drug candidate evaluation toxicogenomics Their utility lies in the ability to simultaneously quantify the relative activity or differential expression of many genes under different biological conditions Some common uses of microarray experiments are to e Classify diseases and their subtypes e Identify and validate new targets for drug discovery e Improve understanding of biological processes e Evaluate drug candidates against drugs with known toxic side effects Develop personalized treatment plans tailored to genotypes It is not our intention to discuss in depth the biology of microarrays If you are new to this area you should investigate the references listed at the end of the chapters in this manual as most chapters provide references with detailed information We give here a brief overview for those new to the area A microarray consists of a slide with genes or active segments of genes attached at spots on a regularly spaced grid There may be anywhere from a few to tens of thousands of genes spotted on a single microarray which may occupy one or more slides At each spot one gene or an active segment of a gene is represented tens of thousands of times by cloning it and fixing all the duplicates to the spot on the slide 11 Chapter 1 Introduction To Microarray Data 12 Figure 1 3 Microarray experiments produce gene expression images like the one pic
234. lcano plot T heatmap plot T chromosome plot T html output F variance plot T smoother df 10 trim 5 OLIGgrpl OLIGgrp0 OLIGgrp2 O0LIGgrp24 var xlabs c A for cg A for cg24 summary name LPESumm open browser F The first six critical arguments to Ipetest graphlet are 1 the expression set object e g LCG N 2 the vector of adjusted p values The Output Table From the Command Line 3 the vector of raw p values D the order vector from mt rawp2adjp for sorting from most differentially expressed to least vector of fold change values 5 6 family wise error rate default 0 05 7 p value adjustment procedure 8 chip name e g hgu95a The rest of the arguments are mostly for controlling output which plots are created whether it should be HTML output or not the smoothing parameters for the loess smoother used in the variance plots etc See the help file for 1petest graphlet for more detail The resulting plot is a java graphlet in an S PLUS graphics device Note that HTML output is turned off html output F The graphlet displayed in S PLUS and the browser are the same except that points are not linked to annotation databases in the S PLUS display The output of the Ipetest graphlet function contains information for annotation and for ranking the genes based on the magnitude of differential expression gt LPESumm 1 10 c 1 6 GeneIndex foldChange Pvalue AdjPvalue 32314_g_at 1608 3 720782 0
235. le and the another for the Graphlet 275 Chapter 7 Differential Expression Testing GUI FOR LPE TESTING LPE Testing The dialog for LPE testing is displayed in Figure 7 7 Open the dialog Dialog Input from the main S PLUS menu by clicking ArrayAnalyzer gt Differential Expression Analysis gt LPE Test The dialog is arranged in five main groups p Data e Options Variance Estimator Graph Options Output ao F OW SS Data The Data group allows you to select the expression object for testing You start by selecting the data type Show Data Type as one of Affymetrix or cDNA and then selecting a data object an expression object created by importing expression summarization for Affy CEL and normalization from the Data drop down list box Differential Expression Analysis LPE Test E a x gt Data Variance Estimation Show Data of Type Smoother D F 10 Affymetrix zi Number of Bins 100 Data DAYSdefense Trim 5 Factor A zi r Output Options 7 Compare Level Day 0 X IV Volcano Plot Compare Level 2 Day 1 Y Axis Orientation Array Name mgu74av2 negative id Line 2 0 r Options FWER FDR 0 05 Adjustment resample X Alt Hypothe l IV Variance Plots IV Top 15 Genes m Output IV Display Qutput in S PLUS I7 Save Output as HTML Save HTML As myLPETesthtmi I Display HTML Gutput Save s myLPETest Can
236. le granularity such that each ordered individual gene expression value is aligned The method assumes there is an underlying common distribution of intensities across all chips in the set and disparate detests can be transformed to the same distribution by transforming the quantiles at the level of individual values of each to have the same value Details of this transformation can be found in the normalize quantiles help file references The draw back of this method is that extreme values in the tails are normalized to the same values thus possibly loosing the differential expression information Empirical evidence however suggests that this is not a problem see Bolstad et al 2002 215 Chapter 6 Pre Processing and Normalization Normalization using VSN VSN is short for variance stabilization normalization In the VSN normalization method the intensities from each array are calibrated by a suitable affine transformation then transformed by a variance stabilizing transformation After this systematic array or dye biases should be removed and the variance should be approximately independent of the mean intensity This is useful for subsequent analyses such as hypothesis tests ANOVA modeling clustering or classification that assume that the variance is the same for all observations Note that VSN only addresses the dependence of the variance on the mean intensity There may be other factors influencing the variance such as gen
237. led information Figure 9 3 shows an example page from LocusLink with annotation for one of the differentially expressed genes from the volcano plot Annotation from the GUI GenBank and Other Browser Metadata Lookups Annotation Libraries e gt NCBI LocusLink PubMed Entrez BLAST OMIM Map Viewer Taxonomy Structure Search LocusLink y Display Brief Organism AIl h Query J Go Clear View Hs RFC2 One of lLoci Save All Loci ABCDEFGHIJKLELMNOPOQRSTUVWX Nez fo p OO EE aS S S S S S S Click to Display mRNA Genomic Alignments spanning 22906 bps HGMD e ucsc Homo sapiens Official Gene Symbol and Name HGNC RFC2 replication factor C activator 1 2 40kDa LocusID 5982 Overview RefSeg Summary The elongation of primed DNA templates by DNA polymerase delta and epsilon requires the action of the accessory proteins proliferating cell nuclear antigen PCNA and replication factor C RFC REC also called activator 1 is a protein complex consisting of five distinct subunits of 145 40 38 37 and 36 5 kD This gene encodes the 40 kD Figure 9 3 Annotation information from LocusLink In this example we show how to launch additional metadata Web sites e g PubMed and GO using gene lists derived from statistical analyses in S ARRAYANALYZER This example also uses the Melanoma data Fox et al 2001 and picks up after the data have been read in through the GUI and analyze
238. lick to browe il ir WildType xl Swirl xl Array Layout Swit ayout z Create Array Layout m Step 3 Save Output Save Data Set As myMarrayh aw I Display Report Cancel k afi entries Figure 4 4 Selecting files for import Creating the The second part of the file association task is specifying a layout file Layout To specify the layout for a cDNA array select an S PLUS layout object that has been previously created or create a new one 131 Chapter 4 Examples Two Color Data Create a layout object for the swirl example by clicking the Create Array Layout button just beneath the file selection grid aia File Selection Options Scanner Layout Browse m Outer Grid number of grid rows and columns Grid Rows Grid Columns m Inner Grid choose columns in data file representing p Inner Grid Row X Inner Grid Col Tl Control Sl Control Value Ce Gene Name Poo Output Save Layout As myLayout Cancel K j current Help Figure 4 5 The Create Layout dialog The Create Layout dialog requires you to fill in the following information 1 Scanner Layout A layout file name Enter the path in the Scanner Layout field or click the Browse button to locate the layout file The file should be a text file with columns for the gene names and control indicator For the example navigate to the fish gal file located in y
239. lting Gene Filtering tab of the Filtering dialog is displayed in Figure 4 25 Note that the report generated by the filtering operation indicates the number of genes dropped Dropped 461 Retained 7283 Out of 7744 Genes 156 Quality Diagnostics Revisited Two Way Reference Design zix Array Filtering Gene Filtering Data Operations Column GreenForeand 4 Transform lt None gt z RedForegnd z Add Logical v Add Add m Column Values _ _ si Show Values Math lt None gt v Value r Add zi In tLeast Value hoo x Add Add m Expression Expression Flags 0 InAtLeast 1 Remove All Remove Last Values selected Keep C Drop Cancel Appt n j current Figure 4 25 Keep only the genes with zero flags After the filtering operation another QC check shows regions of the arrays that had problem or missing values i e flags that were set to non zero To see this go back to Quality Control Diagnostics gt Two Channel and run the image diagnostics again for the filtered object TPMarrayRawFiltered Select the Image Plot checkbox both for the Channel and foreground for Channel Type By choosing a Color Map which has non white low values you can see the regions that have been eliminated across the arrays Figure 4 26 displays one of the post filtering image plots The white strips in the bottom right corner of each print tip group correspond to em
240. ly checking the RMA radio button in the upper right corner of the Affymetrix Expression Summary dialog Open the dialog by clicking ArrayAnalyzer gt Affymetrix Expression Summary from the main S PLUS menu bar Then select the cgAffyBatch object in the CEL Data drop down list and select the RMA checkbox The result of the computation is an expression summary object Set the name to be cgExprSet rma by typing it into the Save As field Figure 3 12 displays the Affymetrix Expression Summary dialog ready to go RMA Output Two Sample Design 2151 Data r Options CEL Data eaAtfyBatch Summary Options GCRMA RMA Save As JoaE xprSet ma C Mix amp Match J avaditt z Normalization quantiles gt r Summary Graphics I MvA Plot I Box Plot Cancel Apply current Help Figure 3 12 Specifying Robust Multichip Analysis with a single checkbox A sequence of graphs is produced as output by the RMA procedure Figure 3 13 displays the MvA plot the expression intensity log ratio M vs the overall average intensity A for the two samples taken at two 2 hours The value in the lower left panel is the inter quartile range IQR of the values of M across all summarized expression values A small value indicates there is little difference on the logy scale for the middle 50 of the expression values for the two chips For replicate chips there is no real differential expression so the IQR is expe
241. m globalMAD gt swirl gMADS lt maNormScale swirl c 2 4 norm globalMAD d printTipMAD gt swirl ptMAD lt maNormScale swirl norm printTipMAD After individual arrays have been normalized for differences in the red and green fluorescence intensities additional normalization can be done between arrays S ARRAYANALYZER offers quantile normalization on the individual R G channels as well as on the average of the two channels A through the GUI for this purpose Normalization between arrays is needed if downstream analyses e g Normalization Methods for Two Channel Data ANOVA compare experimental conditions that vary between arrays The between array normalization options are obtained by selecting a method for the Between Array field in the GUI Refer to section Normalizing with the GUI and Chapter 4 Examples Two Color Data for more information and examples of normalizing through the GUI 231 Chapter 6 Pre Processing and Normalization PRE PROCESSING AND NORMALIZATION FOR AFFYMETRIX PROBE LEVEL DATA Affymetrix data typically arrives as DAT CEL and CHP files The DAT files contain the raw images as processed by the scanner The CEL files contain expression measures for each individual probe on the chip The CHP files contain summaries of the individual probe level data for each gene transcript This section discusses methods for analyzing correcting summarizing and normalizing the CEL probe level dat
242. ment The available methods can be obtained by typing gt pmcorrect methods i tas pmonly subtractmm mas and subtractmm are the pm correction methods performed by Affymetrix MAS 4 0 subtractmm and MAS 5 0 mas software subtractmm returns the difference between the PM and MM intensity values This can lead to negative values for the intensity pmonly returns the PM intensity values from the ProbSet PM slot mas correction allows for the possibility that the MM intensity is larger than the PM intensity for a particular probe pair within a probe set The mas method is described in the Affymetrix Statistical Algorithms Description Document SADD available from Affymetrix The input to the pm correct functions can be either a ProbeSet or AffyBatch object and the return value is a matrix of corrected PM values for each chip in the input object An object of class ProbSet contains the PM and MM data for a probe set from one or more samples ProbeSet objects can be created by applying the method probeset to instances of AffyBatch We illustrate the procedure using the example data affybatch example in the affy library data directory This data set gives a subset of the values read from a HU6800 CEL file gt pps lt probeset affybatch example geneNames affybatch example 1 2 f 1 gt pps subtractmm lt pmcorrect subtractmm pps If no subsetting is desired we can simply use the AffyBatch object in the c
243. mgu74av2 r Output Options Contrasts M Volcano Plot Factor CondTime z Y Axis Orientation IV Baseline negative z Baseline Level METATE Fold Change Line 200 I Sequential I Heat Map J Linear Quadratic gt Chromosome Plot Levels Order I Parallel Coords IV Top 15 Genes r Output IV Display Output in S PLUS T None I Save Output as HTML Apply Contrasts Across Others Save HTML As Within Others IE Display HTML Output Save As MouseSwimANOV Cancel Apply K gt f 1of 3 Help Figure 2 22 The ANOVA dialog with settings for baseline contrasts The Options group allows you to set the family wise error rate FWER or false discovery rate FDR to control the overall Type I error false positive rate based on adjusting individual test p values to account for multiple tests In our swimming mice example there are 7 624 genes so the Type I error is substantial without adjusting the p values See Chapter 7 Differential Expression Testing for more information on the error rate adjustment procedures There are many options for adjusting the p values to achieve the FWER or FDR you want Here we leave the default setting as Bonferroni There are four options in the Output Options group 1 Volcano plot 2 Heat map 3 Parallel Coords 4 Top 15 Genes One Way Design Note that chromosome plots are not available for arrays other than hgu95a Figure 2 23 displays the volca
244. mplements an approach similar to that of the MAS4 0 software This involves forming the differences PM MM for each probe pair calculating the mean and standard deviation sd of these differences removing pairs with a difference of greater than 3 standard deviations from the mean and recalculating the mean from the trimmed set The liwong method fits the model described in LI and Wong 2001a 2001b The default setting gives the current PM only default The reduced model previous default can be obtained using pmcorrect method subtractmm The mas method implements an approach similar to that of the MAS5 0 software This includes forming the differences PM MM for each probe pair and then condensing these within a probe pair set in a robust manner Outlier probe pairs are not dropped as in the avgdiff calculation they are down weighted The median of the probe pair differences within a probe pair set is calculated and each probe pair difference is down weighted as a function of its distance from the median The probe pair differences are then combined in a one step version of the Tukey biweight procedure The medianpolish algorithm works by alternately removing the row and column medians and continues until the proportional reduction in the sum of absolute residuals is less than eps or until there have been maxiter iterations In combination with the bg correct rma background correction method and the quantiles normalization metho
245. mute Student s t test with null distribution and p value estimated by permutation e wilcoxon permute Wilcoxon signed rank test with null distribution and p value estimated by permutation The method names listed in bold are used in the GUI for specifying a particular testing method All the basic methods paired t Welch s t student s t wilcoxon are described in standard introductory statistical textbooks such as Moore and McCabe 1999 or Snedecor and Cochran 1980 The permute versions of the test procedures are based on permuting the intensity scores across treatment conditions repeatedly re computing the test statistic each time to form a null distribution of the test statistic The p value is then obtained by quantifying the frequency of seeing a test statistic as extreme or more so than the one observed for the data Using permutation methods for reasonable samples sizes 10 or more per experimental condition can produce more accurate p value estimates for data which may not satisfy the assumptions of the test procedure In particular tests for skewed intensity values may benefit from computing p values by permutation rather than from the theoretical symmetrical distribution The permute versions of the tests should be used with caution for low replicate studies since the p values are based on the total number of possible test statistics for permuted data For example in a two sample study with two replicates fo
246. n Controlling The False Positive Rate for more details 264 Controlling Type I Error Rates CONTROLLING TYPE I ERROR RATES When testing for differential expression across many genes simultaneously numerous genes may be identified as significantly differentially expressed by chance alone even if there is no real differential expression For example if you test 10 000 genes for differential expression at a significance level of 0 05 you can expect to misidentify about 500 genes as significant even when there is no real difference in gene expression Multiple testing corrections adjust the individual p values to account for the inflated false positive rate due to multiple testing Because there are typically many genes represented in a microarray experiment managing the side effects of multiple statistical tests is important in differential expression testing Consequently a number of procedures have been implemented in S ARRAYANALYZER for controlling family wise error rate FWER and false discovery rate FDR Table 7 1 Errors in statistical testing Truth Significant Test Not Significant Test Differentially S FN False Negative Expressed Type II Error Not Differentially FP False Positive NS Expressed Type I Error Q Total S FP NS FN 265 Chapter 7 Differential Expression Testing Controlling The False Positive Rate Notation FWER Procedures 266 Suppose the significance lev
247. n estimation and various output options Options Other than selecting the data and factor levels for comparison the most critical inputs are in the Options group The Options group allows you to set the family wise error rate FWER or the false discovery rate FDR to control the overall Type I error rate false positive rate based on adjusting individual test p values to account for multiple tests In our melanoma example there are 12 558 genes so the increase in Type I error is substantial if you don t adjust the p values There are many options for adjusting the p values to achieve the FWER or FDR We describe them in more detail in Chapter 7 Differential Expression Testing Here we leave the default setting as Bonferroni Output Options There are five output options in the Output Options group 1 Volcano plot 2 Heat map 3 Chromosome plot 84 4 5 Two Sample Design Variance plots Top 15 genes list Setting Up the Dialog To set up the Local Pooled Error Test dialog follow these steps 1 2 In the Show Data of Type field select Affymetrix In the Data field select cgExprSet rma Select the Factor Compare Level 1 and Compare Level 2 fields to be Time 2hr and 24hr respectively The Array Name field should be automatically filled with hgu95av2 Click the drop down arrow of the Adjustment field in the Options group to set the FWER FDR procedure to Bonferroni Ensure that all the Output
248. n 370 Format Specification 371 Appendix B Importing data 385 Introduction 386 Index 389 vii Contents viii INTRODUCTION TO MICROARRAY DATA Welcome Features Goals Libraries Supported Platforms and System Requirements Installing and Running S ARRAYANALYZER Online Help Online Reference Technical Support Genomics and Differential Expression Microarray Data Affymetrix Arrays Two Color Arrays NNDDDH AA UN m A OO Chapter 1 Introduction To Microarray Data WELCOME S ARRAYANALYZER is an S PLUS module that provides you with a powerful tool for analyzing Affymetrix MAS 5 CHP and CEL data and two channel microarray data Using either the graphical user interface GUI or the Commands window you can perform statistical analysis to determine differential gene expression in microarrays fundamental to the rapidly growing field of functional genomics In S tARRAYANALYZER you can access functions in a Gene Name connective tissue growth factor Volcano Plot Probe Id 36638_at iat o LocusLink R T 2 d 2 e D oh 5 s E o gt 3 l Figure 1 1 Sample volcano plot with Affymetrix CEL data generated by the Differential Expression Analysis LPE Test dialog in S ARRAYANALYZER This plot shows genes that would be false positives and false negatives based on the fold change criteria alone but you can also quickly find significantly differentially expressed genes af
249. n be performed microarray data must be pre processed and normalized Pre processing refers to the process of correcting the measured spot intensities for background signal and non specific binding and for probe level data summarizing the multiply cloned gene expression measurements into one expression measure As described in Parmigiani et al 2003 data from microarrays are subject to many sources of extraneous variability including manufacturing preparation of mRNA from experimental samples hybridization scanning and imaging These sources of variability are often called technical sources of variability The removal and balancing of extraneous technical variability before analysis allows for more confident interpretation of the estimated differential expression effects as true differential expression and not a result of systematic experimental artifacts Pre processing of probe level data primarily involves summarizing data from probe sets into a single measure per gene transcript The Affymetrix MAS software provides a way of doing this based on a one step Tukey biweight procedure Other approaches including the MBEI method of Li and Wong 2001 and the RMA method of Irizarry et al 2003b have been shown to provide improved extraction of biological information from probe level data Irizarray et al 2002 2003b This chapter address the variety of methods available in S tARRAYANALYZER for correcting normalizing and summarizing
250. n the following page in Figure 2 38 MVA plot CGa i i i oO 2 4 6 8 10 12 14 0 4 8 10 12 14 0 839 CGb i a 1 18 1 17 CG24a NG 0 2 4 6 8 10 12 14 1 07 1 19 0 543 CG24b Figure 2 38 M versus A plots for the Melanoma experiment For each graph the vertical axis is M and the horizontal axis is A M is computed from the two arrays found by going horizontally left and vertically down to the first array name you come to For example for the upper left scatterplot just right of the CGa label M is computed as the difference in logged intensities from the CGa and CGb arrays A is the average of the same logged intensities The cone shaped MvA plot shows that variance decreases as a function of the log average expression intensity Given this pattern we will use the LPE test for differential expression since it allows variance to be modeled as a function of the average expression intensity We first have to create a couple of objects that are arguments to the LPE test function The LPE test function requires baseline variance estimates before computing test statistics We compute the baseline variance or error estimate with the baseOLIG function as follows OLIGgrpO lt baseOLIG LCG N 1 2 61 Chapter 2 Examples Affymetrix MAS Data Plotting Differential Expression Results 62 OLIGgrp24 lt baseOLIG LCG N 3 4 The required argument to base0LI1G is t
251. nd Time Rep Array label File Name Swim lwk 1 Swimlwk1 Swimlwl Swim lwk 2 Swim1lwk2 Swimlw2 Swim lwk 3 Swimlwk3 Swimlw3 Swim 2wks 1 Swim2wks1 Swim2wl Swim 2wks 2 Swim2wks2 Swim2w2 Swim 2wks 3 Swim2wks3 Swim2w3 Swim 3wks 1 Swim3wks1 Swim3wl Swim 3wks 2 Swim3wks2 Swim3w2 Swim 3wks 3 Swim3wks3 Swim3w3 Swim 4wks 1 Swim4wks1 Swim4wl Swim 4wks 2 Swim4wks2 Swim4w2 Swim 4wks 3 Swim4wks3 Swim4w3 Swim 4wks lwk 1 Swim4wks 1lwkl Swim4wlwl Swim 4wks 1wk 2 Swim4wks 1lwk2 Swim4wlw2 Swim 4wks 1wk 3 Swim4wks 1wk3 Swim4wlw3 NoSwim 10min 1 NoSwim10min1 NoSwim10min1 NoSwim 10min 2 NoSwim10min2 NoSwim10min2 NoSwim 10min 3 NoSwim10min3 NoSwim10min3 NoSwim 4wks 1 NoSwim4wks 1 NoSwim4wl NoSwim 4wks 2 NoSwim4wks2 NoSwim4w2 19 Chapter 2 Examples Affymetrix MAS Data Setting Up The Analysis 20 Table 2 1 Experimental design and file association for the melanoma cancer study Cond Time Rep Array label File Name NoSwim 4wks 3 NoSwim4wks3 NoSwim4w3 NoSwim 4wks lwk 1 NoSwim4wks 1 NoSwim4wliwl wkl NoSwim 4wkst lwk 2 NoSwim4wks 1 NoSwim4wlw2 wk2 NoSwim 4wkst lwk 3 NoSwim4wks 1 NoSwim4wl1w3 wk3 These data have been obtained from the CardioGenomics PGA Public Data Web site located at http cardiogenomics med harvard edu public data and are used here for the purpose of this example only The data are available for free public download but have also been
252. ng on a distance or similarity structure Example Lymphoma Classification Examples from the Command Line agnes Similar to hclust but the algorithm yields a measure of the amount of clustering structure found mclust Model based clustering with many options The available divisive methods are diana General divisive clustering mona Divisive clustering for only binary variables Distance Metrics All cluster methods are very sensitive to the choice of distance or dissimilarity between points i e samples or genes S PLUS includes two commonly used functions for creating distances or dissimilarities between points namely dist and daisy The correlation function cor may also be used and 1 cor x produces a matrix representing the dissimilarities between columns samples of a matrix x x The dist function simply constructs distances between rows as Euclidean Manhattan maximum and binary If the data are normalized with mean equal to zero and variance equal to one prior to calling dist the resulting matrix is equivalent to a dissimilarity matrix produced using cor In cancer diagnostics there is considerable interest in subpopulations of cancer tissue samples For example distinct subpopulations identified within the collection of samples may have different etiologies and may be candidates for different clinical interventions Alizadeh et al 2000 characterized variability in gene expression among tumor
253. ng the effects of background normalization and summarization on gene expression estimates Unpublished Manuscript Bolstad B M Irizarry R A Astrand M and Speed T P 2002 A comparison of normalization methods for high density oligonucleotide array data based on variance and bias Bioinformatics 19 2 185 193 Cleveland W S 1979 Robust locally weighted regression and smoothing scatterplots Journal of the American Statistical Association 74 368 829 836 Dudoit S and Yang Y H 2003 Bioconductor R packages for exploratory analysis and normalization of cDNA microarray data In The Analysis of Gene Expression Data Methods and Software G Parmigiani E S Garrett R A Irizarry and S L Zeger editors Springer New York Dudoit S Yang Y H Callow M J and Speed T P 2002 Statistical methods for identifying differentially expressed genes in replicated cDNA microarray experiments Statistica Sinica 12 1 111 139 3 4 Fox J W Dragulev B Fox N Mauch C and Nischt R 2001 Identification of ADAM9 in human melanoma Expression regulation by matrix and role in cell cell adhesion Proceedings of International Protelysis Society Meeting Huber W Heydebreck A Sueltmann H Poustka A Vingron M 2002 Variance stabilization applied to microarray data calibration and to the quantification of differential expression Bioinformatics 18 Suppl 1 96 S104 References Huber W Heydeb
254. nizational details such as converting all column names to lower case to make typing easier Data JHHF Change column names to all lower case Manipulation gt names cga lt casefold names cga gt names cgb lt casefold names cgb gt names cg24a lt casefold names cg24a gt names cg24b lt casefold names cg24b Now lets find the control spots All the arrays are the same so we can work off one of the data sets Extracting Probe JHHF Extract probe names Names and gt cg probes lt cga probe set name Finding Controls dHHF Find control spots gt prefix lt substring cg probes 1 4 gt controls lt prefix AFFX You can eliminate genes with few spots used in their summarization by a simple subset operation We repeat it for each array object Removing Genes JHHF Set avg diff to missing wherever pairs used lt 7 With Few Good Spots gt cga avg diff lt ifelse cga pairs used lt 7 NA cga avg diff gt cgb avg diff lt ifelse cgb pairs used lt 7 NA cgb avg diff lt same for the other two arrays gt One example exploratory plot is the comparison of control and non control spots We can generate boxplots as follows Comparing gt par mfrow c 2 2 Controls and gt boxplot list controls logb cga avg diff controls 2 Non controls noncontrols logb cga avg diff controls 2 ylab Log 2 Expression Intensities gt title 0 hr Replicate A 57 Chapter 2 Examples Affymetri
255. no plot with Bonferroni FWER correction Most of the p values for the genes have been adjusted to one Volcano Plot Swim3wks NoSwim4wks Log10 Adjusted p Value Mean Log2 Fold Change Figure 2 23 Volcano plot resulting from Bonferroni FWER correction The Bonferroni correction is very conservative and most of the p value have been pushed to one Let s try a less conservative adjustment procedure We ll use the Benjamini and Hochberg FDR procedure which maintains a small percentage of false positives amongst only those genes which are significant Select BH from the Adjustment drop down list and save the resulting object as MouseANOVANoSwim4wksBH Furthermore turn on the HTML display by checking the Save Output as HTML and Display HTML Output check boxes The resulting volcano plot is displayed in Figure 2 24 There are 21 significant genes in the plot resulting from the BH correction compared to five for the Bonferroni correction Even with a four fold increase in significant genes the BH correction maintains a low false positive rate of 5 amongst the significant genes This translates to on average only a single gene not really differentially expressed amongst those genes tagged as significant by the correction procedure 43 Chapter 2 Examples Affymetrix MAS Data Volcano Plot 44 A volcano plot displays the logarithm of p value versus fold change as shown in Figure 2 24 The vertical lines indicate f
256. normalize across arrays because we will adjust for array error a block effect in the differential expression analysis step The resulting dialog settings are displayed inFigure 4 27 x r Data Normalization Show Data of Type Normalization scalePrintT ipM x Two Channel zi Between Array none x Data TPMarayRawF ha IV MvA Plot Save As TPMarrayRawFilter I Box Plot Probe Set G PH C PM and MM When to Show Before amp After C Only After OK Cancel Apply k current Help Figure 4 27 Normalization dialog settings for the filtered malaria data Click OK or Apply to run the normalization procedure The normalization step produces the before and after boxplots of logo intensity ratios as displayed in Figure 4 28 This plot displays each array as a separate boxplot Note the alignment of the medians and some normalization of the variance across arrays even though normalization was done completely within arrays 159 Chapter 4 Examples Two Color Data 160 Before scalePrintTipMAD Normalization After scalePrintTipMAD Normalization Fme oe Je Jaee oom pasme eese E ame one e Juw eee o mmm o0 Juwa o femmeme woveme coe Jm come TO Jma y 2 oa RH y a t ngg N N fF gd NN g e N N fF d N N fc BS E N N amp g j gt E N N amp g 3 E a 3 3 au gees Wet T 3 Figure 4 28 Intensity b
257. ns _ Limit number of genes to I Genes with maximum fold change fio greater than 2 a Recalculate I Genes with Expression values Number of genes selected by filtering exceeding 10 if at least experiment chips Sort Ord z ort Order a E 4 Fadcrence Move Down OK Cancel Apply d f current Help Figure 9 5 The Filtering Options page of the Annotation dialog with the Contrast Filtering group activated and options fold change set to greater than 2 p value set to significant adjusted p lt 0 05 in this case 343 Chapter 9 Annotation and Gene List Management The Contrast Filtering group is the natural filtering choice in this case since we are working with the results of a differential expression test There are many filtering options available in the Annotation dialog Contrast Filtering e Select genes with at least one fold change greater than a user specified number e g 2 e Significant genes or selects only genes declared significant for the test and FWER FDR procedure rate chosen in the differential expression test e The combination of these two options can be chosen to match the outside regions of the volcano plot or a similarly filtered list of genes Expression Filtering e Select genes with maximum fold change greater than a user specified number e g 2 e Select genes with an expression value exceeding a user
258. ntervals in one dimension of the predictor values using iteratively weighted least squares Yang and Dudoit 2003 write that in the context of microarray experiments robust local regression allows us to capture the non linear dependence of the intensity log ratio M on the overall average intensity A while at the same time ensuring that computed normalization values are not driven to a small number of differentially expressed genes with extreme log ratios Expression data may be variable between chips not only in the median of the data but also in its spread around that median value Variability may be due to such things as scanner settings and different concentrations of mRNA across slides In order to compare expression across slides these extraneous effects must be minimized The spread of the data in some range can be scaled to be the same between groups by specifying that the data between groups match at more than one point For example we could specify that the IQR of the data be the same This requires that the data be scaled so that the spread of the middle 50 of the data is identical across the groups We can extend this idea of normalization so that the data matches at a sequence of points not just one or two points For example deciles 10th 20th 30th 90th percentiles may be aligned via this type of normalization The quantiles method in S ARRAYANALYZER extends the many point approach described above to a fine sca
259. ntrast A multiple plots are produced one for each contrast a single contrast is shown in Figure 7 18 Parallel Coords Plot Day 4 Day 0 Gene Expression Intensity j T T T T T 1 Day 0 Day 1 Day 2 Day 3 Day 4 Day 5 Day 6 Experimental Condition Figure 7 19 Parallel Coordinates Plot for a single ANOVA contrast 293 Chapter 7 Differential Expression Testing DIFFERENTIAL EXPRESSION SUMMARY TABLE OUTPUT Top I5 List Complete Gene List 294 In addition to graphical output the differential expression testing functions generate summary tables ordering the gene list from most to least differentially expressed The graphical Graphlet output provides a list of the top 15 genes but you get the complete gene list as well The 15 genes with the lowest p values are displayed in the Graphlet with the volcano plot heat map and other plots Summary Output for ANOVA Test Day 1 Day 0 with BH Adjustment Top 15 Genes Test Stal Raw p Val Adj p Val neutrophilic granule protein gt 10 0 0 immunoglobulin heavy chain 6 heavy gt 10 0 0 T cell receptor garnma variable 2 lt 10 0 0 interferon activated gene 205 gt 10 lt 0 00 lt 0 00 complement component 3 gt 10 lt 0 00 lt 0 00 CD24a antigen gt 10 lt 0 00 lt 0 00 T cell receptor garnma variable 4 lt 10 lt 0 00 lt 0 00 integrin beta 2 gt 10 lt 0 00 lt 0 00 complement component 4 within H 2 gt 10 lt 0 00 lt 0 00 lymphocyte antigen 6 c
260. object gt pm NCImelanoma lt tmp The default method for normalize is quantiles In this next example we subset the AffyBatch object by treatment 0 and 24 hours normalize each subset then merge the objects into a single normalized AffyBatch object gt mel norm quantiles lt merge normalize NCImelanoma 1 2 normalize NCImelanoma 3 4 We can normalize each replicate set to the median of one of the chips by typing gt mel norm constant lt merge normalize NCImelanoma 1 2 method constant normalize NCImelanoma 3 4 method constant Summarization Methods Pre Processing and Normalization for Affymetrix Probe Level Data Arguments to the normalization methods can be passed through normalize as optional arguments This example normalizes all four chips in the melanoma experiment gt mel norm loess lt normalize NCImelanoma method loess span 5 The corrections and normalization can be done in one step using the expresso function This function also summarizes the probe level data The resulting object is of class exprSet gt melanoma exprSet lt expresso NCImelanoma bgcorrect method mas pmcorrect method mas normalize method constant summary method mas Affymetrix and some other high density oligonucleotide arrays include multiple spots per gene transcript In Affymetrix arrays there are 11 16 or 20 probe pairs in a probe pair set with each probe pair consisting of
261. object but choose Swim4wks NoSwim4wks for the contrast Now check the Union radio button in the Output Options group confirm that the Venn Diagram check box is checked and type in SwimGeneList4wks as the name of the object to save the resulting gene list Figure 2 28 displays the finished dialog and Figure 2 29 displays the resulting Venn Diagram showing the resulting union of three gene lists Note that by selecting the Intersection radio button we can produce the intersection of gene lists with different baseline times This allows us to discover early indicator genes that express early in the heart muscle build up process and continue expressing throughout One Way Design lolx Data Group 1 Data Group 3 Show Data of Type Show Data of Type DiE xpTest lt select gt z Data MouseSwimAN x Data z Array Name mgu74av2 Array Name kundetermined gt Contrast Swim3wks N oS x Contrast 7 Cluster Mem z Cluster Mem x m Data Group 2 Output Options Show Data of Type Choose gene list creation Die xprlest v Data Mouses wim N Y C Intersection Array Name mgu74av2 Union Contrast Swim4wks NoS x Venn Diagram BlusterMem Elea SwimANOVAGene OK Cancel Apply K j current Figure 2 28 Setting up the Gene List Management dialog for merging gene lists Help MouseSwimANOVABs4wkBH MouseSwimANOVABs4wkBH Swim3wks NoSwim4wks i i Swim4wks NoSwim4wks 24 Figur
262. of Insightful Corporation ArrayAnalyzer FinMetrics NuOpt SeqTrial Wavelets and SpatialStats are trademarks of Insightful Corporation All product names mentioned herein may be trademarks or registered trademarks of their respective companies ACKNOWLEDGMENTS Insightful StARRAYANALYZER uses Bioconductor packages that represent state of the art work from a collection of leading statisticians Insightful would like to recognize these contributors affy Rafael A Irizarry Laurent Gautier and Leslie M Cope AnnBuilder Jianhua Zhang annotate Robert Gentleman Biobase Robert Gentleman and Vincent Carey edd Vincent Carey genefilter Robert Gentleman and Vincent Carey geneplotter Robert Gentleman marrayNorm Sandrine Dudoit Yee Hwa Jean Yang marrayClasses Sandrine Dudoit Yee Hwa Jean Yang marrayInput Sandrine Dudoit Yee Hwa Jean Yang marrayPlots Sandrine Dudoit Yee Hwa Jean Yang multtest Yongchao Ge Sandrine Dudoit rhdh5 Byron Ellis Robert Gentleman ROC Vincent Carey iii iv CONTENTS Acknowledgments Chapter 1 Introduction To Microarray Data Welcome Supported Platforms and System Requirements Genomics and Differential Expression Microarray Data Chapter 2 Examples Affymetrix MAS Data Affymetrix Data Analysis Workflow Experimental Design One Way Design From the Command Line References Chapter 3 Examples Affymetrix Probe Level Data Affymetrix Probe Level Data Analy
263. oids analysis summary for the Alizadeh et al 2000 lymphoma data The two clusters are projected onto a biplot of the first two principal components 327 Chapter 8 Cluster Analysis 00 02 0 4 06 08 1 0 Average silhouette width 0 19 Figure 8 15 Silhouette plot for the two subpopulation partitioning around medoids clustering for the Alizadeh et al 2000 lymphoma data 328 References REFERENCES Alizadeh A A Eisen M B Davis R E Ma C Lossos I S Rosenwald A Boldrick J C Sabet H Tran T Yu X Powell J I Yang L Marti G E Moore T Hudson T Jr Lu L Lewis D B Tibshirani R Sherlock G Chan W C Greiner T C Weisenburger D D Armitage J O Warnke R Levy R Wilson W Grever M R Byrd J C Botstein D Brown P O Staudt L M 2000 Distinct types of diffuse large B cell lymphoma identified by gene expression profiling Nature 403 503 511 Eisen M B Spellman P T Brown P O Botstein D 1998 Cluster analysis and display of genome wide expression patterns Proceedings of National Academic Sciences USA 95 25 14863 14868 Kaufman L Rousseeuw P J 1990 Finding Groups in Data An Introduction to Cluster Analysis John Wiley amp Sons New York Kerr M K Churchill G A 2001 Bootstrapping cluster analysis Assessing the reliability of conclusions from microarray experiments Proceedings of National Academic Sciences USA 98 8961 8965 Ross D T Scherf U Eisen M B
264. old change values of plus or minus one and the horizontal line indicates a significant test p value after doing the Benjamini Hochberg correction Points located in the upper outer sextants are those with large absolute fold change and small significant p value Each of those points is active so you can click an individual point to access annotation information from Locus Link or GenBank in Out Rece Fl Gene Name angiotensinogen Volcano Plot Swim3wks NoSwim4wks Probe it 101887_at T 2 D D D 3 T E f gt o a Mean Log2 Fold Change fa fm hdwks Summary Swim4wks NoSwimdwks Summary Swim4wks 1wk NoSwim4wks J Summary NoSwimdwks Twk NoSwimdwks Volcano Swim3wks NoSwimdwks Figure 2 24 A volcano plot which is the logarithm of p value versus fold change The Benjamini Hochberg FDR correction method was used for the swimming mouse data at an FDR of 0 05 The BH correction is less conservative than the Bonferroni procedure yet maintains a small proportion of false positives amongst those genes tagged as significant One Way Design Heat Map A heat map plot shown in Figure 2 25 shows a 2 D plot of the most differentially expressed genes along the vertical axis versus the experimental conditions on the horizontal axis This graph is also hyperlinked to the annotation information Sample Swimdwks lwk3 Gene NA Probe Id 104477_at Exp Value 6 39 Fig
265. om differentiation initiated Aa by addition of RA B D T Promyelocyte Myelocyte Metamyelocyte Band neutrophil Neutrophil 0 1 2 3 4 5 6 days Figure 9 23 Study of granulocyte differentiation In studies such as this there may be a priori interest in a particular pathway or cell response ontology For illustrative purposes only we choose to focus our attention on the defense response Biological Process ontology for our filtering and hypothesis testing exercise We may wish to view the GO hierarchy for defense response e g by using the NetAffx GO browser on the Affymetrix web site This provides a contextual view of the biological processes within this GO node For more information see Figure 9 24 below Annotation Libraries 19884 antigen presentation 19882 antigen presentation 1 exogenous antigen 1 3 7 of 27 6 2 of 16 16066 cellular defense response 6968 cellular defense response 3 sensu Vertebrata 3 10 3 of 29 20 0 of 15 16064 humoral defense mechanism 6959 humoral immune response 1 sensu Vertebrata 1 8955 immune response 11 eS q 2 6 of 38 3 1 of 347 50776 regulation of immune i response 1 50778 positive regulation of 6952 defense response 12 12 5 of 8 Akdi ne ee 1 2 7 of 432 6953 acute phase response 1 20 0 of 5 42742 defense response to bacteria 1 3 7 of 27 45088 regulation of innate 16 6 of 6 immune response 1 45087 innate immune r
266. om 12 488 on the mgu74av2 chip to 540 This allows more focused differential expression testing Note that we are filtering only the exprs slot i e the actual expression data All other slots remain untouched There are several simple ways to do this In the above we start with the full data set DAYSrmaExprSet This data set represents all 7 time points days and 4 chips per time point with the 28 cel files summarized using RMA We make a copy of this object and then subset the exprs slot using the Affy IDs we obtained in Steps 1 and 2 The resulting object DAYSdefense includes only the 540 genes in the defense response branch of the GO Biological Process ontology The new filtered object is available through the S ARRAYANALYZER GUI for analysis e g differential expression and clustering A volcano plot for one of the contrasts is shown in Figure 9 25 Annotation Libraries Gene Name bone marrow stromal cell antigen 1 Volcano Plot Day 6 Day 0 Probe Id 103789_at F OCOD OFDDoIO O o A Number LocusLink T 3 d e 5 5 5 gt 5 l ra D E 2 Mean Log2 Fold Change 4 Volcano Day 2 Day 0 Volcano Day 3 Day 0 Volcano Day 4 Day 0 Volcano Day 5 Day 0 Volcano Day 6 Day 0 Parallel Day 1 Day 0 Paralld Gene List Significant Genes Figure 9 25 HTML volcano plot for the dayO versus day6 baseline contrast for the DAYSdefense object filtered by the GO Biological Process branch
267. omplex locu gt 10 lt 0 00 lt 0 00 cathelicidin antimicrobial peptide gt 10 lt 0 00 lt 0 00 neutrophil cytosolic factor 1 gt 10 lt 0 00 lt 0 00 immunoglobulin heavy chain 6 heavy gt 10 lt 0 00 lt 0 00 proteoglycan 2 bone marrow gt 10 lt 0 00 lt 0 00 complement component 1 q subcompo 9 65 lt 0 00 lt 0 00 Figure 7 20 75 most differentially expressed genes The complete gene list is saved in an S PLUS object For more details see section Output in section Two Sample Dialog Input and in section LPE Testing Dialog Input in this chapter You can access the gene list in three different ways 1 The Data gt Select Data menu item on the main S PLUS menu bar 2 Through the S PLUS Object Explorer From The Data Menu Item From The Object Explorer Differential Expression Summary Table Output 3 The Command line Open the Select Data dialog by selecting Data gt Select Data from the main S PLUS menu bar and select the test summary objects from the Existing Data Name drop down list Select Data Source Existing Data Existing Data Name Imeem C New Data i C Import File New Data PET T Show Dialog on Startup Cancel Apply KE current Figure 7 21 Selecting the complete gene list from the Select Data dialog Clicking OK opens a data sheet containing the summary information a Multlestsumm Read Only 1 2 3
268. on In this section we step through the analysis of an experiment using an MM5 melanoma cell line in which a gel matrix that simulates the in vivo cellular condition and progression of melanoma was added for 0 and 24 hours later Fox et al 2001 This simple experimental design thus involved one factor matrix condition at two levels 0 and 24 hours replicated twice at each time point The main hypothesis of interest involves discovering genes showing differential expression at the two time points because these genes are believed to be relevant to tumor invasion and metastasis The chips and data files are in Table 2 3 Table 2 3 Experimental design and file association for the melanoma cancer study Experimental Condition Repetition chip label File Name 0 hours 1 cga OhA csv 0 hours 2 cgb OhB csv 24 hours 1 cg24a 24hA csv 24 hours 2 cg24b 24hB csv 54 Importing Data General GUI Import Command Line Import From the Command Line S PLUs has several command line functions for importing data as well as a very general facility for importing data through the GUI It is worth spending a little time importing data through the S PLUS GUI because the facility is quite general and easy to use To import a data file though the GUI go to File gt Import Data gt From File When the dialog opens select the File Format and then browse for files Figure 2 34 shows the Data Specs page of the S PLUS
269. oni I GO Website SS Use Affymetrix IDs pe I Save Affy IDs to File Affymetrix ID File A F Open Atfymetriy GO Browser ProbeList txt F Open DAVID EASE Browser Cancel Apply j current Figure 9 16 The General Options page of the Annotation dialog Options chosen Use LocuLink IDs group This writes out a file of LocusLink IDs LocusLinkList txt by default corresponding to the genes selected according to the Annotation dialog options Annotation Using OntoExpress Annotation Libraries SOURCE Batch Search Microsoft Internet Explorer A loj x File Edit view Favorites Tools Help Bak gt A search GaFavorites media C4 B S H Address a http genome www5 stanford edu cgi bin source sourceBatchSearch Go Links kod ad This page is the batch extract interface for SOURCE You can input a list of GenBank Accessions dbEST cloneIDs UniGene ClusterIDs UniGene gene names or UniGene gene symbols and retrieve data from the check list below You will be given a link to the output file when processing is complete You may enter the identifiers as a file or as a list in the text area below If you are entering an input file it must be a text file consisting of a single column containing one of the accepted types of identifiers Please see the help section for further information 1 Input the list of identifiers Input File Browse Or enter a list of
270. ons Filtering Options m Contrast Filteing gt r Gene List Filtering Data on which to Filter Data on which to Filter MouseSwimAN v GeneListLPEAr x Contrast Swimdwks N oS x I Filter on Gene List I Genes with fold change r Cluster Filtering greater than A Data on which to Filter myCluster x I Filter on Cluster Summary m Expression Filtering Data on which to Filter Cluster idan fi z Mouses wimE xp r Gene Sort Order Options Limit number of genes to 4 I Significant genes J Genes with maximum fold change I Genes with Expression values Number of genes selected by filtering exceeding 10 15 in at least experiment chips Ee y Sort Order a Move Up Move Down OK Cancel Appt K o current Help Figure 2 32 Filtering Options settings for annotation of the genes identified by the MouseSwimANOVABs4wksBH ANOVA Clicking OK now will generate the annotation lists from the databases selected Figure 2 33 displays a summary table for the top significant genes identified by the ANOVA tests for the Swim3wks NoSwim4wks contrast About Entrez Help FAQ Entrez Tools ion hi LinkOut pr Search for Genes Submit tao GenBank for full length PubMed Mi 2 3 C4 Cs One Way Design eo Sh Nucleotide Nucleotide Protein Genome Structure Search Nucleotide for L29454
271. or eliminating rows with non zero flags Flags 0 InAtLeast 1 Here are the steps to create it 1 Select a data column Flags and click the Add button in the Data group to add the name Flags to the Expression field at the bottom of the dialog Select the symbol double equal sign from the Logical field of the Operations group and click the Add button right under Logical label to add the symbol to the Expression field at the bottom of the dialog Select the value 0 zero from the list in the Column Values group and click the Add button in that group to add 0 zero to the Expression field at the bottom of the dialog Select 1 one in the InAtLeast Value field of the Operations group and click the Add button right under the InAtLeast Value label to add InAtLeast 1 to the Expression field at the bottom of the dialog Ensure that the Values Selected in the Expression group is set to Keep Click OK to eliminate all genes except those with Flags 0 Note The InAtLeast 1 operation allows you to keep any gene that has a Flags value of zero on one or more arrays Without this operation a gene is eliminated unless the Flags value is zero on all arrays This operation works similarly but in reverse when you are dropping genes rather than keeping them When dropping genes without the InAtLeast1 operation genes will only be dropped when the value is equal to the zero for this case in all samples
272. ore and after normalization Chip Specific Detailed information about these plots for the different chips types are Plots available in the following sections The sections that follow will also describe other types of plots that are available from the command line 219 Chapter 6 Pre Processing and Normalization NORMALIZATION METHODS FOR TWO CHANNEL DATA Normalizing with the GUI 220 There are often systemic variation and imbalances of the red and green fluorescence intensities in two channel data This variation is usually not constant across the spots within or between arrays and can vary according to overall spot intensity location on the array plate origin and possibly other variables Some causes of the imbalances may be the following e Labeling efficiencies and scanning properties of the Cy3 and Cy5 dyes e Amounts of Cy3 and Cy5 labeled mRNA e Scanning parameters such as PMT settings e Print tip spatial and plate effects The GUI performs default setting normalization for a batch of arrays For two channel arrays the chips need to first be normalized to balance differences in the red and green channels on each chip After that if desired the set of arrays can be normalized to clean up undesirable between array variations The GUI includes the within array methods listed in Table 6 3 and Table 6 4 The GUI also allows for between array normalization via quantiles after within normalization is completed N
273. ormalization between arrays is needed if downstream analyses e g ANOVA compare experimental conditions that vary between arrays These options are obtained by selecting a method for the Between Array field shown in Figure 6 3 Notes For Command Line Users Two Channel Diagnostic Plots Normalization Methods for Two Channel Data Please refer to section Normalization on page 210 for examples showing how to use the GUI to produce diagnostic plots and normalize two channel data Normalization lolx m Data r Normalization Show Data of Type Normalization median f Two Channel zl Between Array none x Data myFiltering z none T MvA Plot quantiles on red green Save As myFiltering norm F Box Plot Probe Set Fi PM and titi quantiles on vsn Before amp After C Only After Cancel Apply current Help Figure 6 3 Normalization dialog for two channel data where an optional between array method is chosen in addition to within array normalization When to Show The normalization and plotting functions for two channel data make heavy use of the accessor methods for the different marray classes The input parameters to the functions are labeled x y and z Each function uses the x y and z parameters differently refer to the help files for specifics In general these parameters give the accessor methods for the marrayRaw or marrayNorm class objects These ac
274. orrection procedure gt pmCor mas lt pmcorrect mas affybatch example We can replace the original PM values with the corrected PM values by typing gt affybatch example tmp lt affybatch example gt pm affybatch example tmp lt pmCor mas Normalization normalize Function Pre Processing and Normalization for Affymetrix Probe Level Data Like two channel arrays the spot intensities on Affymetrix arrays include variations due to sample preparation manufacturing of the arrays and array processing labeling hybridization and scanning Many researchers have pointed out the need for normalizing Affymetrix arrays See for example Bolstad et al 2002 and Irizarry et al 2003a S ARRAYANALYZER provides a variety of normalization methods for cell level data Location normalization methods constant e contrasts invariantset e loess Scale normalization methods e qspline e quantiles e quantiles robust e vsn The main function for normalizing AffyBatch objects is normalize The normalize function accepts AffyBatch objects and returns AffyBatch objects AffyBatch objects store the experimental information about the probe level data Please refer to the affy library documentation splus62 library affy affy pdf or the AffyBatch class and exprSet class help files for more details The normalize function is a generic wrapper which calls the normalize AffyBatch method functions These functions extract the int
275. ot can be seen in section Affymetrix Summarized Diagnostic plots on page 234 Other plots such as histograms and Affymetrix qgplots can be obtained for any summarized data data of class Data exprSet by extracting the exprs slot values from the exprSet object We demonstrate this using the previously medi an IQR normalized data The expression values are extracted and the log transform is taken before the box plot is created gt par mfcol c 1 2 two box plots on one page gt boxplot data frame log2 Dilution exprSet exprs ylim c 0 15 style bxp att gt boxplot data frame log2 cbind DilutionEsetNormTmtl DilutionEsetNormTmt2 style bxp att ylim c 0 15 254 Normalization Methods for Affymetrix MAS Data From the plots in Figure 6 7 we can see that after normalization there are differences in the average expression levels between the two treatment groups Intensity Distribution Before Normalization Intensity Distribution After Normalization wo as 15 10 10 X20A X20B X10A X10B X20A X20B X10A X10B Figure 6 7 Before and after normalization box plots of the summarized Dilution dataset 255 Chapter 6 Pre Processing and Normalization REFERENCES 256 Affymetrix 2002 Statistical Algorithm Description Document Affymetrix Santa Clara CA Affymetrix 2001 Affymetrix MicroArray Suite Version 5 0 User s Guide Santa Clara CA Bolstad B M 2002a Compari
276. our splus62 modules Array Analyzer examples directory 2 Outer Grid The print tip group grid size in rows and columns The schematic in Figure 4 6 represents a cDNA array The spots are arranged in large blocks or print tip groups Within each print tip group are spots where the cDNA is fixed To specify the array layout you must specify the size and arrangement of both the print tip groups the outer grid 132 Two Sample Design and the spot matrix the inner grid within each group It is assumed that the spot matrix size is the same for each print tip group Figure 4 6 Schematic of a cDNA array layout The outer grid refers to the layout of the print tip groups You specify this layout in terms of rows and columns In the schematic in Figure 4 6 there are two rows and three columns of print tip groups In the swirl example there are four rows and four columns Enter 4 for both the Grid Rows and Grid Columns in the Outer Grid group Single Large Print Tip Group Note that some arrays have a single large print tip group covering the entire array In this special case the outer grid has just one row and one column 3 Inner Grid The inner grid is specified by selecting columns that correspond to row and column identifiers in the layout file The appropriate column names may be detected automatically and filled in for you If not select the appropriate column name for each of the Inner Grid Row and Inner Gr
277. oxplots before and after normalization You can check the result of the normalization step by going back to the Quality Control Diagnostics for Two Channel dialog and re running the print tip group boxplots for the normalized data in TPMarrayRawFiltered norm Figure 4 29 displays the print tip group boxplots for the same array as in Figure 4 23 Note the removal of the trend we see in the un normalized plot and the alignment of medians and shoulders of the boxes The scalePrintTipMAD method is effective in removing the systematic sampling effects which produced the print tip group biases seen in Figure 4 23 Clustering Expression Two Way Reference Design Print Tip Intensity Box Plot for TP_O1a gpr 1 2 1 3 1 4 2 1 2 2 2 3 2 4 3 1 3 2 3 3 3 4 4 1 4 2 4 3 4 4 1 1 PrintTip Figure 4 29 Boxplots by print tip group after scalePrintTipMAD normalization The normalization has been effective in removing the systematic sampling effects seen in the non normalized data The analysis of Bozdech et al 2003 reveals changing gene expression over the course of the study Groups of genes expressing early in the study no longer do at the end Recall that times ranged from 1 to 48 hours in the original study Our example data set ranges from 1 to 31 hours With a simple hierarchical clustering of expression intensities we can examine the change in expression over time Open the
278. oxplots of logged expression summaries for each array Visual inspection shows the distributions are well aligned at their centers and quartiles Although normalization may be repeated sequentially to summarized expression intensities there is little need to apply more normalization to SurgeryExprSet rma Note The values displayed in the MvA plots in Figure 3 37 depend on the values used for Random seed The samples change from plot to plot so you may see slightly different plots as a result Log of After applying normalization and summarization procedures to the Expression raw expression intensities a log base 2 transformation is applied Intensities Consequently the returned summarized object contains expression intensities on a log scale The log transformation is computed as logo E ifE gt 1 0 if E less than or equal to 1 108 Old Ohr MvA Plots Two Way Design OldOhr1 2 4 6 8 0 12 14 e s o a a 0 205 OldOhr2 Z 6 8 0 12 14 0 273 0 24 OldOhr3 A Figure 3 37 MvA plot for the three replicate samples of old mice measured at O hours The values in the lower left triangle of the plot are the interquartile ranges of M Expression Summaries Intensity 8 1 oe uee A ILLI l J tee J Ieee ma es ef c L
279. p J Linear Quadratic I Chromosome Plot Levels Order IV Parallel Coords IV Top 15 Genes T Output IV Display Output in S PLUS I None J Save Output as HTML Apply Contrasts Across Others G Within Others my4NOVA html F Display HTML Output Save As myANOVa Cancel ate i j current Help Figure 7 10 The ANOVA test dialog Once a data object is selected the chip name is filled in the Chip Name field For custom 2 channel or non Affymetrix oligonucleotide chips the chip name may be lt undetermined gt The Contrasts group contains settings for selecting the types of comparisons done between the levels of the factors The selected contrasts are orthogonal and may be considered as independent sources of information extracted from the experimental data The contrast Factor is chosen first There are 3 choices for the contrasts on this chosen factor Baseline Sequential Linear Quadratic and None 281 Chapter 7 Differential Expression Testing Options 282 The Baseline contrast setting compares the levels of the chosen Factor to the Baseline Level selected This is commonly used in a time course or concentration series experiment in which levels of a Factor representing the different time points are contrasted against the Baseline Level as the initial time point or any other reference time point For example if a factor time had levels t0 t1 t2 t3 an
280. plorer E O x File Edit view Favorites Tools Help tck gt OA A Asearch fairavortes meda G D 3 M A Address http www ncbi nih gov entrez query fegi tool bioconductor amp cmd Search amp db Nucleotide8term X789479 Y Go Links gt gt NCBI Entrez PubMed Search Jj Nucleotide ci for x78947 AF050110 Y11307 AB023206 Limits Preview Index History Clipboard Details About Entrez Display Summary 7 Show 20 S Send to Tex Items 1 4 of 4 One page Help FAQ 1 X78947 Seem H sapiens mRNA for connective tissue growth factor te gil474933 emb X78947 1 HSCTGE 474933 Links M2 AF050110 Links Homo sapiens TGFb inducible early protein and early growth response protein alpha genes complete cds gi 3523 144 gb AF050110 1 AF050110 3523144 3 Y11307 H sapiens CYR61 mRNA gil2791897 emb 11307 1 HSCYR61 2791897 I 4 AB023206 Homo sapiens Amotl2 mRNA for angiomotin like 2 partial cds other name KIAA0989 Figure 9 21 Accession number lookup in Entrez for the four genes identified in the gene filtering analysis described above 362 Filtering Genes Based on GO Categories Annotation Libraries Z Entrez PubMed Microsoft Internet Explorer File Edit Yiew Favorites Tools Help end gt BB A Asah Graos gme A Eh S M A Address http www ncbi nih govjentrez query fcgi tool bioconductor amp cmd Retrieve amp db PubMed dist_uids 12831 7
281. pots on the array When your data files are organized this way you should be able to read the data through the GUI by selecting one of the data files as the layout file for the Create Layout dialog and then reusing that same file on the File Selection page of the Import Data From Two Channel dialog The Create Layout dialog will only pick up the layout probe names and control information When you use the file the second time on the File Selection page the import operation will pick up the expression intensity columns Some arrayers don t have a double grid layout as we describe in the swirl example Agilent with its Inkjet technology for printing arrays produces one large spot matrix In this case set the Grid Rows and Grid Columns to one on the Create Layout dialog Normalization is designed to remove artifacts and systematic variation resulting from the preparation and measurement process The goal is to remove variability not due to differential expression so that differential expression is estimated accurately for each gene Note that we need to be careful not to normalize so aggressively as to wash out signal For cDNA data normalization corrects for various types of dye bias as well as print tip and substratum irregularities Some examples include the following 1 Different labeling efficiencies and scanning properties of the Cy3 and Cy5 dyes 2 Print tip effects Spatial within slide effects 4 Between slide effects T
282. project folder when you start S PLUS your cmd directory is the default working directory gt getenv S_ WORK D Program Files Insightful splus62 cmd You should see two HTML files in your working directory when S PLUs has finished generating the output one for the summary table and the another for the Graphlet 279 Chapter 7 Differential Expression Testing GUI FOR ANOVA TESTING ANOVA Testing The dialog for ANOVA testing is displayed in Figure 9 10 Open the Di alog Input dialog from the main S PLUS menu by clicking ArrayAnalyzer gt Differential Expression Analysis gt ANOVA The dialog is arranged in five main groups Data Contrasts Options Output Options Output GE eee ae IEP 280 Data Contrasts GUI for ANOVA Testing The Data group allows you to select the expression object for testing You start by selecting the data type Show Data Type as one of Affymetrix or cDNA and then selecting a data object created through import probe level summary for Affy CEL and normalization from the Data drop down list box Differential Expression Analysis ANOVA oj x m Data r Options Show Data of Type PWER FDR 0 05 I Protected Data pavs defense a Adjustment Bonferroni x Array Name mgu74av2 Output Options Contrasts Volcano Plot Factor Ja zl Y Axis Orientation M Baseline negative gt Baseline Level Joyo x Fold Change Line ko I Sequential T Heat Ma
283. pter 9 Annotation and Gene List Management ANNOTATION AND GENE LIST MANAGEMENT FUNCTIONALITY 332 S ARRAYANALYZER primarily uses the annotation metadata maintained by the Bioconductor project repackaged as S PLUS libraries and data objects for fast lookup and display Bioconductor maintains Affymetrix chip specific and general annotation data packages these data packages and the process for creating them are described in detail below The Affymetrix chip specific and general annotation data libraries are used by StARRAYANALYZER to do the following e Annotate graphical and tabular reports from statistical analyses using gene lookup metadata sites such as LocusLink and Entrez e Annotate gene lists derived from the statistical analyses via metadata repositories such as LocusLink Entrez Pubmed AmiGO and Source e Connect to gene list analysis sites such as Onto Express and DAVID EASE and initiate gene list analyses e g gene function enrichment and identification of GO categories that are overrepresented in gene lists derived from statistical analyses e Subset microarray datasets according to GO categories prior to differential expression analysis The annotation libraries contain S PLUS objects lists for each metadata type typically with names corresponding to probe ids for the probes on the array and entries corresponding to the metadata mappings The probe ids on an array are typically unique for any manufacturer
284. pty spots that occurred across all arrays 157 Chapter 4 Examples Two Color Data Green Foreground for TP_01a gpr 61000 54000 48000 41000 34000 27000 20000 14000 6900 91 63000 56000 49000 42000 35000 28000 21000 14000 7000 51 Figure 4 26 Image plots of both foreground channels after filtering Note the white strips at the bottom right of each of the print tip groups which were non zero flags for EMPTY spots Normalization Now it s time to normalize the arrays The boxplots in Figure 4 23 clearly show a dependency of expression intensity on print tip group In fact there is even an indication of trending with increasing column number within any print tip row This indicates a need to apply normalization within the print tip groups We can accomplish this with one of the print tip normalization methods available from the Normalization dialog 158 Two Way Reference Design Click ArrayAnalyzer Normalization from the main S PLUS menu bar In the resulting dialog select Two Channel data in the Show Data of Type field and select the TPMarrayRawFiltered object for normalization Note that the Save As field is automatically filled with the object name TPMarrayRawFiltered norm Now choose the scalePrintTipMAD method in the Normalization field of the Normalization group on the right side of the dialog This method effectively equates the median and variance of each print tip group within each array We won t
285. r 3 Examples Affymetrix Probe Level Data Step 3 Saving the Data Object 74 File Type Note that the File Type Probe Level CEL listed below the files is automatically detected once a file is selected The dialog is designed to prohibit mixing file types Array Type The Array Type is automatically detected for probe level data For this experiment it is HG_U95Av2 or hgu95av2 as it is listed in the Array Name drop down list select hgu95av2 and continue to the next step but do not click OK until the following steps have been completed To save the data object type a name in the Save As field near the bottom of the dialog Step 3 Save Output Remember this name as it is used in the other analysis steps such as quality checks filtering and normalization For our example enter cgAffyBatch as the object name The Display Report checkbox indicates whether or not to print summary information into an S PLUS report window Step 3 Save Output Save Data Set As oodttyB atch IV Display Report Figure 3 6 Saving the imported data as cgAffyBatch Saving the Design Once you ve entered all the information on this tab you can save it for later use by clicking the Save Design button at the top of the dialog A txt file is written to the directory of your choice with the number of arrays number of factors number of levels and the full path file names and their associated factor levels Reading Designs This design
286. ray Data All Other Locations Contact the UK office of Insightful Corporation 5th Floor Network House Basing View Basingstoke Hampshire RG21 4HG Telephone 44 0 1256 339800 Fax 44 0 1256 339839 E mail shelp insightful com If you purchased StARRAYANALYZER through our international distributor network contact your local distributor http www insightful com contactus internationaldistributor asp Genomics and Differential Expression GENOMICS AND DIFFERENTIAL EXPRESSION DNA microarrays are the most widely used tools in the analysis of gene expression and the study of functional genomics Microarrays comprise gene specific sequences probes immobilized to a solid state matrix which are queried with mRNA from biological samples under study Since many changes in cells are related to changes in mRNA levels for some genes microarrays can be effectively used in a wide variety of applications including identification and validation of drug targets characterization and screening of drug toxicities exploration of biological pathways and development of molecular diagnostics INSIGHTFUL StARRAYANALYZER WORKFLOW Access Data REIES Prepare ee del g 7 Deleon ae l Probe level Summarization Figure 1 2 Once data are obtained from a microarray experiment several steps are required to prepare and analyze differential expression intensities and annotate the results with gene descriptions available in public data
287. rchical clustering of expression values Standardization The Standardization options in the Response Variable group allows three options for standardizing the data before clustering 1 None raw expression intensities are used 2 Median polish 3 Standard values Values are standardized by subtracting off the mean and dividing by the standard deviation 312 Examples from the GUI EE General Options Filtering Options m Contrast Filteing gt r Gene List Filtering Data on which to Filter Data on which to Filter MouseSwimAN x GeneListLPEAr 7 Contrast NoSwimdwks x I Filter on Gene List I Genes with fold change r Gene Sort Order Options greater than eg Limit number of genes to 200 T Significant genes Recalculate m Expression Filtering Data on which to Filter Number of genes selected by filtering MouseSwimExp Y 51 IV Genes with maximum fold change Sort Order greater than je i I Genes with Expression values Move Up Move Down exceeding 10 in at least experiments chips OK Cancel Appt K j current Help Figure 8 4 Filtering Options ready for simple hierarchical clustering of expression values with maximum fold change greater than 8 The resulting output is displayed in Figure 8 5 The clustering dendrogram for expression values is displayed on the left side of the heat map gene labels are on the right side The dendrogram at the top of the page
288. re 7 8 Setting the p value adjustment procedure for controlling the FWER The Variance Estimation group in the upper right hand corner of the dialog controls optional settings for the LPE estimator The options are Smoother D F The degrees of freedom used by the spline smoother to estimate the baseline variance function for each group Default is 10 277 Chapter 7 Differential Expression Testing Number of Bins Number of bins to compute variance estimates These variance estimates along with an associated average expression intensity is the data used by the loess smoother to estimate the baseline variance function Trim Percent of pooled variances to trim from the low end of expression intensity prior to running the loess smoother Output Options The Output Options group is a list of check boxes for selecting which graphs you want as output The options are described in detail r Output Options IV Volcano Plot Y Axis Orientation negative foa Fold Change Line 2 0 Mi IV Variance Plots I Top 15 Genes Figure 7 9 The Graph Options group in the LPE test dialog in the section Differential Expression Analysis Plots Output The Output group controls where the graphs are displayed and the gene list table is saved after the testing step is complete Display Output in S PLUS Displays the selected graphics in an S PLUS graphic device e Save Output as HTML Saves the S PLUS graphlet with
289. reck A Sueltmann H Poustka A Vingron M 2003 Parameter estimation for the calibration and variance stabilization of microarray data Statistical Applications in Genetics and Molecular Biology Vol 2 No 1 Article 3 Irizarry R A Gautier L and Cope L P 2003a Analysis of Affymetrix Prove level Data In The Analysis of Gene Expression Data Methods and Software Edited by G Parmigiani E S Garrett R A Irizarry and S L Zeger Published by Springer Verlag New York Irizarry R A Hobbs B Collin F Beazer Barclay Y D Antonellis K J Scherf U Speed T P 2003b Exploration Normalization and Summaries of High Density Oligonucleotide Array Probe Level Data Accepted for publication in Biostatistics Irizarry R A et al 2003 Exploration normalization and summaries of high density oligonucleotide array probe level data Biostatistics 4 249 264 Irizarry R A Bolstad B M Collin F Cope L M Hobbs B and Speed T P 2002 Summaries of Affymetrix GeneChip Probe Level Data Nucleic Acids Research Vol 31 No 4 e15 Lazaridis E Sinibaldi D Bloom G Mane S and Jove R 2002 A simple method to improve probe set estimates from oligonucleotide arrays Mathematical Biosciences Volume 176 1 53 58 Lee J K and O Connell M 2003 An S PLUS library for the analysis of differential expression In The Analysis of Gene Expression Data Methods and Software Edited by
290. red for viewing in TreeView Eisen et al 1998 Note that this is not the actual raw data We treat it as raw data to show the cluster methods in S PLUS but the resulting output should not be directly compared with fig3a of Alizadeh et al 2000 Hierarchical Clustering Example Our hierarchical cluster method hclust uses a group average between cluster dissimilarity measure The partitioning method uses the partitioning around medoids method pam We begin by importing the data described above The data can be downloaded from http Ilmpp nih gov lymphoma data shtml From the Figure 3 link download the file named figure3a cdt into the splus62 module ArrayAnalyzer examples directory The first two lines in the following code will import the data and create a data frame which we call mat3a Examples from the Command Line ETIE ElEk ER ew pt Fynt Qaa gaas Gach Options widow Heb a 3 RETF Os OS SHR oe oog Dg Rl i bi fnear J bhi a e Aweza 43h MWEN BG EB No Active Link 1 2 3 4 5 6 ci 8 9 10 u R mecu conz orcos cct 0003 eoLoces DLCLO023 DLCL 0015 OLOL0010 OLCLO030 waw S OLCLODIA 141 0092 1 GES 0 40 0 46 1 74 oes EXE 0 89 007 0 11 0 46 Q o2 2 Gila oa oe 07 154 oe oN aa os on asi o oa 3 GEN 0 23 0 33 Osi 0 48 O29 OSS 0 17 ao 004 0 01 0 29 0 28 4 GREK 029 0 01 008 az 23 0 18 KE Des 00 NA EE 0 15 5 ex 055 ao oa os Is LN ia o 212 ma Oa 6a GENEIOX 0 29 0 60 0 15 009 oo
291. res for doing two sample differential expression analysis They include the t test with or without assuming equal variance for the groups the Wilcoxon rank sum test and several permutations tests In addition the Local Pooled Error LPE test procedure which produces improved error estimates when there is little replication in the design is implemented for two sample problems All these procedures are suitable for doing simple comparisons between two groups treatment versus control tissue 1 versus tissue 2 etc For more details about two sample procedures see Chapter 7 Differential Expression Testing In this section we examine two color microarray data from a developmental biology experiment The data are included with the Bioconductor distribution and were originally provided by Katrin Wuennenberg Stapleton from the Ngai Lab at UC Berkeley The experiment was designed to study the early development of vertebrates using zebrafish as a model organism Zebrafish embryos from two genetic strains were used a swirl mutant and a normal wild type The goal was to identify genes with differential expression between the two strains Refer to the swir1 help file for more details The experiment consisted of two sets of dye swap experiments resulting in a total of four arrays Each pair of experiments swapped the color labels between the swirl and wild type samples Table 4 1 details the experimental conditions and the associated data files
292. res the resulting silhouette plots One can then select the number of clusters yielding the highest average silhouette width If the highest average silhouette width is small e g below 0 2 one may conclude that no substantial structure has been found 309 Chapter 8 Cluster Analysis EXAMPLES FROM THE GUI To access the clustering methods available through the S ARRAYANALYZER GUI click ArrayAnalyzer Cluster Analysis from the main S PLUS menu bar ArrayAnalyzer Import Data gt Quality Control Diagnostics gt Filtering Affymetrix Expression Summary Normalization Differential Expression Analysis gt Annotation Gene List Management Figure 8 1 Opening the Cluster Analysis dialog from the S PLUS menu bar Basic Dialog The resulting dialog has two tabs Description 1 General Options for data and algorithm selection 2 Filtering Options for subsetting the data before clustering The General Options tab is organized into five groups 1 Data to select the data set of interest Response Variable to select the clustering variable Hierarchical Methods to select hierarchical clustering and options 4 Partitioning Methods to select a partitioning methods and options 5 Output display options The Filtering Options tab has four groups 1 Contrast Filtering to filter on a contrast from an ANOVA test object 2 Expression Filtering to filter on expression values Gene List Filtering to filter on
293. rest value not beyond 1 5 IQR from the quartiles where IQR is the inter quartile range i e the difference between the 75 and 25 percentiles Figure 5 6 displays an example Intensity Boxplot Intensity Box Plot ELLELE LL ALLELE 15 10 T T I l Intensity 5 1 __ ce owe oad oo ame ont 0 1 Swim3wks1 Swim3wks2 Swim3wks3 Swim4wks1 Swim4wks2 Swim4wks3 Swim4wks 1wk1 Swim4wks 1wk2 Swim4wks 1wk3 NoSwim4wks1 NoSwim4wks2 NoSwim4wks3 NoSwim4wks 1wk1 NoSwim4wks 1wk2 NoSwim4wks 1wk3 Figure 5 6 Example Intensity Boxplot for comparing distributions of expression intensities 198 Quality Control Diagnostics RNA Degradation The RNA Degradation plot is only available for Affymetrix probe level Plot Prin Comp Plot CEL data The RNA Degradation plot displays the average expression intensity for a set of marker probes at each location of the probe set Separate profiles means at probe locations connected by lines are plotted for each array Trends in the profile lines such as those displayed in Figure 5 7 indicate differential labeling as a function of location in the probe set If there is no differential labeling you should expect to see relatively flat profiles RNA Degradation Plot Mean
294. reted as a special key Whenever a special key is found it is expected that the first occurrence contains the flag START and that there will be another line following this line which contains the same key but with the END flag These special keys define data blocks by the corresponding START and END lines As an example ImportInfo START FileType MAS 5 Summary Data ChipName mgu74av2 CDFPath SaveAs MouseSwimExprSet PrintOutput 1 ImportInfo END In this example the special key is ImportInfo The start of the ImportInfo block is defined by the line containing ImportInfo START and the end by ImportInfo END 371 Appendix A Creating a Design File Any blank lines in between these keys are ignored and any lines that are not blank are interpreted as settings Any line in the block must follow the format Valuename Value where Valuename is the name of the value to set and Value is the value to use Any valuename that is not recognized is ignored Use the tables below to determine the recognized value names You can assign an empty value to a valuename by leaving it blank Valuename Table J 3 contains descriptions rules and examples of the special keys This is followed by a set of tables one for each special key that lists the allowable valuenames for the key along with any rules and examples Table J 3 Table of special keys the data imported such as file type array name etc int
295. rex Browse i Outer Grid number of grid rows and columns Grid Rows 4 Grid Columns fa 4 Inner Grid choose columns in data file representing Inner Grid Row Bow o Inner Grid Col Coum S i Control ID b Control Value EMPTY Gene Name Name x M Output Saye Layout As tpleyout Cancel k current Help Figure 4 18 The layout setup for the malaria data Now shift to the Variable Selection amp Filtering tab and select F532 Mean B532 Mean F635 Mean and B635 Mean for Green Foreground Green Background Red Foreground and Red Background respectively While still on the Variable Selection amp Filtering tab select the Flags variable in the Extra Variables for filtering later list and click on the button to move it into the Keep list on the right side of the dialog The Flags column marks problem spots that should be filtered out before doing any serious analysis The result of selecting expression and filtering variables is displayed in Figure 4 19 Two Way Reference Design Import Data From Two Channel x File Selection MIAME Variable Selection amp Filtering Options m Variable Selection Green Foreground F532 Mean 7 Red Foreground F635 Mean x Green Background B532 Mean x Red Background 6635 Mean X Weights lt none gt hal Extra Variables for filtering later All Variables Keep Cancel x afs entries
296. riment with the same or similar design by modifying the file locations and names and factor levels as needed In fact if you have many arrays in your experiment you can create a text file from scratch with all the design content and then read it with the Read Existing Design button Reading the design in this way will fill the file name fields and their associated factor levels For information about creating a design file see Appendix A Creating a Design File MIAME is an acronym for Minimal Information About a Microarray Experiment and this information can be entered on the second page of the Import Data From Affymetrix dialog This information is not required but it is stored on the resulting object to identify the source of the data Once you ve entered MIAME information for any experiment the first three fields are saved and are filled automatically the next time you open this dialog The MIAME tab of the Import Data From Affymetrix dialog is shown in Figure 2 8 Import Data From Affymetrix a x File Selection MIAME MAS Variables amp Filtering CEL Filtering Options Experimenter s Name Seige Izumo Laboratory Beth Israel Deaconess Medical Center Contact Information Experiment Title E xercise Induced Hypertrophy Experiment Description Eight week old mice were swum in tanks with diameter of 50 cm and a surface area of 2000 cm2 Phsical interaction between the mice discouraged floating and unlike rats Kap
297. rl 4 94 94 swirl 4 spot wild type experiment Cy5 date comments 1 wild type 2001 9 20 NA 2 swirl 2001 9 20 NA 3 wild type 2001 11 8 NA 4 swirl 2001 11 8 NA Number of labels 4 Dimensions of maInfo matrix 4 rows by 6 columns Notes C PROGRAM FILES INSIGHTFUL splus62 module ArrayAnalyzer examples SwirlSample txt C Summary statistics for log ratio distribution lst Qu Median Mean 3rd Qu Max Min swirl l spot 2 74 Swirl 2s Spot 2 72 Swi lsd Spot 2 29 swirl 4 spot 3 21 0 79 0 58 0 48 0 29 4 42 Dela OG M02 Gel wise 0 75 0 46 0 42 0 12 2 65 0 46 lt 0 26 0 27 0 06 2 90 D Notes on intensity data Normalization We can extract the controls with the subsetting method for marrayRaw objects as follows JHH Extract controls gt swirl raw controls lt swirl rawLcontrols control gt swirl raw lt swirl rawLcontrols control Comparative plots are produced with maPlot and maBoxplot We can create the plots either within print groups for a single chip or for all chips disregarding print tip groups From the Command Line JHHF Boxplots of controls vs noncontrols by print tip group gt graphsheet gt par mfrow c 1 2 gt maBoxplot swirl raw controls 3 main Controls by Print Tip Group srt 90 gt maBoxplot swirl rawL 3 main Non controls by Print Tip Group srt 90 Figure 4 42 displays the resulting graph Controls by Print T
298. rms of differential expression testing An example of the The volcano plot complete with hyperlinks can be sent to an HTML file for later viewing It can also be sent to an S PLUS graphics window Figure 7 13 shows a typical volcano plot with the interactive menu generated by clicking a point in the differential expression region of the plot 287 Chapter 7 Differential Expression Testing 288 When the plot is viewed in an S PLUS graphics window the active points are not hyper linked to the annotation databases However hovering the mouse over active points shows the gene name ID in the upper right corner of the graphic as it does for the HTML display as shown in Figure 7 14 Gene Name 38428_at Figure 7 14 Finding the gene name on the graph for differentially expressed genes Heat map Differential Expression Analysis Plots A heat map plot shows a 2 D image plot of the 300 genes with lowest p values by default along the vertical axis versus the experimental conditions on the horizontal axis See Figure 7 15 This graph is also hyperlinked to public annotation databases and displays the gene identifier in the upper right corner of the plot Left clicking in a colored rectangle exposes the menu for making an annotation database choice The left and top margins of the graph contain a dendrogram resulting from applying hierarchical clustering to the expression intensity values and treatment conditions respectively ao
299. rwise combinations of chips invariantset prd td c 0 003 0 007 The chip with the median mean intensity for the set is chosen as the reference chip Based on the rank of the intensities a group of invariant genes is chosen for each chip A smooth spline is fit to this invariant set of genes This is a pairwise normalization for each chip in the set to the reference chip 242 Pre Processing and Normalization for Affymetrix Probe Level Data Table 6 6 Normalization methods available through the normalize function Normalization Methods Default Function Values Description loess subset Normalizes the chips with respect sample 1 dim mat 2 to each other by forcing log ratios 5000 epsilon 10 2 eee ie to be scattered around the same maxit 1 log it T ods verbose T span 2 3 constant curve This is family loess symmetric accomplished on more than two arrays by averaging the pairwise loess curves ScaleNormalization Methods qspline target NULL samples The quantiles from each array and NULL fit iters 3 the target are used to fit a system of min offset 5 A s 2 TA cubic splines to normalize the data spline method natural smooth TRUE spar 0 p min 0 p max 1 incl ends TRUE converge FALSE verbose TRUE na rm FALSE quantiles Assuming an underlying common distribution the set of chips are normalized so that their quantiles have the
300. s Two Color Data ax Data Specs Options Filter r General j Additional Colnames row 6 Wweikshest nuniber Row name col auto x E Penta 7 Start col 1 I Strings as factors End col KEND gt z I Sort factor levels Start row fe I Labels s numbers Ende eND gt gt Century cutoff fis30 Format string en nis inane Delimiter fabs t and orspaces i CS Yd Decimal Point Peio gt 1000s Separator None T Separate Delimiters Date format M d yyoy X Time format frimmsstt xf Cancel Apply i current Figure 4 38 Options for importing gene ID data Now from the S PLUS command line create the named list as follows tplayoutPF lt as list TPLayoutComplete ID names tplayoutPF lt TPLayoutComplete Name tplayoutPF is the named list with gene ID s used to construct the URLs for annotation Note Creating the To create the data frame that is used by the graphlet for type the Information Data following command into the command line window Frame 170 The first part of the name of the named list tplayout must match the name of the layout object you created when you read in the data Also the layout object name must be all lower case AA Annotation Columns lt data frame Two Way Reference Design ANNODATA c PF COLNAMES c PF ID URLS c http malaria ucsf edu ligolink php 0LIGO MENULABEL c PF ID
301. s available through the ArrayAnalyzer GUI and opens the door to additional analyses The flexibility and feature rich S PLUS language make it an ideal platform for exploratory analysis statistical testing and modeling of gene expression data This section is designed to expose you to the critical functions for differential expression testing of microarray data If you do not plan to run your analyses from the command line you can skip this section The relevant information for a cDNA microarray is 1 layout of the chip 2 experimental design 3 gene ID s 4 expression intensities All of this information must be read into S PLUS and assembled into a single object for further analysis The primary convenience of importing data through the GUI is the coordination of the following three functions which read the above information read marrayLayout This function creates objects of class marrayLayout to store layout parameters for two color cDNA microarrays read marrayInfo This function creates objects of class marrayInfo The marray Info class is used to store information regarding the target mRNA samples co hybridized on the arrays or the spotted probe sequences e g data frame of gene names annotations and other identifiers read marrayRaw This function reads in cDNA microarray data from a directory and creates objects of class marrayRaw from spot quantification data files obtained from image analysis software or databases Re
302. s in lymphoma patients using a customized cDNA lympho chip This chip included genes expressed in lymph cells and genes that play an important role in cancer They ran samples from the three most common adult lymphomas on the lympho chip namely diffuse large B cell lymphoma DLBCL follicular lymphoma FL and chronic lymphocytic leukemia CLL and a variety of other lymphoma and leukemia cell lines Each chip had a reference sample with cy5 labeling used for the experimental samples and cy3 for the reference samples Alizadeh et al 2000 identified two distinct subtypes of DLCBL from a hierarchical cluster analysis of the resulting data the relevant heat map and dendrogram from this analysis are given in Figure 3a of Alizadeh et al 2000 321 Chapter 8 Cluster Analysis 322 Note that Alizadeh et al 2000 focused their attention on B cell differentiation genes based on visual examination of hierarchical cluster analysis and heat map visualization of 96 samples run on arrays of more than 10 000 genes We do not recommend this qualitative approach rather we suggest genes be included in cluster analyses based on their differential expression according to a reliable statistical hypothesis testing procedure We provide two analyses of the subset of data presented in Figure 3a of Alizadeh et al 2000 using hierarchical and partitioning cluster routines We actually use the data as summarized by Cluster Eisen et al 1998 and prepa
303. s required to use GC RMA for background correction In S PLUS the information is provided in named lists stored in libraries S ARRAYANALYZER combines the Affymetrix CDF and probe libraries into one library called lt chipname gt cdf The complete installation of S ARRAYANALYZER includes several libraries for common chips in the S PLUS library directory For example hgu95av2cdf and hgul33acdf hgul33bcdf mgu74av2cdf moe430acdf rae230acdf rgu34acdf If you are working with these chips hgu95av2 hgu133a hgul33b mgu74a rae230a then you do not need to do anything the S ARRAYANALYZER functions that operate on the CEL data will find the named lists If you used the custom or standard install and you need to manually install cdf libraries for other chips copy the cdf libraries into the S PLUS library directory Each cdf library contains four objects e lt chipname gt cdf derived from the Affymetrix Chip Definition File gives the pm and mm information for each gene represented on the chip This is used by a variety of summarization methods in the affy library e lt chipname gt probe derived from the Affymetrix probe_tab files gives the DNA sequence for each spot This is used by the GC RMA background correction method and by other functions in the matchprobes library e lt chipname gt xy2i and lt chipname gt i2xy are little functions that map from the x y position on the chip to the linear index of a row in a CEL file
304. s with experimental conditions we need to set up the experimental conditions in S ARRAYANALYZER The easiest way to do this is through the Create Modify Design dialog Open the Create Modify Design dialog by clicking on the Create Modify Design button on the File Selection page of the Import Data From Affymetrix dialog 23 Chapter 2 Examples Affymetrix MAS Data 24 Number of Arrays ig Number of Factors fi Figure 2 3 The default Create Modify Design dialog Use the Create Modify Design dialog to specify the following 1 The number of arrays to be read 2 The number of factors in the experiment Currently one or two are allowed 3 The name number of levels and level values for each factor Start by incrementing the Number of Arrays to 15 To modify the default factor Name of Levels and Level Values type them into the appropriate field For this example click the row in the Name field once and enter CondTime enter 5 for the of Levels field and click the Level Values field and enter Swim3wks Swim4wks Swim4wks 1wk NoSwim4wks and NoSwim4wks 1wk One Way Design For the one way analysis of the Mouse Swimming data we read 15 arrays and one factor with five experimental conditions The resulting dialog is displayed in Figure 2 4 Create Modify Design Number of Arrays fis Number of Factors fi Name of Levels Level Values CondTime 5 Swim3wks Swim4wks Swimdwks 1w Cancel Hep
305. scuss the diagnostic plots and specific methods for normalizing two channel and Affymetrix data 213 Chapter 6 Pre Processing and Normalization IDEAS IN NORMALIZATION Normalizing to One Point Normalization Using Loess 214 Normalization typically involves adjusting distributional summaries of data from each chip to common reference values Sometimes reference values are supplied by the user Alternatively some methods assume one chip is the reference chip and the other chips in the set are then normalized to the target chip s reference values A median is a robust estimate of the center of the data distribution where just under 50 of the data on either side of the median can be moved to infinity and the median value will not be affected It isa quantity that defines the center of the data 50 of the data are above the median and 50 are below the median Consequently the median is often used as a reference value The inter quartile range IQR estimates the spread or variability of the data and is computed as the range of the middle 50 of the data The IQR is a robust estimator of spread in that you can move just under 25 of the data at either end of the distribution to infinity and the IQR remains unchanged The robust properties of the median and IQR make them good reference values in normalization procedures There are also methods which do not require a target or reference chip These methods use the information from th
306. sed in the plotting functions 223 Chapter 6 Pre Processing and Normalization 224 Table 6 1 Plotting functions available for two channel data Plotting Function Default Parameter Settings Description maPlot x maA y maM a z maPrintTip Produces scatter plots of microarray spot statistics for the classes marrayRaw marrayNorm and marray Two Creates plot for first chip given maBoxplot x maPrintTip y maM Produces box plots of microarray spot statistics for the classes marrayRaw marrayNorm and marrayTwo Plots by print tip groups if given one chip Creates one box plot for each chip if given more than one ma Image x maM subset TRUE col contours FALSE bar TRUE Creates spatial images of shades of gray or colors that correspond to the values of a statistic for each spot on the array The statistic can be the intensity log ratio M a spot quality measure e g spot size or shape or a test statistic This function can be used to explore whether there are any spatial effects in the data for example print tip or cover slip effects Creates plot for first chip given maDotPlots data x list maA id ID pch col nrep 3 A dot plot showing the values of replicated control genes Location and Scale Normalization Within Array Normalization Loess Normalization Normalization Methods for Two Channel Data df Additional e
307. separating the fields in each line of the data files Normally this can be detected automatically but it is provided as an option for unusual cases where auto detection can not determine the field delimiter 76 Two Sample Design Import Data From Affymetrix E x File Selection MIAME MAS Variables amp Filtering CEL Filtering Options Import Options Header lines to skip auto X Data delimiter fauto Cancel x afs entries Figure 3 9 Options page of the Import Data From Affymetrix dialog Press OK when you have completed the dialog and the data are imported It is now ready for use in S ARRAYANALYZER Normalization Normalization procedures may be applied to both raw probe set intensities and to summarized expression intensities For examples of normalizing expression summary data see Chapter 2 Examples Affymetrix MAS Data and Chapter 6 Pre Processing and Normalization In this section we focus on normalizing probe set data without summarizing it first 77 Chapter 3 Examples Affymetrix Probe Level Data Normalization Dialog 78 Open the Normalization dialog by selecting Normalization from the ArrayAnalyzer drop down menu slob Data r Normalization Show Data of Type Normalization ha Affymetrix CEL PEEPAR M y Data eaAffyBatch v I MvA Plot Save As caAffyBatch norm PM Box Plot Probe Set C PM PM and MM When to Show Before amp After C Only After
308. sequence information targets object of class marrayInfo containing target sample information Note that besides file names and location and the columns indicating foreground and background intensities we need to supply the objects we just created swirl layout swirl gnames and swirl samples The command line call for reading the four data files is as follows gt fnames lt paste swirl 1 4 spot sep gt swirl raw lt read marrayRaw fnames path AApath name Gf Gmean name Gb morphG name Rf Rmean name Rb morphR layout swirl layout gnames swirl gnames targets swirl samples gt swirl raw Pre normalization intensity data Object of class marray Raw Number of arrays 4 arrays A Layout of spots on the array Array layout Object of class marrayLayout Total number of spots 8448 Dimensions of grid matrix 4 rows by 4 cols Dimensions of spot matrices 22 rows by 24 cols Currently working with a subset of 8448 spots Control spots There are 2 types of controls Control N 768 7680 Notes on layout C PROGRAM FILES INSIGHTFUL splus62 module ArrayAnalyzer examples fish gal B Samples hybridized to the array 179 Chapter 4 Examples Two Color Data Filtering Out Controls Quality Diagnostics 180 Object of class marraylInfo maLabels of slide Names experiment Cy3 1 81 81 swirl 1 spot swirl 2 82 82 swirl 2 spot wild type 3 93 93 swirl 3 spot swi
309. sign Top 15 Genes The first page of the output is the Top 15 Gene list Included in the Summary Table table is the test statistic the raw p value the adjusted p value and the fold change When the output is displayed in HTML and annotation data is available for the chip each gene identifier at the beginning of each row of the table is hyperlinked to one or more annotation database replication factor C activator 1 2 40kDa Summary Output for LPE Test with Bonferroni Adjustment Top 15 Genes Test Statistic Raw p Value Adj pValue Fold Change EDL CANON TACI an 0 0 0 89 ornithine decarbc ee 0 119 protein phosphata 0 0 92 interferon alpha i 4 77 0 73 bone morphogenetic 4 solute carrier fami eukaryotic translat 8 4 NA 9 roteasome prosome 5 6 4 6 0 79 1 39 0 76 a 9 9 1 54 0 94 0 73 0 92 1 08 eukaryotic translat 0 73 cathepsin K pycnod E 0 71 omosome condensa cyclin dependent ki 35P2 and RIPK1 dom i transforming growth 7 7 6 9 9 99 6 0 9 4 3 1 2 Figure 3 18 Summary of top 15 differentially expressed genes Each gene is hyperlinked to annotation databases 87 Chapter 3 Examples Affymetrix Probe Level Data Volcano Plot 88 A volcano plot displays the logarithm of adjusted p value versus fold change as shown in Figure 3 19 The vertical lines indicate fold change values of plus or minus one and the horizontal line indicates a significant test p value af
310. sis Workflow Two Sample Design Two Way Design References Chapter 4 Examples Two Color Data Two Color Data Analysis Workflow iii Hu Dna N 15 16 17 18 54 65 67 68 69 93 123 125 126 Contents vi Two Sample Design Two Way Reference Design From the Command Line 127 147 174 Chapter 5 Quality Control Diagnostics and Filtering 191 Quality Control Diagnostics Filtering Chapter 6 Pre Processing and Normalization Introduction Normalization Ideas in Normalization Diagnostic Plots Normalization Methods for Two Channel Data 192 203 207 209 210 214 218 220 Pre Processing and Normalization for Affymetrix Probe Level Data Normalization Methods for Affymetrix MAS Data References Chapter 7 Differential Expression Testing Introduction Statistical Tests Controlling Type I Error Rates GUI for Two Sample Testing GUI for LPE Testing GUI for ANOVA Testing Differential Expression Analysis Plots Differential Expression Summary Table Output References 232 251 256 259 260 261 265 270 276 280 286 294 298 Contents Chapter 8 Cluster Analysis 301 Introduction 302 Hierarchical Methods 304 Partitioning Methods 308 Examples from the GUI 310 Examples from the Command Line 320 References 329 Chapter 9 Annotation and Gene List Management 331 Annotation and Gene List Management Functionality 332 Annotation Libraries 334 Appendix A Creating a Design File 369 Introductio
311. soft Internet Explorer File Edit View Favorites Tools Help esk gt amp A A Reach fagravorites media D S fel S S NCBI Dibran of Medicine Entrez P h ote Genome Structure OMIM PMC Preview Index History Clipboard Details Items 1 20 of 50 Page i of 3 Next Hashimoto O Ueno T Kimura R Ohtsubo M Nakamura T Koga H Related Articles Links Totimura T Uchida 5 Yamashita K Sata M Inhibition of proteasome dependent degradation of Weel in G2 arrested Hep3B cells by TGF beta 1 Mol Carcinog 2003 Apr 36 4 171 82 PMID 12669309 PubMed indexed for MEDLINE Masaki T Shiratori Y Rengifo W Igarashi K Yamagata M KurokohchiK Related Articles Links Uchida N Miyauchi Y Yoshiji H Watanabe S Omata M Kuriyama 5 Cyclins and cyclin dependent kinases comparative study of hepatocellular carcinoma versus cirrhosis Hepatology 2003 Mar 37 3 534 43 PMID 12601350 PubMed indexed for MEDLINE Yuan H Xie YM Chen IS Related Articles Links Depletion of Wee 1 kinase is necessary for both human immunodeficiency virus type 1 Vpr and gamma irradiation induced apoptosis J Virol 2003 F eb 77 3 2063 70 PMID 12525641 PubMed indexed for MEDLINE Strausberg RL Feingold EA Grouse LH Derge JG Klausner RD Collins Related Articles Links FS Wagner L Shenmen CM Schuler GD Altschul SF Zeeberg B Buetow KH Schaefer CF Bhat NK Hopkins RF Jordan H Moore T Max SI
312. specific association hgu95av2G0 Maps probe ids to GO data ids evidence code and ontology hgu95av2GRIF Maps probe ids to the unique PubMed id 335 Chapter 9 Annotation and Gene List Management 336 Table 9 1 S PLUS Affymetrix chip specific library objects Continued S PLUs Annotation object Description hgu95av2HGID Maps probe ids to internal HomoloGenlds hgu95av2LOCUSID Maps probe ids to LocusLink ids hgu95av2MAP Maps probes to cytobands hgu95av2NM Maps probe ids to RefSeq accession numbers for mRNA records hgu95av2NP Maps probe ids to RefSeq accession numbers for protein records hgu95av20MIM Maps probe ids to MIM numbers hgu95av20RGANISM The name of the organism hgu95av2 in this case hgu95av2PATH2PROBE Maps KEGG pathway ids to probe ids hgu95av2PATH Maps probe ids to KEGG pathway ids hgu95av2PMID2PROBE Maps PubMed ids to probe ids These annotation objects contain some redundancies the objects most commonly used include lt gt ACCNUM lt gt LOCUSID lt gt GENENAME lt gt GO lt gt PATH lt gt PMID and lt gt UNIGENE The general annotation libraries e g GOAnnoData KEGGAnnoData may be manually attached as follows gt library lt chipname gt AnnoData Annotation Libraries S PLUS objects for some of these libraries are shown in Table 9 2 Table 9 2 S PLUS general annotation library objects S PLUS Annotation object Description
313. t J Pubmed Correction Bonferroni z J GO Website m Use Affymetrix IDs IV Save Affy IDs to File Affymetrix ID File Probe ist tat Browse I Open Affymetrix GO Browser Username Password Cancel Apply if f current Figure 9 10 The General Options page of the Annotation dialog Options chosen Use Affymetrix IDs group This writes out a file of Affymetrix IDs ProbeList txt by default corresponding to the genes selected according to the Annotation dialog options This ProbeList txt file can be uploaded 351 Chapter 9 Annotation and Gene List Management File Edit ols Help Back gt gt OA A Qsearch Favorites meda 4 B S M S Address ja https www affymetrix comfanalysis query go_analysis affx z fed Go Links a 6 X home logout register your profile contact site index search site gt AFFYMETRIX rh Register Now at a a PRODUCTS EUOSSIES SUPPORT TECHNOLOGY RESEARCH COMMUNITY CORPORATE NetAffx Gene Ontology Mining Tool Upload a probe set list into the NetAfic Gene Ontology GO Mining Tool to review either a graph of GO terms that are associated with those probe sets or a listing of GO terms that are at the same distance from the root GO node suitable for uploading into Data Mining Tool For more information please see the Gene Ontology Mining Too E GO GRAPH E DMT GO ANNOTATION FILE Ontology Branch GeneC
314. t NoSwim w3 NoSwim4wks C Program Files Insig NoSwim4wiwi txt NoSwim wiw1 NoSwim4wks 1lwk xl C Program Files Insig NoSwim4wiw2 txt NoSwimgwiw2 NoSwim4wks 1lwk xl C Program Files Insig NoSwim4w1w3 txt NoSwim w1w3 NoSwimgwks 1wk xl Type Filename or right click to bre aaga lenia z Swim3wks xl File Type Mas 5 Summary Data Array Type lt required gt x CDF m Step 3 Save Output Save Data Set As mye xprSet I Display Report Cancel RI if Figure 2 5 Browsing for data files You can find the swimming mice example data by navigating to your splus62 module ArrayAnalyzer examples directory and selecting the Swim3w1 txt file Repeat for the other 14 txt files entering one file per field A more efficient option is to select Swim3w1 txt left click and scroll to the last sequential file in this list Swim4w3 txt repeat for the NoSwim data Alternatively read the design file named MouseSwimDesign txt in the examples directory to load the design and create file associations File Type Note that the File Type entry MAS5 Summary Data listed below the files is automatically detected once a file is selected The dialog is designed to prohibit mixing file types Step 3 Save The Data Object One Way Design Array Type Array Type is a required field You must select the name that corresponds to the Affymetrix array you used for your experiment Some common examples are hgul33a and hgu95a
315. t each iteration the two closest groups are merged The compact method also known as the complete linkage or farthest neighbor method is similar except that the distance between any two groups is defined to be the largest distance between any two members from different groups The average weighted linkage method uses the average of the distances between the objects in one group and the objects in the other group These are all heuristic criteria Hierarchical Methods Distance Metrics The agglomerative hierarchical methods use several measures for defining between cluster distance or dissimilarity These methods proceed by merging the two clusters with the smallest between cluster variability based on the chosen between cluster dissimilarity measure at each stage of the process Four distance metrics are available as options for heirarchical clustering through the StARRAYANALYZER GUI Euclidean The square root of the sum of squares of differences between points in the two groups Maximum The maximum difference between points in the two groups Manhattan The sum of absolute differences between points in the two groups Binary The proportion of non zeros that two vectors do not have in common the number of occurrences of a zero and a one or a one and a zero divided by the number of times at least one vector has a one Handling Missing Values The distance methods handle missing values in different ways Following is a bri
316. ter adjustment for FWER FDR family wise error rate family detection rate With the annotation turned on in an HTML display you can click specific genes to determine which are in the significant zone collection of libraries based on the Bioconductor project an open source and open development software project developed by leading statisticians with the goal of providing tools for current microarray and genomics research S ARRAYANALYZER 2 0 is based on Bioconductor version 1 3 with some libraries ported from 1 4 Welcome Features The S tARRAYANALYZER module helps you analyze microarray data using these built in features Enhanced microarray data import Supported file types now include Affymetrix MAS 5 ASCII or Excel CHP binary and CEL ASCII or binary and any two channel microarray data ASCII or Excel Improved pre processing Updated Affymetrix probe level methods e g affy gcrma matchprobes from the Bioconductor libraries and added QC diagnostics PCA RNA degradation and filtering genes arrays Added cluster analysis methods A palette of supervised and unsupervised learning methods e g mclust are included in this release Added annotation for derived genelists For example you can import Affymetrix NetAffx annotation data and your customers can update the data from the Affymetrix Web site You can also connect to online databases for genelist annotation analysis and integrate with OntoExpress and D
317. ter doing the Bonferroni correction Points located in the upper outer sextants are those with large absolute fold change and small significant p value Each of those points is active so you can click on a point to access annotation information from Locus Link or GenBank Gene Name connective tissue growth factor Volcano Plot Probe ld 36638_at oO LocusLink k T 2 T gt a D a a 1 D o pa Figure 3 19 A volcano plot which is the logarithm of p value versus fold change Points below the horizontal line are hyperlinked to annotation databases Heat Map Plot Two Sample Design A heat map plot shown in Figure 3 20 shows a two way layout of the most differentially expressed genes along the vertical axis versus the experimental conditions on the horizontal axis This graph is also hyperlinked to the annotation information Tal oa Re Sample cg2b CEL Gene 33543_s_at Exp Value 1 26 Accession Number LocusLink 2 Summary Volcano Plat Variance Plot Figure 3 20 A heat map plot shows differentially expressed genes as a function of experimental conditions The map is hyperlinked to annotation databases 89 Chapter 3 Examples Affymetrix Probe Level Data Chromosome Plot A chromosome plot displays the entire chromosome with differential expression marked up for positive down for negative for each gene represented on the chip The top 15 dif
318. text of few replicates the results may be misleading For more information see the cautionary note in the section Within Gene Two Sample Comparisons of section Statistical Tests GUI for Two Sample Testing Output Options The Output Options group is a list of check boxes for selecting which graphs you want as output r Output Options IV Volcano Plot Y Axis Orientation Eold Change Line 20 f IV Heat Map I Chromosome Plot V Top 15 Genes Figure 7 5 The Graph Options group in the Two Sample Test dialog Each of these options is described in detail in the section Differential Expression Analysis Plots 273 Chapter 7 Differential Expression Testing Output 274 The Output group controls where the graphs are displayed and the gene list table is saved after the testing step is complete Output I Display Output in PLUS I Save Output as HTML Save HTML F Display HTML Gutput Save As mM ultTest Figure 7 6 The Output group of the Two Sample Test dialog Display Output in S PLUS Displays the selected graphics in an S PLUS graphic device Save Output as HTML Saves the S PLUS Graphlet with selected graphs and the significant gene list to HTML files to view later Display HTML Output View the S PLUS Graphlet with selected graphs in a browser The displayed Graphlet has a hyperlink to the significant genes table Points on the Graphlet and entries in the significant gene list are
319. the distances between points in the column and row spaces and fit the hierarchical cluster models Note that since the data are normalized with mean zero and variance one prior to calling dist the resulting matrix is equivalent to a dissimilarity matrix produced using cor gt module ArrayAnalyzer gt fileName lt file path getenv SHOME module ArrayAnalyzer examples figure3a cdt gt mat3a lt importData fileName rowNamesCol 1 colNameRow 1 drop c 2 4 startRow 3 type ASCII gt stand norm lt function x x mean x na rm T sqrt var x na method available gt aliz cmat lt apply mat3a 1 stand norm cluster rows 323 Chapter 8 Cluster Analysis 324 d i i d gt gt aliz distl lt dist t aliz cmat aliz hclustl lt hclust dist aliz distl method average cluster cols eliz disi2 lt dist as matrix aliz cmat aliz hclust2 lt hclust dist aliz dist2 method average color 6 GC B like color 5 Activated B like color 1 GC centroblasts array3a colors lt e rep o 16 reptl 2 repls 6 rept5 23 plot heat map and dendrograms par mai c 0 0 0 0 omi c 0 2 7 1 4 1 1 image aliz cmatLaliz hclust2 order aliz hclustl order axes F bty n par new T omi c 6 55 2 75 0 1 15 plclust2 fn aliz hclust2 cex 1 rotate me F 1ty 1 colors array3a colors aliz hclust2 order par new T omi c 0 02 0 95 1 42 7 75 plclust2 fn aliz h
320. the heat map We can now change the clustering method simply by selecting one of the partitioning methods and re running the analysis Partitioning around medoids automatically estimates the number of clusters and generates three graphics Examples from the GUI PAM PC Biplot Genes Component 2 Component 1 These two components explain 72 13 of the point variability Figure 8 6 Partitioning around medoids plot of the first two principal components of expression values with clustering boundaries specified PAM Silhouette Plot Genes r T T T T 1 0 0 0 2 0 4 0 6 0 8 1 0 Silhouette width Average silhouette width 0 33 Figure 8 7 Partitioning around medoids silhouette plot of expression values Separation of the silhouettes for each group indicates good separation of clusters 315 Chapter 8 Cluster Analysis Clustering Two Channel Expression Data 316 PAM Parallel Coords Plot Genes Gene Expression Intensity Experimental Condition Figure 8 8 Partitioning around medoids parallel coords plots expression values across experimental conditions for each cluster Clustering two channel data is not that much different from single channel expression intensity data The key difference is that you now have a choice of variables to cluster on You can choose between 1 M the logs ratio of the two channels or 2 both red and green channels In the second cas
321. the sample appears at the bottom of the list rather than in the original order 203 Chapter 5 Quality Control Diagnostics and Filtering Gene Filtering 204 Figure 5 12 displays the Filtering dialog set up to drop three arrays before running an analysis ox Array Filtering Gene Filtering r Data Show Data of Type Atfymetris Sumr Data MouseSwimE xp Array Name mgu74av2 m Output IV Display Report Save s myFiltering Filter Columns Samples to Keep Swim4wks1 Swim4wks2 Swim4wks3 Swim4wks 1wk1 Swim4wks 1wk2 roo Samples to Drop Swim3wks1 Swim3wks2 Swim3wks3 OK Cancel Apply d j current Figure 5 12 Filtering dialog set up to remove three arrays before running analyses To filter on genes you create an S PLUS logical expression that specifies the condition or conditions for keeping or deleting genes Some examples include the following Expr Intensity gt 2 RedForegnd gt 5000 amp M gt 2 Flags 0 InAtLeast 1 You can type expressions directly into the Expression field on the Gene Filtering tab or build it sequentially In the following example we run a filtering experiment from Chapter 4 Examples Two Color Data Refer to the chapter to create the data set TPMarrayRaw and follow the instructions below for creating the filtering expression Filtering For this example use the Flags column to create expressions f
322. the three replicate pairs observed at week three of the Mouse Swimming study 34 EML MPWIMS TML Mpls One Way Design LAL Mpls Emp TMYWIMS LMpUIMS EMEWIMS MEWIMS LMEWIMS EMLMPUIMSON Genes Present Plot ZTMLMPUIMSON LMLMPUIMSON EMpWIMSON TMYWIMSON LMPUIMSON JUasald Intensity Box Plot Genes present plot of all arrays in the MouseSwimExprSet data set Each bar corresponds to the percent of genes detected as indicated by the Detection variable in the Affymetrix MAS summary data Figure 2 15 EYMLHSAMPUMSON ZAMLSAMPUMSON DIML S PINSON ESMPLUIMS ON ZSYMPLIMSON LSMYULIMSON EYAL SHMPLUIMS DM LESHMPWINS DM b SYM plums ESIMPWIMS ZSYMPWIMS LampumS ESYMELUIMS ZSYMEWIMS o Lomewms 35 Boxplots of expression intensities for all arrays in the MouseSwimExprSet data set Each box represents the distribution of expression intensities for one array Figure 2 16 Chapter 2 Examples Affymetrix MAS Data Normalization 36 Swim3wks Swim4wks Swim4wks 1wk NoSwim4wks eu NoSwim4wks 1wk OX D0 Figure 2 17 Principal components plot of all arrays in the MouseSwimExprSet data set Each point corresponds to a different array Different symbols represent different experimental conditions The diagnostic plots
323. the two sample test procedures These are made available in two dialogs the Two Sample Tests and the LPE Test dialogs To demonstrate the idea open the LPE Test dialog by clicking ArrayAnalyzer gt Differential Expressing Analysis gt LPE Test In the resulting dialog select Affymetrix data type the MouseSwimExprSet norm data set and the levels for comparison in Compare Level 1 and Compare Level 2 fields Save the result in MouseSwimLPEtTest The settings are displayed in Figure 2 27 Differential Expression Analysis LPE Test a ol x m Data mYariance Estimation Show Data of Type Smoother D F 10 Jatt metriz a Number of Bins ho H Data MouseSwinE vt Z Trim pa 5s Factor CondTime z r Output Optins Compare Level 1 Swim3wks x IV Volcano Plot Compare Level 2 NoSwimdwks x Y Axis Orientation Array Name mgu74av2 negative i Fold Change Line 2 0 Options ger FWER FDR 005 Heat Map Adjustment BH IF Chromosome Plot Alt Hypothes Not equal z Variance Plots IV Top 15 Genes r Output IV Display Output in S PLUS I Save Output as HTML Save HTML MouseSwimLPETe F Display HTML Gutput Save As MouseSwimLPETe OK Cancel Apply k j current Help Figure 2 27 The LPE Test dialog Setting the adjustment to BH as we did for the ANOVA yields 22 significant genes when comparing Test Input MouseSwimExprSet norm Test Input Class
324. these objects is an S PLUS list with names corresponding to unique probe ids and entries the actual accession numbers For reverse mapping lookups the values of the annotation element are the names keys and the entries are the probe ids Using this approach S PLUS annotation objects can be simply built for non Affymetrix and custom arrays Once these S PLUS annotation objects are in place they are automatically used by the graphical and tabular reports in S tARRAYANALYZER In the first example we show how the S ARRAYANALYZER default tabular and graphical plots are linked to LocusLink and UniGene Web sites for simple metadata annotation This example uses the Melanoma data Fox et al 2001 and revisits results from reading in the cel files summarizing with RMA and performing differential expression testing using LPEtest with Bonferroni FWER control Refer to Chapter 3 Examples Affymetrix Probe Level Data for details on this process Three of the tabular graphical reports from the LPE analysis include annotation metadata links 1 Top 15 Genes 2 Volcano plot 3 Heat map We first review these reports and then show how to link to additional annotation information from the Annotation dialog and via the command line scripting environment Annotation Libraries The volcano plot shows the genes p value v fold change Each gene is active so you can click on a gene to access annotation information using its LocusLink ID or
325. tle Before Normalization gt boxplot LCG N style bxp Tatt gt title After Normalization 59 Chapter 2 Examples Affymetrix MAS Data M vs A Plots 60 The resulting boxplots are displayed in Figure 2 37 Before Normalization After Normalization i ia aiai io a 7 p an ar ae 5 aia ian my T ry T T NJ i o a 1 P a i e r e e J e el ad J i la I ala cda od Loa tebe eel ey d P CGa CGb CG24a CG24b CGa CGb CG24a CG24b Figure 2 37 Before and after normalization plots for the Melanoma data logged expression intensities We can do an M vs A plot of the logged expression intensities in LCG N with the mva pairs function For MAS4 5 data this function plots all pairwise scatter plots of M vs A for each treatment condition and replicate combination Because there are over 12 000 probes on each array we randomly sample 2 000 of them before plotting and because the intensities have already been logged we turn that off in the plotting function dHHE First remove missing values gt LCG N lt na exclude LCG N gt graphsheet gt mva pairs LCG NEsample dim LCG N 1 2000 J log F Differential Expression Testing LPE Test From the Command Line The resulting plots are displayed o
326. to the slide Differential expression is computed as the difference between the color intensities of the two samples Prepare cDNA target Proves hllerdcrrcay Figure 1 5 Custom two color cDNA microarrays compare treatments on each array by tagging them with different colors This two color design provides a way of estimating differential expression independent of chip to chip variability cDNA microarrays may be customized both gene content and layout by the experimenter Consequently the layout and gene content must be provided at the time of the analysis This makes data import more complex than for Affymetrix chips for which there are many standard fixed layout descriptions We provide a cDNA example in Chapter 4 to illustrate the steps involved in the analysis EXAMPLES AFFYMETRIX MAS DATA Affymetrix Data Analysis Workflow Experimental Design Two Sample Design One Way Design One Way Design Swimming Mice Data Importing Data Quality Diagnostics Normalization Differential Expression Analysis Gene Lists Management Annotation From the Command Line Importing Data Data Manipulation Normalization Differential Expression Testing References 16 17 17 17 18 18 21 31 36 40 48 50 54 55 57 59 6l 65 15 Chapter 2 Examples Affymetrix MAS Data AFFYMETRIX DATA ANALYSIS WORKFLOW The entire process of analyzing gene expression data with Affymetrix MAS 4 5 or cel file data can be done through the
327. tured here These images must be converted to numbers the quantification step before analysis can proceed Scanners like those from GenePix and Agilent produce raw intensity data files which form the starting point for differential expression analysis in S ARRAYANALYZER A gene expression experiment entails washing microarrays with concentrated cellular material and quantifying how much cellular substance binds to the gene spots A great deal of binding at a spot indicates that gene is active in the cell that is the gene is being expressed in that cell or tissue Knowing which genes are being expressed or not expressed and how that expression changes under different experimental conditions is of great importance in functional genomics and in developing new diagnostics therapeutics or treatment strategies S tARRAYANALYZER is designed to work with data from different commercial microarrays In particular it works with data from Affymetrix microarrays and from custom cDNA microarrays available through several suppliers We describe in more detail the Affymetrix Arrays Microarray Data differences between these two basic types of microarrays through examples in Chapters 2 through 4 Here we introduce them briefly to aid in understanding the examples that follow Affymetrix GeneChip microarrays represent each gene with an oligonucleotide 25 mer probe spotted at typically 11 20 pairs of spots 22 40 spots in all Each pro
328. ur arrays the total number of different test statistics is at most six This means that the smallest possible p value for a two sided alternative is 0 333 When the number of replicates increases to 10 per sample the minimum p value drops to 0 00001 and when there are 20 arrays per sample the minimum p value drops to 0 00000000001 107 The local pooled error LPE test is an experimental procedure designed for low replicate studies When there are few replicates in a study the degrees of freedom for estimating the standard error of differential expression within genes may be as low as one or two In this context estimates of within gene standard errors are imprecise resulting in increased Type I and Type II errors In particular with the large number of genes on the chip there will always be genes with low within gene error estimates by chance so that some signal to Raw P Values Statistical Tests noise ratios will be large regardless of mean expression intensities and fold change The local pooled error test attempts to avert this by combining within gene error estimates with those of genes with similar expression intensity In this sense the LPE approach is similar to the SAM method of Tusher et al 2001 and the B statistic of Lonnstedt and Speed 2002 LPE estimates used for differential expression testing are formed by pooling variance estimates for genes with similar expression intensities The LPE is derived by first
329. ure 2 25 A heat map plot shows differentially expressed genes as a function of experimental conditions 45 Chapter 2 Examples Affymetrix MAS Data Graphical Annotation 46 Clicking one of the hyper linked points in either the volcano plot or the heat map pops up a menu for selecting the database to query for annotation information Selecting either one opens an HTML page in your default web browser displaying a brief description of the gene with a hyper link to more detailed information Figure 2 26 shows an example page from LocusLink with annotation for one of the differentially expressed genes in the swimming mice example Taxonomy Structure View Mm Romi One of 1 Loci Save All Loci ABCDEFGHIJKLMNOPQRSTUVWXYZ Click to Display mRNA Genomic Alignments spanning 1796 bps Mus musculus Official Gene Symbol and Name MGI Roml rod outer segment membrane protein 1 LocusID 19881 Overview Locus Type gene with protein product function known or inferred Product rod outer segment membrane protein 1 Alternate Rom 1 Symbols Function Submit GeneRIF All Pubs Gene Ontology Term Evidence Source Pub e G protein coupled photoreceptor activity TEA MGI e cell adhesion TEA MGT Figure 2 26 Annotation information from LocusLink The LPE Test One Way Design Alternative to the ANOVA dialog which tests multiple contrasts at once you can test a particular contrast of interest directly with one of
330. usts the location and scale of the data so expression values on all arrays have equal medians and equal inter quartile ranges i e the spread between the 25th and 75th percentiles is the same The data normalization plots can be viewed before and after normalization or just after by selecting the appropriate choice Note that the Probe Set radio buttons are disabled for Affymetrix MAS data These will be discussed in Chapter 3 Examples Affymetrix Probe Level Data One Way Design Click OK or Apply to produce the normalized data and generate the pre and post normalization MvA pairs plots and box plots as shown in Figures 2 20 and 2 21 Swim3wks After medianlQR Normalization Swim3w1 7 j ki ki a 5 10 15 _ 5 10 15 1 44 Swim3w2 i ki 5 10 15 1 39 0 574 Swim3w3 Figure 2 20 MvA scatter plot matrix Each plot is M vs A for two arrays To determine which arrays are used for each plot go down vertically and left horizontally from the plot to the first array names you encounter The numbers in the boxes below the diagonal are values of the interquartile range of M for the pair of arrays obtained by going up vertically and right horizontally to the first array names you encounter 39 Chapter 2 Examples Affymetrix MAS Data Differential Expression Analysis 40 Before medianlQR Normalization After medianlQR Normalization 15 10
331. usts the probe level data by e Subtracting the global background signal and noise as described in section mas on page 246 Summarizing the 11 20 mismatch MM and perfect match PM values using a simple trimmed average difference procedure see section avgdi ff on page 246 The output intensity for MAS 4 0 data is termed Avg Diff 251 Chapter 6 Pre Processing and Normalization Normalization Methods medianIQR 252 The summarized Affymetrix data from both MAS4 and MASS have not been suitably normalized for differential expression testing Note that the MAS software allows a very simple global scaling in which the user enters a target value TGT value With this method the average signal across all probes on each chip is calculated for each chip and a scale factor SF is determined for each chip such that chip mean SF TGT Thus the signals on each chip are scaled by a single number for each chip a crude form of normalization S ARRAYANALYZER provides three methods for normalizing this summarized data from the command line medianIQR vsn and affy scalevalue exprSet Quantile normalization is also available from the GUI Dilution exprSet is a sample exprset object available in the S ARRAYANALYZER database Dilution exprSet is a summarized version of the Dilution experiment object Dilution Please refer to the help files for more details We use Dilution exprSet to demonstrate normalization of summarized microarray dat
332. utes P t min Np 1 N 1 P 9 ETE P N 1 Controlling Type I Error Rates Py min N 1 pq N 2 P 3 gt P N 1 Pn 1 min 2py 1 PiN 1 Pin min piyy 1 and stops at the first adjusted p value pq that exceeds a Holm The Holm 1979 step down correction is pa Max 1 min N k 1 pqy 1 The procedure sequentially computes Pi max min Np 1 P g max min Np 1 min N 1 pq 1 and stops at the first adjusted p value pq that exceeds a Sidak SS The Sidak single step SS correction is N p 1 1 p All genes with adjusted p value p less than Ua are significant with an overall FWER of at most Ua Sidak SD The Sidak free step down SD correction is N i 1 p 1 1 po All genes with adjusted p value p less than Lia are significant with an overall FWER of at most Ha minP The Westfall and Young 1993 minP step down procedure is computed as Pa max Pr min in k N P lt Pa Ho For each pj pq is the resampling based probability of obtaining a p value no larger than Pii from simulated probability distributions generated by the decreasing sets Pa Pin Pia Pin Pips Pin 267 Chapter 7 Differential Expression Testing FDR Procedures 268 maxT The Westfall and Young 1993 maxT step down procedure is computed as Pa maxy y Pr max in c npITjl gt lto Hod For each py pq is the resampling b
333. vailable from the One Way Design GUI in the General Annotation group For the example we leave the three defaulted databases checked Figure 2 31 displays the resulting settings oix General Options Filtering Options r Data m Use LocusLink Ds Show Data of Type T Save LocusLink IDs to File DifExprTest zi LocusLink File LocusLinkList txt Data MouseSwimA Ng hens prey Heme muiaa IF Open Stanford Source r General Annotation eer er M LocusLink pen aS I OntoExpress IV Unigene Distribution Binomial z J Pubmed HEDY V GO Website ee a Use Affymetrix IDs a eee p IT Save Affy IDs to File SEN Affymetrix ID File JProbeLi ttxt Browse metris GO Browser F Open F Open DAVID EASE Browser Cancel Appt K j current Figure 2 31 General Options settings for annotation of the genes identified by the MouseSwimANOVABs4wksBH ANOVA Now on the Filtering Options tab in the Contrast Filtering group select the Swim3wks NoSwim4wks contrast and check the Significant genes checkbox Clicking the Recalculate button in the Gene Sort Order Options group will show you how many genes are selected by the filtering Note the Limit number of genes to field which puts a cap on the number of gene ID s that will be sent to the databases for annotation extraction Figure 2 32 displays the resulting settings 51 Chapter 2 Examples Affymetrix MAS Data 52 General Opti
334. ving neutrophil production one may be specifically interested in expression of genes associated with the GO term defense response In the case of such targeted interest restriction to these pathways a priori can have considerable advantages compared to the approach of considering all genes simultaneously In particular FWER FDR adjustments are more contained and smaller but real effects can be detected more often The GO filtering process includes the following four components 1 Select the GO term s of interest 2 Identify the set of assayed genes that are annotated at that GO term Identifiers can be for example the Affy IDs for genes on the chip with the GO annotation of interest 3 Subset the expression data set using the appropriate identifiers 363 Chapter 9 Annotation and Gene List Management 364 4 Perform the analysis e g differential expression on the expression data subset We illustrate this a priori filtering pipeline with a study on the progression of granulocyte differentiation The study was done using Affymetrix mgu74av2 arrays as a time course experiment in which a model cell line was used to elucidate mechanisms by which retinoic acid signaling causes promelocytes to stop dividing and turn into mature neutrophils The progression under study is shown below in Figure 9 23 A more detailed explanation and analysis of these data is available by e mailing a request to S PLUS pharma insightful c
335. x Chips annotation is readily available on public databases see Chapter 9 Annotation and Gene List Management for more detail so we have automated annotation through interactive S PLUS graphlets and the GUI For examples see section Graphical Annotation on page 46 or section Annotation on page 50 We can set up graphical annotation for any two channel array as long as the annotation information is available via a typical http internet protocol and the URL s to individual gene annotation are systematically structured The URL structure required is a base URL plus a gene ID which when concatenated together creates the unique URL to the specific gene annotation page The TP is associated with a web site dedicated to Malaria parasite annotation with appropriate URL naming structure so we ll use it as an example of setting up annotation for custom non standard arrays To create graphical annotation for the TP test results proceed as follows 1 Create an S PLUS named list containing all the gene ID s which are part of the gene specific URL 2 Create a small data frame with the following information ANNODATA character string which is appended onto the chip name to create the name of the named list created in 1 COLNAMES column name of gene names in the layout file URL base URL to the annotation database Gene ID must be appended to complete the URL MENULABEL label displayed in the drop down menu of the graphlet DISPLA
336. x MAS Data Creating an Expression Intensity Data Frame Logging Expression Intensities 58 The above expression creates a pair of boxplots for the first 0 hour replicate By repeating the commands for the other three arrays we produce the remaining plots in Figure 2 36 0 hr Replicate A 0 hr Replicate B wo 10 Log 2 Expression Intensities 5 l in Log 2 Expression Intensities controls noncontrols controls noncontrols 24 hr Replicate A 24 hr Replicate B 15 15 10 10 in Log 2 Expression Intensities 5 Log 2 Expression Intensities 5 e 5 Q 8 OERI Siedlacy a controls controls noncontrols Figure 2 36 Boxplots of control versus noncontrol spots for the melanoma data Now extract the expression intensities from each array in preparation to normalization and differential expression testing Extract the avg diff column and add it to a data frame named CG For MAS5 data this is the signal column gt CG lt data frame CGa cga avg diff CGb cgb avg diff CG24a cg24a avg diff CG24b cg24b avg diff Compute the base 2 log transformation of the intensity values as follows Any intensity values less than one will be negative or missing after taking logs so we set them explicitly to one in the ifelse function call JHH Threshold and log adjusted average differences gt LCG lt CG gt for i in names
337. x Probe Level Data Setting Up The Analysis 94 Table 3 2 Experimental design and file association for the mouse surgery study Age Time Rep Array label File name Old 4hr 2 Old4hr2 Old4hr2 Old 4hr 3 Old4hr3 Old4hr3 Young Ohr 1 YoungOhr1 YoungOhr 1 Young Ohr 2 YoungOhr2 YoungOhr2 Young Ohr 3 YoungOhr3 Young0hr3 Young 1hr 1 Younglhr1 Younglhr1 Young 1hr 2 Young1hr2 Younglhr2 Young 1hr 3 Young1hr3 Younglhr3 Young 4hr 1 Young4hr1 Young4hr 1 Young 4hr 2 Young4hr2 Young4hr2 Young 4hr 3 Young4hr3 Young4hr3 The fundamental question posed by this study is Are there any differences in gene expression between the young and old age groups A secondary question is Do the differentially expressing genes change over time The Two Way Design Importing Data Import Affymetrix Data Dialog Two Way Design To answer these questions the example focuses on data for the young and old mice collected at 0 hours 1 hour and 4 hours The experimental design is a balanced two way design as displayed in Table 3 3 Table 3 3 Experimental conditions for the examining differential expression in healing for young versus old mice Young Old 3 reps 0 hours 3 reps 0 hours 3 reps 1 hour 3 reps 1 hour 3 reps 4 hours 3 reps 4 hours Start by reading in all the arrays To import Affymetrix data from the main S PLUS menu select ArrayAnalyzer gt
338. xamples of diagnostic plots Boxplots of all chips in swirl dataset pre normalized swirl dataset gt maBoxplot swirl Boxplots of pre normalization red foreground intensities for each grid row for the Swirl 81 array gt maBoxplot swirl 1 x maGridRow y maRf main Swirl array 81 pre normalization red foreground intensity J MvA plot of chip 93 overlaid with loess curves for each print tip group default gt maPlot swirl 3 Image plots of chip 81 the first chip in the object gt malmage swirl x maGf green foreground plot gt malmage swirl x maM log intensity ratio plot S ARRAYANALYZER supports both location and scale normalization for two channel data These normalization methods help normalize the red and green channels on one chip within chip normalization One of the most common location normalization methods is loess normalization This method normalizes the data to the loess curve on the MvA plot Please refer to the section Normalization Using Loess on page 214 for more details about loess curves For two channel data the intensity log ratio M log and the overall intensity A log RG are most commonly used to create the loess regression curve Like the median normalization which shifts the median of each chip to zero the loess normalization effectively shifts 225 Chapter 6 Pre Processing and Normalization the loess curve of the data to zero
339. y The y axis is typically the intensity log ratios and the x axis is the grouping variable Figure 6 1 on page 211 shows an example of a typical box plot for an experiment with two replicate slides for each dye swap condition Scatter plots of spot statistics allow the user to highlight and annotate subsets of points on the plot and assess patterns of differences in intensities between channels or chips Such patterns may be visualized via fitted curves from robust local regression or other smoothing procedures The MvA plot shows the log ratio of the intensities difference of the log intensities usually termed M between channels or chips to the average of the log intensities usually termed A for the channels chips Figure 6 2 shows an MvA plot for one chip in the swirl dataset two channel data with loess curves overlaid for each print tip group Diagnostic Plots Note MvA plots from the GUI plot a maximum of 2000 genes If there are more than 2000 genes in the experiment then only 2000 randomly sampled genes are plotted MvA Hexbin Hexbin plots provide another dimension to plotting data Unlike the Plots MvVA scatter plots which use no more than 2000 points hexbin plots use ALL the data and show its density or concentration in a particular area by coloring that area So hexbin plots not only let you see ALL the data but they let you see its concentration Hexbin MvA are important to show a global picture of the data bef
340. y estimate a set of parameters rma2 is only available for the command line and is not recommended The rma2 method is available via bg correct rma2 object bgtype 1 or bg correct object method rma2 Setting bgtype 2 will result in the original rma method being applied mas The mas background method performs the noise correction described in the Statistical Algorithms Description Document SADD a white paper from Affymetrix This method divides the chip into a given number of zones and uses the lowest 2 of the intensity values An Example With bg correct Background Correction with gcrma PM Correct Methods Pre Processing and Normalization for Affymetrix Probe Level Data to compute the background intensity within each zone Smoothing across zones is done by computing a zone weight which is based on the distances of spots to zone centers The background at each cell location x y is computed using these weights A similar computation is made for the noise at each cell The background corrected value is computed as a function of the background at x y noise at x y and the threshold and floor noise values at each x y cell location based on the noise at x y such that the cell intensity remains positive bg correct takes an AffyBatch object and returns an AffyBatch object Following are some examples of background correcting a sample extracted from the Dilution experiment Please refer to the Dilution
341. ypically the silhouette value s i is computed and then represented in the plot as a bar of s i If A denotes the cluster to which object i belongs we define Partitioning Methods a i average dissimilarity of i to all other objects of A Now consider any cluster C different from A and define d i C average dissimilarity of i to all objects of C After computing d i C for all clusters C not equal to A we take the smallest of them b i ming d i C The cluster B that attains this minimum namely d i B b i is called the neighbor of object i This is the second best cluster for object l The value s i can now be defined gt a _ _ b i a i i max a i b i wy We see that s i always lies between 1 and 1 The value s i may be interpreted as follows sli 1 gt object iis well classified s 7 0 gt object i lies between two clusters sli 1 gt object iis badly classified The silhouette of a cluster is a plot of the s i ranked in decreasing order The silhouette plot shows the silhouettes of all clusters next to each other so the quality of the clusters can be compared The average silhouette width of a partitioning cluster analysis is the average of all the s i from every cluster This is a measure of quality or goodness of the cluster analysis One typically runs pam several times using a different number of clusters within a specified range appropriate for the number of samples and compa

Download Pdf Manuals

Related Search

Related Contents

SPB - Oxford Laboratory Balances    v rI` 8 _5 - e    USER MANUAL - Conference Room AV  ASSMANN Electronic KVM Cable  Samsung M1618N Manual de Usuario  Cisco Systems 2130 User's Manual  IDRO  3 - My Soehnle  

Copyright © All rights reserved. Failed to retrieve file