User`s Manual for CNAG (Copy Number Analyser for GeneChip
Contents
1. case Note that in principle the precise ploidity can be determined only by cell based assay e g FISH cytogenetics or FACS analysis because the ploidity information is lost forever when cell membrane is disrupted although its minimum value could be inferred form the distribution of heterozygous SNP calls Eea Cancel Eile Data Parameter View Analyze Help i DREH mat io Bmw ome e e ba Sample PC11_X compared with Samples Display average number 10 Max and Min deleted ploidy 2 744731 SD 0 107396 Jina MAMAS 6 2 Changing the mode of averaging and parameter of HMM analysis To change the mode of averaging click ave icon and select the mode of averaging as already described You can change the parameters of HMM analysis by clicking the HMM icon This requires when the tumor sample is contaminated by normal cells in primary samples or when tumor heterogeneity exists You need to have some insights into basic principle of copy number inference using HMM and see more details in our paper Icons I RbaRef gt CHAG File Data Parameter View Analyze Help D cm AS sor SF Wl vh SS BO ii sm e gt Store ave HMM Chromosome 6 3 Viewing the results of the previous analysis and reanalysis You can save the results by selecting Store from the Data menu or click Store icon The res2. for haploid genome is 1 Within the upper most graph the copy numbers inference from the hidden Markov Model HMM analysis are given in green bars In the next lower panel chromosome cytobands are presented together with green bars indicating heterozygous SNP calls In the bottom panel the maximum likelihood of LOH is calculated for each contiguous region For this LOH inference algorithm in detail see appendix A The thicker the blue bar the higher the Kbakef GHAG File Data Parameter View Analyze Help orale H Oe AAA gt ge Xba Sample PC11_X compared with Samples Display average number 10 Max and Min deleted ploidy 2 124506 SD 0 182304 11 probability of the region having LOH Note that this is only inference based on the observed frequencies of heterozygous SNP calls and not showing true LOH It fails if the individual have the same chromosome by descent for example by consanguinity It also fails in the regions where the allelic imbalance becomes extreme or when the tumor cells show heterogeneity In this view you can specify the regions expected to be diploid more closely by clicking the Range Select icon as shown in chromosome 13 in this example Following step is identical except that you have more freedom of specifying ploidity value for the selected region Finally the SD value is improved to 0 1073 in this
3. items selected as references 9 After the CNAG dialog asking to confirm the selection you will see the following dialog in which you choose the mode of taking moving averages and the mode of the copy number graph In CNAG moving averages of the copy numbers are taken for several adjacent SNPs In the Exclude Max and Min mode averages are taken for the specified number of adjacent SNPs with the most deviated values being excluded Alternatively you can include all the SNPs in calculations in Include All Data mode The number of SNP loci to be averaged can be specified by entering the number from 1 to 10 1 means no averages are taken In Line mode the graph is drawn with every adjacent points being connected with a line while it is not in the Point mode Click OK to obtain a genome wide view of the copy number Male Enter Averaging Number 1 10 ho Points Lines Averaging Mode Draw Method Include All Data 10 In this view the copy numbers are shown for all the chromosomes In the upper panel of each chromosome each red spot represents a log2 ratio for each SNP locus and the middle blue lines or points in Point mode shows the averaged log2 ratios Calibrations of the graph are provided so that each interval of horizontal lines be 0 5 in log2 scale In pure tumors where all the tumor samples have haploid the expected CNAG output is 0 5 although the theoretical value
4. rh A i sp EE EV RRE AREA A Eee pe EE rr ee ee ee rro ds crono rro cr rro coords rr rro rr rr ii ii 5 5 fH 5 He ES ee gt 2 ee y From 64819 Ch 2 position 138381343 Db to 67294 Ch 3 position 3547242 selecte Calculate StDev eee ETS s z qm ai it A A AA ee ee ee Ready pr 8 CNAG compute log2 ratios for SNP probes within the specified region against all the reference samples included in the Ref file and ask you to select the reference samples to be used for further In the dialog names of the reference samples are displayed with pre calculated SD values You may select as many alos samples as you want to but on average no Sample Load Emea more improvement of the SD value will be me expected with more than five references h sampleCRL2362 0 164509 br 0 166322 e rd Moreover note that using inappropriate Y sample317 0 168731 references 1 e those with high SD values can A ES only degrade the result Unfortunately this E sample316 0 171500 optimization step 1s not automated although C sample104_3 0 172976 the selection of optimized references is usually sample318 0 173456 pe se el er eer not difficult When you choose the references C sample132_2 0 174226 click OK to move on to the next step samplel70 3 0 175082 O sampleckl2339 0 176078 O sampleckl2363D 0 179546 sample307 0 180301 sample308 0 180923 samplels7 0 182403 FT era oe ee NE 6
5. strongly recommend preparing your own Ref files because too much differences in experimental conditions between one lab and the other will not compensated even with CNAG algorithms You should understand that uniform experimental conditions between samples and references are the most important factor to obtain high quality copy number analysis results using Affymetric GeneChip platform For theoretical discussions see our paper CancerRes 10 14 1 10 2005 If the experimental condition of your tumor sample is too different from those of reference samples you are not likely to obtain high quality results Below is the list of known parameters that may affect experimental conditions 1 PCR devices PCR conditions are the primary determinant of the kinetics of amplification reactions If a total PCR time is significantly different from one machine to another you should NOT compare data between both machines 2 Lot of Tag polymerase 3 Amount of DNA to be analyzed 4 Quality of genomic DNA All the experiment should be performed carefully to make each experimental condition as uniform as possible 4 2 Creating reference files The following procedure is not really cool and in the future versions it may be improved However please be patient for the time being Also we recommend saving the preexisting Ref files if ever as XbaRef_bak or HindRef_bak To create your own reference files 1 Export intensity files for reference sample
6. these four files and click OK TSG Enter Averaging Number 1 10 10 Averaging Mode Draw Method Points Include All Data Lines Enter the number of SNPs to be averaged and choose an averaging mode and an Draw Method in the same manner described before and click OK FIERE CNPG HE File Data Parameter View Analyze Help Ose RS RA eH ee Ble gt el Xba Sample 5929 compared with the self reference Display average number 10 Max and Min deleted SD 0 370449 HK HO mme E ee eee IE 1 tebe cee oe ee ma S S ee ke es ls ee ate ee Cr um DI ee ee ee mn ee he s a des ee Ye ele ee e y de le e E AAA A AAA ER mme er DE NE a RN EG 0 7 You should adjust the ploidity of the tumor samples by specifying a region having known ploidity as in the analysis using non self references Note that in this mode of analysis signal ratios are plotted without taking logarithm Red lines and green lines correspond to alleles having larger or lesser deviated signal ratios respectively and blue lines are copy number analysis in which analysis is performed with A and B probe signales being combined and identical to the previous analysis using non self references except that no logarithm is taken Each pink bar is a discordant SNP call between tumor and reference represe
7. 1Gb or more is recommended for quick response 3 Preparing Data files CNAG requires two kinds of files for analysis an intensity file containing the intensity data of all SNP probes and the SNP file corresponding to the same tumor sample both of which can be obtained from the GDAS software using the export to txt file option Note that you should not sort or rearrange the SNPs in GDAS before exporting the data Each file should be saved as sample name followed by _X txt or H txt according to the array used Any names can be used for sample name but they must be unique and we strongly recommended that both should be easily linked together sample name X txt and SNP sample name X txt for files from Xba arrays are examples Intensity files should be placed in the tumorsampleX folder or tumorSampleH files according to the array type Similarly SNP files are expected to be either in SNPtumorsampeleX or SNPtumorsampleH folder If a self reference sample is available both intensity file and SNP file for the self reference should be placed into the sampledataX folder for Xba arrays or sampledataH folder Hind arrays 4 Createing reference files 4 1 Recommendation Ref files are used by CNAG to calculate the copy number in tumor samples when a self reference is unavailable and you must have at least one reference file for each array Xba or Hind You may use attached Ref files if available but we
8. M ee gn I n s ul toe SI Noam mia
9. User s Manual for CNAG Copy Number Analyser for GeneChip Ver 1 1 1 Introduction CNAG is the software developed specifically for analyzing copy number alterations and LOH in cancer cells using Affymetrix GeneChip platform It implements a robust algorithm to correct between chip variations introduced by the subtle difference in experimental conditions enabling high quality copy number analysis When the constitutive DNA is available it also allows for allele based copy number estimation to sensitively detect allelic imbalances Copy number calls are automated according to the copy number calls based on the hidden Markov model analysis and the LOH inference is also enabled even when constitutional DNA is not available 2 Installation To install CNAG simply put the CNAG and CNAGView folders directly under the YC directory The CNAG folder will contain a CNAG excutable file and other files and folders named MSVCR71D dlP MFC71D dll and MSVCP71D DLL files sampledataX sampledataH SNPtumorsampleX SNPtumorsampleH tumorsampleX tumorsampleH temporary files and data folders In addition it should contain XbaRef and HindRef files and if it is not found you need to create these reference files by yourself according to the instruction described below using CNAG software We confirmed that this program runs on Windows XP and 2000 At least 512Mb RAM is required but
10. ew Ref file using Make Reference option 1 Make a text file having entries to be added in just the same manner as described in the previous section and save it 2 From the Data menu select the AddReference option and specify the file you just saved 3 CNAG asks you the name of the Ref file you are about to updated Make sure that the Ref file specified is the file you really want to update Do not confuse Xba and Hind Click OK and the Ref file updated 5 Start and setup application for dataanalysis 1 To launch application simply double click the CNAG executable file 2 From file menu choose Load Ref and specify the reference file for the array you want to analyze For example open the XbaRef file for Xba arrays and the HindRef file for Hind arrays Both reference files can be read in at the same analyzing session If a reference file is successfully recognized by the application it will return a message requesting you confirmation Click OK Now you are ready to analyze your data 6 Data analysis To start analysis select New from the file menu and a window appears that request you to choose a mode of analysis as below Daine Type of Analysis Type of Arrays Self reference xba Hind Cancel Xba Hind Temp files of both types should be available in selecting Xba Hind analysis 6 1 Analysis using non self references 1 Select the t
11. nd g23 3 Position 114310434 point 89881 signal 0 131989 point 89882 signal 0 916303 point 89883 signal 2 736584 point 89884 sienal 2 140147 point 89885 signal 0 579132 Goto Web Information i il it UN When you click the right button at the location of your interest information around that location is shown up including SNP numbers cytoband and the exact position of the SNP marker together with log2 ratios near around Click OK and you go to the UCSC browser Home Genomes Blat Tables Gene Sorter PCR DNA Convert PDFIPS Help UCSC Genome Browser on Human July 2003 Assembly move zoom in zoom out position chr8 114 310 334 114 310534 jump clear size 201 bp Base Posi Ti 431 SSR PINK iS VE E a VIF SN FETO Bands Localized by FISH Mapping Clo Chromosome ET Oo STS Markers on Genetic blue and Radiation Hybrid black Maps STS Markers Clones Placed on Cytogenetic Map Using FISH Gap Locations Gap Known Genes Based on SWISS PROT TrEMBL mRNA and RefSeq CSMDS RefSeq Genes RefSeq Genes Ensemb1 Gene Predictions Ensemb1 Genes AceView Gene Models With Alt Splicing CSMDS bDeces CSMDS hDeces CSMD3 cDec 3 CSMDS aDec 3 CSMDS eDeces Genscan Gene Predictions NT_665646 422 Human mRNAS from GenBank AB114604 AY210419 AB114605 Bx645252 Human ESTs That Have Been Spliced Spliced ESTs Human Ch imp Mouse Rat Chicken Multiz Alignments amp PhyloHMM C
12. nting true LOH Green bars are heterozygous SNP calls Any deviation of red and green lines indicates that allelic imbalance exists in that region and really this is more sensitive to detect LOH than discordant calls when contamination of normal cells or tumor heterogeneity is extensive We can predict existence of LOH even without discordant SNP calls in contaminated tumor samples Of course in principle allele based copy number analysis is possible for SNPs showing heterozygous calls in the reference sample Thus significant reduction in resolution may occur where heterozygous SNP calls are thin Note that on average heterozygous SNP calls are expected in 30 of all SNPs To obtain a chromosome view move on to the Chromosome View by clicking Chromosome View icon in just the same manner as explained above sections Closer inspection is also available from the Explorer View Pe use CNAG File Data Parameter View Analyze Help o o i D RSG THE SHAN OO ce 2 Xba Sample 5929X compared with the self reference Display average number 10 Max and Min deleted SD 0 250843 OC A ee E RE ee ee eee ee ee eee a ee ee eS es ee ee le oe ee dt an 7 CE ee ee Lee ee eee JA REE mr 000000 00000000 002 En EE me eee namam OJ SS ER EEE one oma me SS SS ER SS EE M 0 EG GE OG O OG o o 000 o a n P mam oe OG OI m m ma BOGE ANS ae aro EA S GY 1 I E
13. ons Conserva tion Takifugu rubripes Aug 2962 fri Translated Blat Alignments Fugu Blat Chicken Feb 208042316412 Alignment Net Level 1 Level 2 Level 3 Level 4 Level 5 Level 6 Simple Nucleotide Polymorphisms SNPS SNP_A 1799495 rs750262 Repeating Elements by RepeatMasker SINE LTR unknown move start Click on a feature for details Click on base position to zoom in move end Jo C around cursor an on left mini e for track ia options Jo C e iti Color oe 6 5 Printing CNAG outputs I bake gt CHAG File Data Parameter View Analyze Help De ORF 1 ee IEEE To output the result to printer change the view mode to the Printer View and click the Printer icon Sorry for this inconvenience since we are not so familiar with the computer programming Make sure the expected output in preview before printing Any CNAG view is required to be changed to Printer View mode to be properly printed 6 6 Analysis using self references The other mode of copy number analysis is one using self reference Of course the information about the alterations of tumor genomes is greatly increased with analysis using this mode because copy number alterations in the two alleles can be separately analyzed based on the genotyping information in constitutive DNA of the tumor for details see our paper In this mode comparison between corresponding SNP probes becomes more accurate because
14. s from GDAS using export options save them following the rule as described in section 3 2 Place all of the intensity files you want to include into the sampledataX or sampledataH folder as described 3 Make a text file describing the names of the intensity files in which each line contains a single entry with the format of sample name for an intensity file named as sample name_X H txt 4 Save the text file within the SampledataX H folder You can use any file name you like for saving 5 From the Data menu select MakeReference and specify the file you just saved 6 Wait a moment and CNAG ask you the name of the Ref file you are about to create Enter XbaRef or HindRef and select the directory you want to the file to be saved in Now a Ref file for Xba or Hind arrays is created If you want to create another Ref file just repeat these procedures 4 3 Adding new references to the existing Ref file You can add the new intensity files in the existing Ref file Note that the Ref file thus updated may be different form what is newly created if the order of file list is different However since CNAG do not know the names of the reference files but only track them using the internal IDs a replacement of the Ref file with another will lead to disastrous result because the saved temporally files may now indicate totally different reference files Use the add Reference option to add refereces instead of n
15. the intensity of each SNP in tumor sample is compared to what it would be in normal diploid genome To see this in general A and B probes show different dynamics in hybridization signals So for examples even in the ideal experimental conditions intensities of diploid signals would be different between AA AB and BB genotypes Thus a single normal reference frequently provides a better SD value than multiple best fit non self references although the uniform experimental conditions between tumor and reference is the primary determinant of the SD value So we strongly recommend to always perform both experiments simultaneously 1 To start the analysis select New from the File Menu and choose Self reference and an appropriate enzyme Waluz Type of Analysis Type of Arrays fr xba Non self reference Hind Cancel ba Hind Temp files of both types should be available in selecting Xba Hind analysis 2 Check New Sample to analyze a new sample Alternatively you check Edited Sample for reviewing and editing the result of the previous analysis Dialga Select Sample New Sample Tumor Signal File 5929x Cancel Normal Signal File 5969x Tumor SNPs File SnP5929 ile SNP 5969 X Edited Sample In this mode of analysis you should specify four files intensity files and SNP files for tumor and reference samples By clicking each tag select
16. ult is stored in temp file format within Temporary file folder in CNAG directory You can reproduce the result of the previous analysis and if you want reanalyze the sample 6 4 Working with chromosome view To identity all the copy number changes in detail and more precisely fix those abnormalities move on to the chromosomal view by clicking Chromosome icon and choosing chromosome you want to review Fjell Select a Chromosome C chg Bo ch2 C cho C Chis ch3 C chi chi C Ch4 Cch2 ch20 C chs SN a yd C ch O ch4 ch22 chis Ch end Cancel Click OK to move to chromosome 8 Baker CHAG File Data Parameter View Analyze Help D a se id HAT VIN ana e gt Xba Sample PC11_X compared with 6 Samples Display average number 10 Max and Min deleted ploidy 2 745635 You can move between chromosomes by clicking Next ch and Previous ch icons There exist a number of gains and losses in this chromosome To get a closer look at these changes move to an Explorer view from a Monitor view by clicking Explorer view icon You can crawl on the chromosome using Move Right or Move Left icons i ZbaRef CNAG Eile Data Parameter View Analyze Help Do Sle wa Hm ee Bale gt k l Xba Sample PC11 X IE with ER Delay average nu TE Max and Min deleted ploidy 2 745635 SNPrumber 89883 Chromosomes 8 Cytoba
17. ype of the analysis according to which analysis you want to perform using self reference see our paper or Non self references Also select the type of array s You can analyze either Xba or Hind file or both together 2 3 Click OK and the following Dialog box will appears alas Select Sample Signal File Edited Sample Cancel 4 Choose New Sample click Signal File button and select an intensity file you want to analyze Similarly click SNP File and select the SNP file that corresponds to the intensity file you just selected Utils Select Sample amp New Sample Signal File pct 1 i ile SNP PC 11_ Edited Sample Then click OK The program calculates log2 ratios of signal intensities between the tumor sample and a reference sample for each SNP signal After a while CNAG will output the result Click TempCHGView icon I KbaRef gt GNAG File Data Parameter View Analyze Help De R amp W sus 3 i mi E CI ii i 7 When raw copy number plots appeared click Range Set icon to specify the region that is expected be diploid with the mouse button continuing to be hold down and click OK haRef CHAG File Data Parameter View Analyze Help SEE ee oe oe ee OT baRef CNAG BAR File Data Parameter View Analyze Help D am Sle ee ee a ara P CGH after compensation Log base
Download Pdf Manuals
Related Search