"user manual"
Contents
1. manual last update on July 8 2008 1 Introduction The geNorm VBA applet for Microsoft Excel determines the most stable reference genes from a set of tested genes in a given cDNA sample panel and calculates a gene expression normalization factor for each tissue sample based on the geometric mean of a user defined number of reference genes geNorm calculates the gene expression stability measure M for a reference gene as the average pairwise variation V for that gene with all other tested reference genes Stepwise exclusion of the gene with the highest M value allows ranking of the tested genes according to their expression stability The underlying principles and calculations are described in Vandesompele et al Genome Biology 2002 Accurate normalization of real time quantitative RT PCR data by geometric averaging of multiple internal reference genes The full article can be read at http genomebiology com 2002 3 7 research 0034 Please check the geNorm website at http medgen ugent be genorm for updates of the applet and user manual The current geNorm version is 3 5 An accompanying discussion group can be found at http groups yahoo com group genorm Ghent University recommends the use of a geNorm kit alongside the geNorm software geNorm kits for a wide variety of species can be purchased from PrimerDesign Ltd Each kit contains a panel of high quality real time PCR primers for candidate reference genes Kits can be2. Security level in Excel is not set at High but rather at Medium in Excel 2000 XP 2003 Tools Macro Security 3 Use a point as decimal separator important for continental Europe where a comma iS normally used for this purpose To change the decimal separator used in Excel go to Tools gt Options gt International Tab Alternatively you could change the decimal separator symbol in your operating system please consult the appropriate manual or help file for your system in Windows Start button Settings Control Panel Regional Settings 4 The input file should be an Excel data table with the first column containing the sample names and the first row containing the gene names The first cell of the first row and column cell A1 should be empty The other cells contain the relative gene expression levels Empty cells are NOT allowed The input file should be saved in the InputData directory where also an example data file is located Raw expression levels are needed for input these are the quantities NOT Ct values obtained from a real time RT PCR run either trough a standard curve or via the delta Ct method also called comparative Ct method see FAQ section for more information A prerequisite is that the raw data are comparable between the samples This is easily achieved for samples tested in the same plate If you want to compare quantities between plates then you need a few controls either a dilution ser
3. Fit for determination of propagated errors http www angelfire com rnb labfit A more comprehensive overview of all formulas and error propagation procedures including the error on the estimated PCR efficiency can be found in our qBase manuscript Hellemans et al Genome Biology 2007 The Ghent University geNorm and qBase technology are embedded in the professional real time PCR data analysis software qBasePlus http www biogazelle com 16 16
4. SD sampleCt 3 InE natural logarithm of the amplification efficiency SD sampleCt standard deviation Ct values of the sample replicates standard deviation for normalized expression levels Suppose n reference genes REF and one gene of interest GOI each with their own SD values calculated as outlined above are measured and the geometric mean of the n housekeepers is calculated as a reliable normalization factor NF gene of interest GOI SD GOI housekeeper 1 REF SD REF housekeeper n REF SD REF The Normalization Factor based on n reference genes is NF REF REF REF 4 geometric mean The standard deviation for this Normalization Factor is SP REF sp REF 4 4 SPREE es SD NF NF n REF n REF n REF 7 16 The standard deviation for the normalized GOI is 6 SD GOI GOI reese GOI SD NF SP Gory NF It might be more interesting to use standard error SE values instead of standard deviations SD as the latter is the error on a single measured value and the former is the error on the mean of repeated measurements Furthermore the SE value adds confidence to the calculated mean the true mean has a 95 chance of lying between the measured mean 1 96 times the SE SD SE m number of measurements i e 3 for triplicates in a PCR ym The error propagation rules are identical when using SD or SE values just replace
5. the SD values with SE values in formula 5 and 6 Note however that the above described procedure only provides the error for the normalized expression level of a gene of interest for a single sample mainly reflecting technical and pipetting variation and variation among the different reference genes used for normalization If you however average multiple samples e g biological replicates same cells independently grown or harvested technical replicates testing the same sample in different runs or grouping samples with similar properties e g diseased versus healthy tissue samples different rules apply This is a typical ANOVA analysis of variance problem which can in most cases not be solved using standard ANOVA statistics because the number of replicates repeated experiments to be grouped samples is often too small and because the results of the different PCR runs are most often not easily linked to each other due to the experimental variation this is only possible after correcting for reference samples which are tested in both experiments e g the same standards or samples run on both plates In fact there are 2 kind of variances a so called within variance this is the variance for the normalized gene expression level of one sample in a single experiment and a between variance for biological or technical replicates or sample groupings based on similar characteristics The between variance is typically much higher than
6. the within variance For the final between variance you don t need the within variance in this kind of small experiments you just consider your replicates or multiple samples as independent measurements and calculate the standard deviation on these measurements Just make sure that you can actually compare the different results which is natural if samples were analysed on the same plate but which is not trivial when samples are analysed on different plates See also section 7 8 16 example GOI SD experiment 1 after normalization 11 0 1 5 GOI SD experiment 2 after normalization 13 0 1 8 GOI SD experiment 3 after normalization 10 0 1 9 mean GOI SD 11 33 1 53 SD 11 13 10 A detailed example illustrating all calculations normalization and error propagation is available as Excel file on the geNorm website http medgen ugent be jvdesomp genorm example_calculations xls 7 Requirements 1 We have tested the applet only in Microsoft Excel version 2000 XP and 2003 on a Windows platform We cannot guarantee that the VBA applet works on other platforms or other Microsoft Excel versions Please let us know if you successfully use geNorm on another platform Excel version geNorm does not work in Excel 2007 due to a change in the VBA code base 2 Macro s need to be enabled in Excel If the message prompt enable macro s is not displayed while opening geNorm please check that the
7. Delete column button Repeating this process leads to stepwise elimination of the least stable reference genes until you end up with the two best reference genes which can t be further ranked Automated analysis NOU close all running instances of Microsoft Excel start up the geNorm applet in Excel Open File or double click on the geNorm xls file enable macro s when prompted load the expression data matrix raw data this means not yet normalized expression levels see requirements for data file press the Calculate button press the Automated analysis icon in the menu bar A first chart is generated Figure 2 equivalent from the Genome Biology paper indicating the average expression stability value of remaining reference genes at each step during stepwise exclusion of the least stable reference gene Starting from the least stable gene at the left the genes are ranked according to increasing expression stability ending with the two most stable genes at the right which can t be further ranked Example fibroblast data first step an M value for all 10 genes is calculated GAPD ACTB B2M HMBS HPRT1 RPL13A SDHA TBP UBC YHWAZ 0 687 0 864 1 076 1 128 0 651 0 873 0 762 0 928 0 700 0 651 with an average M value for the 10 genes of 0 832 see Figure below second step the gene with the highest M value is excluded HMBS and new M values are calculated for the remaining 9 genes GAPD ACTB B2M HPRT1
8. RPL13A SDHA TBP UBC YHWAZ 0 669 0 795 1 050 0 619 0 851 0 735 0 853 0 644 0 607 with an average M value for the 9 remaining genes of 0 758 see Figure below etc this process is repeated until only 2 genes remain 4 16 0 832 0 758 Average expression stability M HMBS B2M RPLI3A SDHA TBP ACTB UBC YHWAZ GAPD HPRT1 lt x Least stable genes Most stable genes gt 8 A second click on the chart icon generates a figure equivalent to Genome Biology Figure 3a indicating the pairwise variation V between two sequential normalization factors containing an increasing number of genes A large variation means that the added gene has a significant effect and should preferably be included for calculation of a reliable normalization factor Based on the Genome Biology data we proposed 0 15 as a cut off value below which the inclusion of an additional reference gene is not required For example if the V3 4 value is 0 22 then the normalization factor should preferably contain at least the 4 best reference genes Subsequently if the V4 5 value is 0 14 then there s no real need to include a 5 gene in the normalization factor An optimal number of reference genes for normalization in this example would therefore be 4 Note Please bear in mind that the proposed 0 15 value must not be taken as a too strict cut off The second graph is only intended to be guidance for determination of the optimal number of reference genes Sometimes the o
9. bserved trend of changing V values when using additional genes can be equally informative Anyway just using the 3 best reference genes and ignoring this second graph is in most cases a valid normalization strategy and results in much more accurate and reliable normalization compared to the use of only one single reference gene 5 16 5 Normalization flow chart In the following example 5 reference genes HK and one gene of interest GOI are quantified in 4 different samples by means of real time RT PCR It s highly recommended that the genes are quantified on the same batch of cDNA to minimize experimental variation in large part due to cDNA synthesis We routinely make cDNA from 2 yg of total RNA which is sufficient to quantify 50 different genes including a number of reference genes in duplicated reaction tubes see Vandesompele et al 2002 Analytical Biochemistry Furthermore we strongly advice to test the same gene on the different samples in the same PCR run to exclude further variation see also section 7 The Ct values are transformed to quantities either by using standard curves or the comparative Ct method Here the highest relative quantities for each gene are set to 1 These raw not yet normalized reference gene quantities are the required data input for geNorm In this example geNorm analysis would indicate that HK1 HK2 and HK3 are the most stable genes Hence after calculation of a normalization
10. factor either by geNorm or by manually calculating the geometric mean of these 3 reference genes the normalized GOI expression levels can be calculated by dividing the raw GOI quantities for each sample by the appropriate normalization factor Ct values J m e e a J 32 10 27 00 34 90 22 20 31 20 sample B 33 30 28 40 36 10 24 60 29 70 sample C 31 00 27 50 34 00 27 80 24 80 30 50 28 20 33 00 21 60 23 90 quantities geNorm input quantity GOI C R e e a J e sample B 0 144 0 379 0 117 0 125 0 018 sample C 0 707 0 707 0 500 0 014 0 536 1 000 0 435 1 000 1 000 1 000 sample A coi sample A normalization normalized factor GOI 6 16 6 Calculation standard deviation on normalized expression levels To calculate the standard deviation SD on the normalized gene of interest GOInorm expression levels the error propagation rules for independent variables have to be applied standard deviation on a relative expression value the delta Ct formula for transforming Ct values to relative quantities with the highest expression level set to 1 E deltaCt i Q g minCt sampleCt ies 2 Q sample quantity relative to sample with highest expression E amplification efficiency 2 100 minCt lowest Ct value Ct value of sample with highest expression The SD for this relative quantity Q is see addendum for derivation of formula SDQ E ln E
11. ies of the same standard or say 3 5 experimental samples which are run on both plates using these standards or samples you can link the data sets of both plates Please bear in mind that real time PCR is all about relative quantification relating the quantity of one sample to another and that it is therefore better to test different samples in the same run which you can then 9 16 easily and reliably compare You want to compare several samples for the same gene not several genes for the same sample 10 16 8 Frequently Asked Questions Qi A1 Q2 A2 Q3 A3 Q4 A4 Q5 A5 Q6 A6 When I load geNorm the menu bar is not visible Close all open instances of Microsoft Excel and reload geNorm What should I do if I have an empty cell Remove sample OR remove gene which contains an empty value and recalculate To remove the sample click on the empty cell and then click on the Delete row button To remove the gene click on the empty cell and the click on the Delete column button How many samples should I analyze to determine which reference genes are most stable In principle any number of samples higher than 2 would be sufficient However the more samples you use the more reliable are the conclusions We propose to use at least 10 samples Do I always have to retest and determine which genes are the most stable and should be used for normalization This depends on your experimenta
12. is available on the geNorm web site under citations http medgen ugent be genorm qBasePius software Biogazelle s professional successor of qBase based on qBase s universal quantification model and incorporating geNorm technology for systematic assessment of expression stability of reference genes http www biogazelle com 15 16 11 Addendum error propagation rule for functions in words The standard deviation on an arbitrary function of x is obtained by taking the derivative of that function evaluating it at the value of your measurement taking the absolute value and multiplying the result by the standard deviation on x in formula Consider function y E with SD being the standard deviation on x The standard deviation on y SD is given by SD 2 sp E InE SD lIn natural logarithm i X Given the comparative Ct or delta Ct function see formula 1 SD Q E n E SD deltaCt SD Q E Jn E SD minCt sampleCt SD Q E In E SD sampleCt SD deltaCt SD sampleCt because there s no error on the minimal Ct value it s just a fixed rescaling factor In practice minCt can be any value The minimal Ct value is just a practical rescaling factor because after transformation the highest expression level is set to 1 with relative rescaling of all other expression values Thanks to Wilton Pereira da Silva for help with derivation of this formula See also his software package LAB
13. l setup Once you have determined which genes and how many are required for accurate and reliable normalization you can use this information for future experiments as long as no significant changes in the experimental setup have been introduced e g once you have determined that HPRT1 GAPD and YWHAZ are the most stable reference genes for short term cultured human fibroblasts you can use these genes for normalization of all future fibroblast samples as long as you keep the culture conditions harvesting procedures etc identical Importantly the real time PCR data analysis software qBasePlus http www biogazelle com has built in geNorm technology for systematic assessment of expression stability of reference genes in each experiment Whom should I contact if I have further questions Please do not ask questions directly to one of the authors Instead use the geNorm discussion forum to ask your question or give feedback This forum intends to be a venue where geNorm users can interact and help each other Additionally information can be posted which genes are most suited for which biological system The geNorm discussion forum can be found at http groups yahoo com group genorm The geometric mean I calculate is different from the normalization factor geNorm displays The normalization factors calculated by geNorm are subsequently divided by the geometric mean of all normalization factors This additional step is only performed t
14. lysis determination of the optimal number of reference genes for normalization Automated analysis extended FAQ and data requirement section extended and slightly modified error propagation procedure September 6 2004 updated URLs compliant with geNorm version 3 4 bug fix release of 3 3 August 14 2006 FAQ section extended March 13 2007 Minor bug fix geNorm kit section added qBase article reference added compliant with geNorm version 3 5 July 8 2008 explicit mentioning that geNorm does not work in Office2007 minor updates reference to qBaseP us software based on Ghent University s geNorm and qBase technology 14 16 10 References Vandesompele J De Preter K Pattyn F Poppe B Van Roy N De Paepe A Speleman F 2002 Accurate normalization of real time quantitative RT PCR data by geometric averaging of multiple internal control genes Genome Biology 3 34 1 34 11 Vandesompele J De Paepe A Speleman F 2002 Elimination of primer dimer artifacts and genomic coamplification using a two step SYBR green I real time RT PCR Anal Biochem 303 95 98 Hellemans J Mortier G De Paepe A Speleman F Vandesompele J 2007 qBase relative quantification framework and software for management and automated analysis of real time quantitative PCR data Genome Biology 8 R19 http medgen ugent be qbase A growing list of articles citing the geNorm method Vandesompele et al Genome Biology 2002
15. o distribute the normalization factors around value 1 but has NO effect on the net result of your gene of interest due to the relative nature of the expression levels Both normalization factors are equivalent 11 16 Q7 A7 Q8 A9 Q9 A9 When or why would you use the Criteria settings option Two strategies are available to calculate a normalization factor In a first strategy you simply import the raw expression levels of e g 10 reference genes and adjust the expression stability threshold so that a user defined number of genes all with an expression stability value below the threshold are included in the calculation of a normalization factor Doing so you simply exclude a number of genes to take part in the normalization factor A second strategy is partially explained in section 4 It s a stepwise exclusion of the least stable reference gene until you end up with 3 or more if necessary stable reference genes Then you make sure that the stability threshold value is higher than the stability values of the remaining reference genes you intend to use for normalization Doing so all remaining genes are used to calculate the normalization factors We prefer the second strategy as also outlined in the accompanying article Vandesompele et al 2002 Genome Biology What is the difference between the delta delta Ct and delta Ct method The delta delta Ct method transforms Ct values into normalized relative expressi
16. on levels by relating the Ct value of your target gene in your sample to a calibrator control sample AND to the Ct value of a reference gene in both samples Note that in the original publication of the delta delta Ct method Applied Biosystems technical bulletin there s no correction for a difference in amplification efficiency between the target and reference gene only the underlying requirement that the efficiency of target and reference gene should be similar In the delta Ct method you don t use any reference gene you just relate the Ct value of your gene either target or reference to a control calibrator This control calibrator can be any sample e g a real untreated control or the sample with the highest expression lowest Ct value The delta Ct method generates raw not normalized expression values which need to be normalized by dividing with a proper normalization factor Doing 3 times delta delta Ct between your gene of interest and 3 reference genes and then taking the geometric mean of the 3 relative quantifications is the same as first transforming the Ct values of your 4 genes to quantities using delta Ct and dividing the gene of interest by the geometric mean of the reference genes Although both approaches yield the same result I favour the delta Ct method because a it s much easier to do in Excel b it s very easy to take different amplification efficiencies for the different genes into account just replace val
17. reference genes for normalization appears empty This problem generally occurs if your decimal separator settings are not correct see 7 3 for required settings or if you used the copy paste method for loading data into geNorm in contrast to the preferred way of preparing an input file and importing this file into geNorm see 7 4 There is no gene expression stability value for the least stable gene in the Average expression stability of remaining reference genes graph This is caused by not clicking on the Calculation button top left cell of the expression data matrix 13 16 9 Manual Version History July 21 2002 m first version August 19 2002 compliant with geNorm geNorm3 2 successfully tested for the XP version of Microsoft Excel Manual Version History section added References section added September 6 2002 Calculation standard deviation after normalization section added February 17 2003 compliant with newest geNorm version 3 2c bug fix release of 3 2 Save report now works Criteria settings in Excel XP works Show matrix does not display anymore Spearman rank correlation values relic from the first geNorm versions besides to the intended pairwise variation V values which is a much better and robust measure November 10 2003 compliant with geNorm version 3 3 automatic ranking of the reference genes according to their expression stability Automated ana
18. sample Insert row insert sample Lf Delete column remove gene Insert column insert gene Show matrix displays the pairwise variation V values for each gene with all other genes click Return to leave the matrix view S Print Save report iy ll Save input data Automated analysis automatic ranking of reference genes according to their expression stability chart 1 and determination of optimal number of reference genes chart 2 101 E 230m user adjustable zoom level e g allows to view all genes and or samples by highlighting the cells of interest and selecting fit to selection Clear screen About geNorm brief information on contact address method input file menu icons and link to the most recent manual on the web O Exit geNorm quit the application 3 16 4 How to determine the most stable reference genes Manual method 1 2 ul close all running instances of Microsoft Excel start up the geNorm applet in Excel Open File or double click on the geNorm xls file enable macro s when prompted load the expression data matrix raw data this means not yet normalized expression levels see requirements for data file press the Calculate button The M values of the least and most stable genes are now highlighted in red and green respectively To eliminate the gene with the highest M value this is the least stable gene click on the gene name top row and subsequently click the
19. shipped anywhere in the World More details can be found at http www primerdesign co uk geNorm asp PrimerDesign is an innovative biotechnology company founded within Southampton University s School of Medicine and focused on developing improved solutions for gene quantification by real time PCR 1 16 2 Installation Windows version Unzip the downloaded geNorm_3 5 zip file After unzipping a geNorm directory is created which contains the geNorm x s applet and an InputData directory and OutputData directory The InputData directory contains a demo data file fibroblast xls described in Vandesompele et al 2002 Genome Biology and the OutputData directory contains the user manual as a PDF file geNorm requires Microsoft Excel version 2000 XP or 2003 on a Windows platform geNorm does not work in Excel 2007 due to a change in the VBA code base 2 16 3 Menu bar Be ee Se dy om 9 20 A fl Load input data loads Excel data file see Requirements for data format m Manual data input provides possibility to type the data manually indicate the number of samples and reference genes to be analyzed pt Criteria settings adjusts the expression stability threshold below which genes are included in the calculation of a normalization factor genes expression values that are used for normalization are displayed in black while genes in grey inactive are not used to calculate the normalization factors Delete row remove
20. ue 2 with the actual efficiency of the gene e g 1 95 for 95 in the formula of delta Ct and c it allows easy inclusion of multiple reference genes for normalization When using replicated tubes in the same run should I average first the Ct values and then transform to quantities or vice versa Transforming the arithmetic mean Ct value to a quantity is equivalent to transforming each Ct value to a quantity and then calculating the geometric mean of the individual replicate quantities However for determination of the error propagation the first procedure is much more 12 16 Q10 A10 Q11 All Q12 A12 straightforward and therefore is used in the example calculation file on the geNorm web site What is the difference between relative and absolute quantification Both comparative Ct methods delta delta or delta and standard curves based on serial dilutions of a template of which you do not know the exact copy number can be considered as relative quantification methods in which you relate the normalized expression level of one sample to another For absolute quantification you definitely need standard curves based on a template of which your measured the absolute number of molecules It s still a matter of debate however if you can extrapolate copy numbers from a standard dilution most often PCR product or plasmid to the number of molecules in a cDNA sample The graph for Determination of the optimal number of
Download Pdf Manuals
Related Search