LogFC query in Limma
1
0
Entering edit mode
@roopa-subbaiaih-5490
Last seen 9.4 years ago
United States
Hi All, Thanks for the reply, I could pull out the the whole information for differentially expressed genes. The criteria used was adjust="fdr", p=0.05. I came up with ~ 20,000 genes to be differentially expressed. Since I wanted to analyze these genes for deregulated pathways I had to come up with fold change values for further analysis. I assume that for each gene FC= FCmean(test)-FCmean(blank). LogFC is log2 of FC values. When I convert the FC values (test/blank) to foldchanges using IF function I get lesser number of genes to be deregulated. The criteria was =>2 foldchanges and =<-2 fold changes. My question is am I doing it right? Why does the number drastically reduce? Is there a way to do it? Thanks, Roopa On Thu, Jan 31, 2013 at 5:36 AM, Maciej Jończyk <mjonczyk@biol.uw.edu.pl>wrote: > Hi Roopa, > > > results <- decideTests(fit2, adjust="fdr", p=0.05) >> summary(results) >> write.table(results,file="**myresults.txt") >> >> The results table shows ~10,000 genes to be upregulated and ~12,000 genes >> to be down regulated. >> >> My question is how can I get fold change values associated with these >> genes >> in the results file? >> > > You can create object with needed columns binded: > > e.g. > > x=cbind(fit$coefficients,fit$**p.value,p.adjust(fit$p.value,"** > BH"),separate) > > Where "fit" is the result of lmFit and then eBayes commands - it contains > "coefficients" column with (mean) log2 fold change for each gene. > > *In fact it haven't been clear for me at first - but I compared it to the > output of topTable (logFC column) and it is equal.* > > Back to the table - proposed example will give you: > log2 fold change, > raw p-value, > corrected p-value (here Benjamini-Hochberg, equall to "fdr"), > and finally change direction for significant genes (where separate is > result of decideTests) > > Before "write.table" you can change column names with colnames(x)=c(...) > > > HTH > > -- > Maciej Jonczyk, > Department of Plant Molecular Ecophysiology > Faculty of Biology, University of Warsaw > 02-096 Warsaw, Miecznikowa 1 > Poland > > > -- > This email was Anti Virus checked by Astaro Security Gateway. > http://www.astaro.com > [[alternative HTML version deleted]]
Pathways convert Pathways convert • 3.2k views
ADD COMMENT
0
Entering edit mode
@steve-lianoglou-2771
Last seen 13 months ago
United States
Hi, On Thu, Jan 31, 2013 at 2:54 PM, Roopa Subbaiaih <rss115 at="" case.edu=""> wrote: > Hi All, > > Thanks for the reply, I could pull out the the whole information for > differentially expressed genes. The criteria used was adjust="fdr", p=0.05. > I came up with ~ 20,000 genes to be differentially expressed. Hmm ... surely 20k cannot be correct? > Since I wanted to analyze these genes for deregulated pathways I had to > come up with fold change values for further analysis. > > I assume that for each gene FC= FCmean(test)-FCmean(blank). LogFC is log2 > of FC values. > > When I convert the FC values (test/blank) to foldchanges using IF function > I get lesser number of genes to be deregulated. The criteria was =>2 > foldchanges and =<-2 fold changes. I'm missing previous context to this email, so -- not sure what the "IF function" is, but if you're using limma, the log2fold changes are reported for you in the logFC column that is returned from `topTable(...)` -steve -- Steve Lianoglou Graduate Student: Computational Systems Biology | Memorial Sloan-Kettering Cancer Center | Weill Medical College of Cornell University Contact Info: http://cbio.mskcc.org/~lianos/contact
ADD COMMENT
0
Entering edit mode
Hi Steve, This was the script I used- getwd() setwd(dir="/CRSP 406-11/DEMOS/GSE14905-a") ls() #-----------------------------------------------# library(affy) eset = justRMA() f <- factor(c(1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1, 2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2), labels=c("Healthy", "unaffected")) design <- model.matrix(~ 0 + f) design colnames(design) <-c("Healthy","unaffected") design library(limma) fit <- lmFit(eset, design) library(hgu133plus2.db) fit$genes$Symbol <- getSYMBOL(fit$genes$ID,"hgu133plus2.db") contrast.matrix <-makeContrasts(affected-Healthy,levels = design) fit2 <- contrasts.fit(fit, contrast.matrix) fit2 <- eBayes(fit2) topTable(fit2,coef=1,p=0.05, adjust="fdr") results <- decideTests(fit2, adjust="fdr", p=0.05) summary(results) write.table(results,file="myresults.txt") write.fit(). I had identified ~54,000 genes of which ~ 20K were differentially expressed. But when I use these genes for pathway analysis the software asks for fold change values but not p value so it is easier to analyze the data. What I did was - I used the differentially expressed gene table for further analysis. That is I converted logFC values to FC(test/control) assuming that FC= FCmean(test)-FCmean(blank) and LogFC is log2 of FC values. Once I got test/control values I converted them to fold changes using "IF" function in excel sheet to eliminate genes with fold changes between -2 to +2. Once I did this the number of significant genes drastically reduced to ~ 2,000. Is this the right method? Please advice, thanks, Roopa On Thu, Jan 31, 2013 at 3:23 PM, Steve Lianoglou < mailinglist.honeypot@gmail.com> wrote: > Hi, > > On Thu, Jan 31, 2013 at 2:54 PM, Roopa Subbaiaih <rss115@case.edu> wrote: > > Hi All, > > > > Thanks for the reply, I could pull out the the whole information for > > differentially expressed genes. The criteria used was adjust="fdr", > p=0.05. > > I came up with ~ 20,000 genes to be differentially expressed. > > Hmm ... surely 20k cannot be correct? > > > Since I wanted to analyze these genes for deregulated pathways I had to > > come up with fold change values for further analysis. > > > > I assume that for each gene FC= FCmean(test)-FCmean(blank). LogFC is log2 > > of FC values. > > > > When I convert the FC values (test/blank) to foldchanges using IF > function > > I get lesser number of genes to be deregulated. The criteria was =>2 > > foldchanges and =<-2 fold changes. > > I'm missing previous context to this email, so -- not sure what the > "IF function" is, but if you're using limma, the log2fold changes are > reported for you in the logFC column that is returned from > `topTable(...)` > > -steve > > -- > Steve Lianoglou > Graduate Student: Computational Systems Biology > | Memorial Sloan-Kettering Cancer Center > | Weill Medical College of Cornell University > Contact Info: http://cbio.mskcc.org/~lianos/contact > -- --------------------------------------- Roopa Shree Subbaiaih Post Doctoral Fellow Department of Dermatology School of Medicine Case Western Reserve University Cleveland, OH-44106 Tel:+1 216 368 0211 [[alternative HTML version deleted]]
ADD REPLY
0
Entering edit mode
Hi Roopa, On 1/31/2013 3:45 PM, Roopa Subbaiaih wrote: > Hi Steve, > > This was the script I used- > getwd() > setwd(dir="/CRSP 406-11/DEMOS/GSE14905-a") > ls() > #-----------------------------------------------# > library(affy) > eset = justRMA() > f<- factor(c(1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1, > > 2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2), > labels=c("Healthy", "unaffected")) > design<- model.matrix(~ 0 + f) > design > colnames(design)<-c("Healthy","unaffected") > design > library(limma) > fit<- lmFit(eset, design) > library(hgu133plus2.db) > fit$genes$Symbol<- getSYMBOL(fit$genes$ID,"hgu133plus2.db") > contrast.matrix<-makeContrasts(affected-Healthy,levels = design) > fit2<- contrasts.fit(fit, contrast.matrix) > fit2<- eBayes(fit2) > topTable(fit2,coef=1,p=0.05, adjust="fdr") > results<- decideTests(fit2, adjust="fdr", p=0.05) > summary(results) > write.table(results,file="myresults.txt") > write.fit(). > > I had identified ~54,000 genes of which ~ 20K were differentially expressed. > > But when I use these genes for pathway analysis the software asks for fold > change values but not p value so it is easier to analyze the data. > > What I did was - I used the differentially expressed gene table for further > analysis. That is I converted logFC values to FC(test/control) assuming > that > > FC= FCmean(test)-FCmean(blank) and LogFC is log2 of FC values. > > Once I got test/control values I converted them to fold changes using "IF" > function in excel sheet to eliminate genes with fold changes between -2 to > +2. > > Once I did this the number of significant genes drastically reduced to ~ > 2,000. > > Is this the right method? No. Note that the range of fold changes after 'unlogging' will be 0-INF, and the down-regulated genes will be in the range 0-1 whereas the upregulated genes will be in the range 1-INF. (e.g. two fold up will be 2, whereas 2 fold down will be 1/2 or 0.5). The easiest way to filter is to keep the logFC and filter on -1 and 1. Or you can use the lfc argument to decideTests(). Best, Jim > > Please advice, thanks, Roopa > > On Thu, Jan 31, 2013 at 3:23 PM, Steve Lianoglou< > mailinglist.honeypot at gmail.com> wrote: > >> Hi, >> >> On Thu, Jan 31, 2013 at 2:54 PM, Roopa Subbaiaih<rss115 at="" case.edu=""> wrote: >>> Hi All, >>> >>> Thanks for the reply, I could pull out the the whole information for >>> differentially expressed genes. The criteria used was adjust="fdr", >> p=0.05. >>> I came up with ~ 20,000 genes to be differentially expressed. >> Hmm ... surely 20k cannot be correct? >> >>> Since I wanted to analyze these genes for deregulated pathways I had to >>> come up with fold change values for further analysis. >>> >>> I assume that for each gene FC= FCmean(test)-FCmean(blank). LogFC is log2 >>> of FC values. >>> >>> When I convert the FC values (test/blank) to foldchanges using IF >> function >>> I get lesser number of genes to be deregulated. The criteria was =>2 >>> foldchanges and =<-2 fold changes. >> I'm missing previous context to this email, so -- not sure what the >> "IF function" is, but if you're using limma, the log2fold changes are >> reported for you in the logFC column that is returned from >> `topTable(...)` >> >> -steve >> >> -- >> Steve Lianoglou >> Graduate Student: Computational Systems Biology >> | Memorial Sloan-Kettering Cancer Center >> | Weill Medical College of Cornell University >> Contact Info: http://cbio.mskcc.org/~lianos/contact >> > > -- James W. MacDonald, M.S. Biostatistician University of Washington Environmental and Occupational Health Sciences 4225 Roosevelt Way NE, # 100 Seattle WA 98105-6099
ADD REPLY
0
Entering edit mode
... what Jim said. But also, this 20k differentially expressed (likely probe sets, not genes) is raising a red flag for me, no? Am I alone here? That's .. what's the word I'm looking for ... "a lot". -steve On Thu, Jan 31, 2013 at 3:56 PM, James W. MacDonald <jmacdon at="" uw.edu=""> wrote: > Hi Roopa, > > > On 1/31/2013 3:45 PM, Roopa Subbaiaih wrote: >> >> Hi Steve, >> >> This was the script I used- >> getwd() >> setwd(dir="/CRSP 406-11/DEMOS/GSE14905-a") >> ls() >> #-----------------------------------------------# >> library(affy) >> eset = justRMA() >> f<- factor(c(1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1, >> >> 2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2), >> labels=c("Healthy", "unaffected")) >> design<- model.matrix(~ 0 + f) >> design >> colnames(design)<-c("Healthy","unaffected") >> design >> library(limma) >> fit<- lmFit(eset, design) >> library(hgu133plus2.db) >> fit$genes$Symbol<- getSYMBOL(fit$genes$ID,"hgu133plus2.db") >> contrast.matrix<-makeContrasts(affected-Healthy,levels = design) >> fit2<- contrasts.fit(fit, contrast.matrix) >> fit2<- eBayes(fit2) >> topTable(fit2,coef=1,p=0.05, adjust="fdr") >> results<- decideTests(fit2, adjust="fdr", p=0.05) >> summary(results) >> write.table(results,file="myresults.txt") >> write.fit(). >> >> I had identified ~54,000 genes of which ~ 20K were differentially >> expressed. >> >> But when I use these genes for pathway analysis the software asks for fold >> change values but not p value so it is easier to analyze the data. >> >> What I did was - I used the differentially expressed gene table for >> further >> analysis. That is I converted logFC values to FC(test/control) assuming >> that >> >> FC= FCmean(test)-FCmean(blank) and LogFC is log2 of FC values. >> >> Once I got test/control values I converted them to fold changes using "IF" >> function in excel sheet to eliminate genes with fold changes between -2 to >> +2. >> >> Once I did this the number of significant genes drastically reduced to ~ >> 2,000. >> >> Is this the right method? > > > No. Note that the range of fold changes after 'unlogging' will be 0-INF, and > the down-regulated genes will be in the range 0-1 whereas the upregulated > genes will be in the range 1-INF. (e.g. two fold up will be 2, whereas 2 > fold down will be 1/2 or 0.5). > > The easiest way to filter is to keep the logFC and filter on -1 and 1. Or > you can use the lfc argument to decideTests(). > > Best, > > Jim > > > >> >> Please advice, thanks, Roopa >> >> On Thu, Jan 31, 2013 at 3:23 PM, Steve Lianoglou< >> mailinglist.honeypot at gmail.com> wrote: >> >>> Hi, >>> >>> On Thu, Jan 31, 2013 at 2:54 PM, Roopa Subbaiaih<rss115 at="" case.edu=""> wrote: >>>> >>>> Hi All, >>>> >>>> Thanks for the reply, I could pull out the the whole information for >>>> differentially expressed genes. The criteria used was adjust="fdr", >>> >>> p=0.05. >>>> >>>> I came up with ~ 20,000 genes to be differentially expressed. >>> >>> Hmm ... surely 20k cannot be correct? >>> >>>> Since I wanted to analyze these genes for deregulated pathways I had to >>>> come up with fold change values for further analysis. >>>> >>>> I assume that for each gene FC= FCmean(test)-FCmean(blank). LogFC is >>>> log2 >>>> of FC values. >>>> >>>> When I convert the FC values (test/blank) to foldchanges using IF >>> >>> function >>>> >>>> I get lesser number of genes to be deregulated. The criteria was =>2 >>>> foldchanges and =<-2 fold changes. >>> >>> I'm missing previous context to this email, so -- not sure what the >>> "IF function" is, but if you're using limma, the log2fold changes are >>> reported for you in the logFC column that is returned from >>> `topTable(...)` >>> >>> -steve >>> >>> -- >>> Steve Lianoglou >>> Graduate Student: Computational Systems Biology >>> | Memorial Sloan-Kettering Cancer Center >>> | Weill Medical College of Cornell University >>> Contact Info: http://cbio.mskcc.org/~lianos/contact >>> >> >> > > -- > James W. MacDonald, M.S. > Biostatistician > University of Washington > Environmental and Occupational Health Sciences > 4225 Roosevelt Way NE, # 100 > Seattle WA 98105-6099 > -- Steve Lianoglou Graduate Student: Computational Systems Biology | Memorial Sloan-Kettering Cancer Center | Weill Medical College of Cornell University Contact Info: http://cbio.mskcc.org/~lianos/contact
ADD REPLY
0
Entering edit mode
What you said was true. They are probe sets. When I upload the information into IPA 54k probe sets gives ~18k genes identified. I was concerned with fold changes for further analysis. Thanks, Roopa On Thu, Jan 31, 2013 at 4:02 PM, Steve Lianoglou < mailinglist.honeypot@gmail.com> wrote: > ... what Jim said. > > But also, this 20k differentially expressed (likely probe sets, not > genes) is raising a red flag for me, no? Am I alone here? > > That's .. what's the word I'm looking for ... "a lot". > > -steve > > On Thu, Jan 31, 2013 at 3:56 PM, James W. MacDonald <jmacdon@uw.edu> > wrote: > > Hi Roopa, > > > > > > On 1/31/2013 3:45 PM, Roopa Subbaiaih wrote: > >> > >> Hi Steve, > >> > >> This was the script I used- > >> getwd() > >> setwd(dir="/CRSP 406-11/DEMOS/GSE14905-a") > >> ls() > >> #-----------------------------------------------# > >> library(affy) > >> eset = justRMA() > >> f<- factor(c(1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1, > >> > >> 2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2), > >> labels=c("Healthy", "unaffected")) > >> design<- model.matrix(~ 0 + f) > >> design > >> colnames(design)<-c("Healthy","unaffected") > >> design > >> library(limma) > >> fit<- lmFit(eset, design) > >> library(hgu133plus2.db) > >> fit$genes$Symbol<- getSYMBOL(fit$genes$ID,"hgu133plus2.db") > >> contrast.matrix<-makeContrasts(affected-Healthy,levels = design) > >> fit2<- contrasts.fit(fit, contrast.matrix) > >> fit2<- eBayes(fit2) > >> topTable(fit2,coef=1,p=0.05, adjust="fdr") > >> results<- decideTests(fit2, adjust="fdr", p=0.05) > >> summary(results) > >> write.table(results,file="myresults.txt") > >> write.fit(). > >> > >> I had identified ~54,000 genes of which ~ 20K were differentially > >> expressed. > >> > >> But when I use these genes for pathway analysis the software asks for > fold > >> change values but not p value so it is easier to analyze the data. > >> > >> What I did was - I used the differentially expressed gene table for > >> further > >> analysis. That is I converted logFC values to FC(test/control) assuming > >> that > >> > >> FC= FCmean(test)-FCmean(blank) and LogFC is log2 of FC values. > >> > >> Once I got test/control values I converted them to fold changes using > "IF" > >> function in excel sheet to eliminate genes with fold changes between -2 > to > >> +2. > >> > >> Once I did this the number of significant genes drastically reduced to ~ > >> 2,000. > >> > >> Is this the right method? > > > > > > No. Note that the range of fold changes after 'unlogging' will be 0-INF, > and > > the down-regulated genes will be in the range 0-1 whereas the upregulated > > genes will be in the range 1-INF. (e.g. two fold up will be 2, whereas 2 > > fold down will be 1/2 or 0.5). > > > > The easiest way to filter is to keep the logFC and filter on -1 and 1. Or > > you can use the lfc argument to decideTests(). > > > > Best, > > > > Jim > > > > > > > >> > >> Please advice, thanks, Roopa > >> > >> On Thu, Jan 31, 2013 at 3:23 PM, Steve Lianoglou< > >> mailinglist.honeypot@gmail.com> wrote: > >> > >>> Hi, > >>> > >>> On Thu, Jan 31, 2013 at 2:54 PM, Roopa Subbaiaih<rss115@case.edu> > wrote: > >>>> > >>>> Hi All, > >>>> > >>>> Thanks for the reply, I could pull out the the whole information for > >>>> differentially expressed genes. The criteria used was adjust="fdr", > >>> > >>> p=0.05. > >>>> > >>>> I came up with ~ 20,000 genes to be differentially expressed. > >>> > >>> Hmm ... surely 20k cannot be correct? > >>> > >>>> Since I wanted to analyze these genes for deregulated pathways I had > to > >>>> come up with fold change values for further analysis. > >>>> > >>>> I assume that for each gene FC= FCmean(test)-FCmean(blank). LogFC is > >>>> log2 > >>>> of FC values. > >>>> > >>>> When I convert the FC values (test/blank) to foldchanges using IF > >>> > >>> function > >>>> > >>>> I get lesser number of genes to be deregulated. The criteria was =>2 > >>>> foldchanges and =<-2 fold changes. > >>> > >>> I'm missing previous context to this email, so -- not sure what the > >>> "IF function" is, but if you're using limma, the log2fold changes are > >>> reported for you in the logFC column that is returned from > >>> `topTable(...)` > >>> > >>> -steve > >>> > >>> -- > >>> Steve Lianoglou > >>> Graduate Student: Computational Systems Biology > >>> | Memorial Sloan-Kettering Cancer Center > >>> | Weill Medical College of Cornell University > >>> Contact Info: http://cbio.mskcc.org/~lianos/contact > >>> > >> > >> > > > > -- > > James W. MacDonald, M.S. > > Biostatistician > > University of Washington > > Environmental and Occupational Health Sciences > > 4225 Roosevelt Way NE, # 100 > > Seattle WA 98105-6099 > > > > > > -- > Steve Lianoglou > Graduate Student: Computational Systems Biology > | Memorial Sloan-Kettering Cancer Center > | Weill Medical College of Cornell University > Contact Info: http://cbio.mskcc.org/~lianos/contact > -- --------------------------------------- Roopa Shree Subbaiaih Post Doctoral Fellow Department of Dermatology School of Medicine Case Western Reserve University Cleveland, OH-44106 Tel:+1 216 368 0211 [[alternative HTML version deleted]]
ADD REPLY
0
Entering edit mode
If you just use the expression values from the original authors, I get just under 9K probesets for this comparison at an FDR of 0.05 and no fold change criterion. It drops to just under 900 with a 2-fold difference added in. So yeah, seems like a lot to me as well. Best, Jim On 1/31/2013 4:02 PM, Steve Lianoglou wrote: > ... what Jim said. > > But also, this 20k differentially expressed (likely probe sets, not > genes) is raising a red flag for me, no? Am I alone here? > > That's .. what's the word I'm looking for ... "a lot". > > -steve > > On Thu, Jan 31, 2013 at 3:56 PM, James W. MacDonald<jmacdon at="" uw.edu=""> wrote: >> Hi Roopa, >> >> >> On 1/31/2013 3:45 PM, Roopa Subbaiaih wrote: >>> Hi Steve, >>> >>> This was the script I used- >>> getwd() >>> setwd(dir="/CRSP 406-11/DEMOS/GSE14905-a") >>> ls() >>> #-----------------------------------------------# >>> library(affy) >>> eset = justRMA() >>> f<- factor(c(1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1, >>> >>> 2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2), >>> labels=c("Healthy", "unaffected")) >>> design<- model.matrix(~ 0 + f) >>> design >>> colnames(design)<-c("Healthy","unaffected") >>> design >>> library(limma) >>> fit<- lmFit(eset, design) >>> library(hgu133plus2.db) >>> fit$genes$Symbol<- getSYMBOL(fit$genes$ID,"hgu133plus2.db") >>> contrast.matrix<-makeContrasts(affected-Healthy,levels = design) >>> fit2<- contrasts.fit(fit, contrast.matrix) >>> fit2<- eBayes(fit2) >>> topTable(fit2,coef=1,p=0.05, adjust="fdr") >>> results<- decideTests(fit2, adjust="fdr", p=0.05) >>> summary(results) >>> write.table(results,file="myresults.txt") >>> write.fit(). >>> >>> I had identified ~54,000 genes of which ~ 20K were differentially >>> expressed. >>> >>> But when I use these genes for pathway analysis the software asks for fold >>> change values but not p value so it is easier to analyze the data. >>> >>> What I did was - I used the differentially expressed gene table for >>> further >>> analysis. That is I converted logFC values to FC(test/control) assuming >>> that >>> >>> FC= FCmean(test)-FCmean(blank) and LogFC is log2 of FC values. >>> >>> Once I got test/control values I converted them to fold changes using "IF" >>> function in excel sheet to eliminate genes with fold changes between -2 to >>> +2. >>> >>> Once I did this the number of significant genes drastically reduced to ~ >>> 2,000. >>> >>> Is this the right method? >> >> No. Note that the range of fold changes after 'unlogging' will be 0-INF, and >> the down-regulated genes will be in the range 0-1 whereas the upregulated >> genes will be in the range 1-INF. (e.g. two fold up will be 2, whereas 2 >> fold down will be 1/2 or 0.5). >> >> The easiest way to filter is to keep the logFC and filter on -1 and 1. Or >> you can use the lfc argument to decideTests(). >> >> Best, >> >> Jim >> >> >> >>> Please advice, thanks, Roopa >>> >>> On Thu, Jan 31, 2013 at 3:23 PM, Steve Lianoglou< >>> mailinglist.honeypot at gmail.com> wrote: >>> >>>> Hi, >>>> >>>> On Thu, Jan 31, 2013 at 2:54 PM, Roopa Subbaiaih<rss115 at="" case.edu=""> wrote: >>>>> Hi All, >>>>> >>>>> Thanks for the reply, I could pull out the the whole information for >>>>> differentially expressed genes. The criteria used was adjust="fdr", >>>> p=0.05. >>>>> I came up with ~ 20,000 genes to be differentially expressed. >>>> Hmm ... surely 20k cannot be correct? >>>> >>>>> Since I wanted to analyze these genes for deregulated pathways I had to >>>>> come up with fold change values for further analysis. >>>>> >>>>> I assume that for each gene FC= FCmean(test)-FCmean(blank). LogFC is >>>>> log2 >>>>> of FC values. >>>>> >>>>> When I convert the FC values (test/blank) to foldchanges using IF >>>> function >>>>> I get lesser number of genes to be deregulated. The criteria was =>2 >>>>> foldchanges and =<-2 fold changes. >>>> I'm missing previous context to this email, so -- not sure what the >>>> "IF function" is, but if you're using limma, the log2fold changes are >>>> reported for you in the logFC column that is returned from >>>> `topTable(...)` >>>> >>>> -steve >>>> >>>> -- >>>> Steve Lianoglou >>>> Graduate Student: Computational Systems Biology >>>> | Memorial Sloan-Kettering Cancer Center >>>> | Weill Medical College of Cornell University >>>> Contact Info: http://cbio.mskcc.org/~lianos/contact >>>> >>> >> -- >> James W. MacDonald, M.S. >> Biostatistician >> University of Washington >> Environmental and Occupational Health Sciences >> 4225 Roosevelt Way NE, # 100 >> Seattle WA 98105-6099 >> > > -- James W. MacDonald, M.S. Biostatistician University of Washington Environmental and Occupational Health Sciences 4225 Roosevelt Way NE, # 100 Seattle WA 98105-6099
ADD REPLY

Login before adding your answer.

Traffic: 637 users visited in the last hour
Help About
FAQ
Access RSS
API
Stats

Use of this site constitutes acceptance of our User Agreement and Privacy Policy.

Powered by the version 2.3.6