[BioC] limma: results from different coefficients

[James W. MacDonald]

kfbargad at ehu.es wrote:
> Dear list,
> 
> I have three groups of samples (Tto, NoTto, C) and have made the three 
> possible pairwise contrasts as follows:
> 
> design.tto <- model.matrix(~-1+factor(c
> (1,1,2,2,2,3,2,1,3,1,2,2,3,1,1,2,1,1,3,1)))
> colnames(design.tto) <- c("C","Tto", "NoTto")
> 
> fit.tto <- lmFit(eset,design.tto)
> contrast.matrix.tto <- makeContrasts(Tto-C,NoTto-C, Tto-NoTto, levels 
> = design.tto)
> fit2.tto <- contrasts.fit(fit.tto, contrast.matrix.tto)
> fit3.tto <-eBayes(fit2.tto)
> 
> If I select all d.e (p<0.05) genes from each contrast of interest I 
> get the same number of genes, 463, although different

Do you mean that you get the same number of genes, but they aren't 
necessarily the same genes?

> 
> topTable(fit3.tto, coef = X(where X =1,2 or 3)#, adjust="none", 
> sort.by="P", number=5506); NoTto_vs_C_y <- NoTto_vs_C[x$P.Value < 
> 0.05,]

I assume this is just a snippet of code intended to give the general 
idea of what you have done? If so, is it possible that your actual code 
did the same exact thing three times?

> 
> dim(Tto_vs_C_y)
> [1] 463   7
> 
>>dim(NoTto_vs_C_y)
> 
> [1] 463   7
> 
>>dim(Tto_vs_NoTto_y)
> 
> [1] 463   7
> 
> 
> Is this normal?

I wouldn't say it is normal, but it certainly is possible. These days 
limma has code to do exactly what you have done that you can use to 
check your results.

rslt <- decideTests(fit3.tto, adjust="n")
vennDiagram(rslt)

HTH,

Jim


> 
> Thanks in advance
> 
> David
> 
> _______________________________________________
> Bioconductor mailing list
> Bioconductor at stat.math.ethz.ch
> https://stat.ethz.ch/mailman/listinfo/bioconductor
> Search the archives: http://news.gmane.org/gmane.science.biology.informatics.conductor


-- 
James W. MacDonald
University of Michigan
Affymetrix and cDNA Microarray Core
1500 E Medical Center Drive
Ann Arbor MI 48109
734-647-5623



**********************************************************
Electronic Mail is not secure, may not be read every day, and should not be used for urgent or sensitive issues.