[BioC] Help on experiment design

Mark Cowley m.cowley at garvan.org.au
Thu Aug 19 06:29:51 CEST 2010


hi Marcelo,

If you use that design matrix, then you'll need to also fit contrasts
> designEsp <- cbind(controle=c(1,1,1,0,0,0),experimento=c(0,0,0,1,1,1))
> fitEsp <- lmFit(exprs(summarized),designEsp)
contr.matrix <- c(-1,1)
fitEsp <- contrasts.fit(fitEsp, contr.matrix)
> fitEsp <- eBayes(fitEsp)
> topTable(fitEsp,coef=1, adjust = "fdr", n = 10)

or use this design
> designEsp2 <- cbind(intercept=c(1,1,1,1,1,1),expvscon=c(0,0,0,1,1,1))
> fitEsp2 <- lmFit(exprs(summarized),designEsp2)
> fitEsp2 <- eBayes(fitEsp2)
> topTable(fitEsp2,coef=2, adjust = "fdr", n = 10)

cheers,
Mark
On 19/08/2010, at 10:59 AM, Marcelo Brandão wrote:

> Hello all!
> I am analyzing a set with 6 microarrays, 3 controls and 3
> experimentals. I am interested in look after differentially expressed
> genes. My doubt, among a lot of else, is if my experiment is designed
> correctly. I am currently using the following comands:
> 
> designEsp <- cbind(controle=c(1,1,1,0,0,0),experimento=c(0,0,0,1,1,1))
> fitEsp <- lmFit(exprs(summarized),designEsp)
> fitEsp <- eBayes(fitEsp)
> topTable(fitEsp,coef=2, adjust = "fdr", n = 10)
> 
> is it the best way to infer different expression? I am using nimblegen
> microarrays.
> 
> Thanks in advance.
> 
> Marcelo
> 
> -- 
> Marcelo Mendes Brandão
> Postdoc fellow
> Laboratório de Biologia Molecular de Plantas - ESALQ/USP
> Website: http://bioinfo.esalq.usp.br
> AtPIN: http://bioinfo.esalq.usp.br/atpin
> SKYPE: mmbrand
> Tel: (+55) 19 3429 4442
> 
> _______________________________________________
> 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



More information about the Bioconductor mailing list