[BioC] limma design for tech and biol replicates (1 color design)

[Gordon K Smyth]

Dear Florence,

Your code looks to be correct.

Best wishes
Gordon

> Date: Fri, 24 Jun 2011 15:50:27 +0200
> From: Florence Combes <fcombes at gmail.com>
> To: Bioconductor <bioconductor at stat.math.ethz.ch>
> Subject: [BioC] limma design for tech and biol replicate (1 color
> 	design)
>
> Dear Bioconductor mailing-list,
>
>
> I tried to analyze one-color experiments with limma, but I am really not
> sure about the way to define tech and biol replicates.
> Any answer very welcome for me to know of my results are correct or not ...
>
> The experiment is as follow :
> I have a mutant (mut) and a wild type (wt) strain x 2 biol. rep x 4
> technical replicates
> So I have :
> mut1.1 ; mut1.2 ;  mut1.3 ; mut1.4 ; mut2.1 ; mut2.2 ; mut2.3 ; mut2.4 ;
> wt1.1  ; wt1.2 ;  wt1.3 ; wt1.4 ; wt2.1 ; wt2.2 ; wt2.3 ; wt2.4
>
> So I have :
> - 2 (biol rep) mutants : mut1 and mut2, with 4 tech rep each => 4x2=8 arrays
> - and, similarly : 2 (biol rep)  wt : wt1 and wt2, with also 4 tech rep each
> => 4x2=8 arrays
>
> ==> 16 arrays, one color design.
>
> I (very classically) want to know what is the most significantly different
> between the mutant and the wt.
>
> My limma code is this one :
>
> #1 and 2 are biol rep for the mutant, and 3 and 4 are biol rep for the wt
> # I am not sure this is the right way to code this properly ?
> biolrep<-c(1,1,1,1,2,2,2,2,3,3,3,3,4,4,4,4)
> design<-cbind(MU=c(1,1,1,1,1,1,1,1,0,0,0,0,0,0,0,0),
> WT=c(0,0,0,0,0,0,0,0,1,1,1,1,1,1,1,1))
> corfit<-duplicateCorrelation(my.data, design, ndups=1, block=biolrep)
> #corfit$consensus = 0.7658993
> fit<-lmFit(my.data, design, block=biolrep, cor=corfit$consensus)
> cont.matrix<-makeContrasts(MUvsWT=MU-WT, levels=design)
> fit2<-contrasts.fit(fit, cont.matrix)
> fit2<-eBayes(fit2)
> anadiff.limma<-topTable(fit2, adjust="BH")
>
>
> Am I right ?
>
> Again, thanks a lot for your help,
>
> Florence.

______________________________________________________________________
The information in this email is confidential and intend...{{dropped:4}}