Hi Simon,

For edgeR I was using this code :

rawdata<-read.delim(file="ALL_MTcD_9.5_NO_MTsD1",header=TRUE,row.names=1)
y<-DGEList(counts=rawdata,genes=rawdata[,0])
y<-calcNormFactors(y)
plotMDS(y,xlab="Mouse factor",ylab="Treatment Factor",xlim=c(-0.4,0.4),ylim=c(-0.4,0.4))

Mouse<-factor(c("WT","WT","WT","WT","WT","WT","MT","MT","MT","MT","MT"))
Treatment<-factor(c("Untreated","Untreated","Untreated","Treated","Treated","Treated","Untreated","Untreated","Treated","Treated","Treated"))

data.frame(Sample=colnames(y),Mouse,Treatment)
design<-model.matrix(~Mouse+Treatment+Mouse:Treatment)
rownames(design)<-colnames(y)

#Overall dispersion of the dataset
y<-estimateGLMCommonDisp(y,design,verbose=TRUE)
#Estimation of the gene-wise dispersion
y<-estimateGLMTrendedDisp(y,design)
y<-estimateGLMTagwiseDisp(y,design)
#Diff expresion
fit<-glmFit(y,design)
lrt<-glmLRT(fit,contrast=c(1,1-1))
top<-topTags(lrt,n=100)
head(top$table)

but it gives me an error :

lrt<-glmLRT(fit,contrast=c(1,1-1))
Error in glmfit$coefficients %*% contrast : non-conformable arguments



On 4 Jun 2013, at 12:18, "Simon Anders" <anders@embl.de> wrote:

> Hi Eduardo
> 
> On 04/06/13 12:13, Eduardo Andrés León wrote:
>> About ANOVA, I did some experiments with the edgeR anova-like test using
>> the glmLRT function and the coef parameter.
>> 
>> But as usual, both results were quite different.
> 
> Just to clarify: What you do with DESeq's GLM facilities is also ANOVA (or actually, both edgeR and DESeq do ANODEV, but this is just a technicality), and you should get roughly similar results. Are you sure you tested _for_the_interaction_ with edgeR?
> 
>> About the next step, the time course analysis, can this be done with
>> DESeq, using the time as a new covariate (as with the drug) ?
> 
> In principle, yes. But now things get really complicated because you need to choose which time points to compare (as you have more than two).
> 
>  Simon
> 

===================================================
Eduardo Andrés León
Tlfn: (+34) 91 732 80 00 / 91 224 69 00 (ext 5054/3063)
e-mail: eandres@cnio.es        Fax: (+34) 91 224 69 76
Unidad de Bioinformática       Bioinformatics Unit
Centro Nacional de Investigaciones Oncológicas
C.P.: 28029                Zip Code: 28029
C/. Melchor Fernández Almagro, 3    Madrid (Spain)
http://bioinfo.cnio.es	http://bioinfo.cnio.es/people/eandres
===================================================


	[[alternative HTML version deleted]]