View Single Post
Old 08-07-2013, 05:53 AM   #2
dpryan
Devon Ryan
 
Location: Freiburg, Germany

Join Date: Jul 2011
Posts: 3,480
Default

You could use DESeq or edgeR still and just not do the pair-wise comparisons. Using DESeq syntax:
Code:
fit1 <- fitNbinomGLMs(cds, count~treatment*time)
fit0 <- fitNbinomGLMs(cds, count~time + treatment:time)
pvals <- nbinomGLMTest(fit1, fit0)
...assuming you're interested in genes that are differentially expressed by treatment when accounting for time point and a possible time-point:treatment interaction.
dpryan is offline   Reply With Quote