[BioC] edgeR: very low p-value and very high variance within the group of replicates

Simon Anders anders at embl.de
Wed Jun 19 20:45:01 CEST 2013

Hi Valentina

> I obtained a very odd results: in some cases I had a very low p-value
> and FDR but looking at the raw data it is obvious that the difference
> between the two groups can't be significant.

Maybe also give DESeq2 a try. We recently implemented some changes to 
improve its robustness against outliers of the kind that you are 

Don't be too surprised, though, if you don't find any useful hits: 
Usually, it takes samples from dozens of patients to find genes which 
gene which differ between sub-groups with different prognosis. With only 
6 samples, you are severely underpowered.


More information about the Bioconductor mailing list