I have an RNA Seq experiment looking at the impact of a stressor on several fish species. The analysis in edgeR went smoothly in all but one species where I could only identify a few deferentially expressed genes. This fish has been shown to show muted transcriptional responses in the past and it's not a complete surprise to see a small response in this experiment, but looking at the p-value histogram showed a rising hill (shown in the link below) rather than the flat distribution that I would traditionally expect :
Is this the sign of a missing covariate? Outside of gender (which attempted to block on without impact) I'm hard pressed to think of any. The PCA for this particular species showed no clustering at all and nothing that would suggest another impact. Any advice on how to troubleshoot the analysis would be appreciated.
thanks in advance,