EdgeR or Ttest/ANOVA for non normal RNAseq data?

Question

[datanerd]

HI All,

I have RNAseq data and analyzed it with both EdgeR (Exactest) and usual ANOVA (on library normalized CPM count matrix)

However my data is not normal so Iam worried about using ANOVA.

Also, I ran EdgeR as in the tutorial (do not know much how to inset design here). But what I observe is (i have loss of function disease group and I know it should be dramatically downregulated in the disease group as compared to Normal) that this mRNA as expected in the ANOVA results but not in Exacttest (tried poisson, tagwise, common dispersion).

Would anyone recommend what test to use on normalized CPM matrix if not excattest?

Thanks in advance!!!

Mamta

[eric.kern13]

How many rows of data do you have? Linear models such as ANOVA rely heavily on independence of errors and constant variability, but with enough data, normality is not as crucial.