I am looking for critique on the methods I've used to find DEGs in my data.
I initially used Ebayes with 'BH' FDR adjustment but found this stringency wasn't enough. I switched to trying bonferroni but had issues with my decidetests and toptable results not matching?
dt <-decideTests(efit, method = "global", adjust.method = "bonferroni", p.value = 0.01)
Degs <- topTable(efit, number = Inf, adjust.method = "bonferroni", p.value = 0.01)
When I subset the write.fit(efit, dt) in regards to the toptable results its apparent that there are many genes being selected by TopTable that have all 0's in the decidetests.
Instead of using TopTable should I just subset by rowsums!=0 from the decidetests?