Hi,

Since edgeR appears to be designed to deal with count values, I was wondering if it could be applied to roughly detect differences in gene counts in gene families of two groups, for example (numbers below represent counts):

I'm assuming that each species in such matrix could be considered a replicate within each group, that gene families would be the equivalent to loci, and counts are absolute.

Any suggestion/advice are will be welcome.

Eliza

In principle, edgeR could be applied here. However we tend to find that taxa abundances have more technical variability than the negative binomial model would imply. I don't know the best method. We analysed microbiome abundances in a recent paper, so you can read in the methods what we did:

https://microbiomejournal.biomedcentral.com/articles/10.1186/s40168-021-01104-y