I have a project with 20 samples of one tissue, and 20 of another.  (also, there are two different genotypes, and 2 different treatments) I was wondering if I should be keeping them together as a single DEseq object if the investigators are not really interested in comparing between tissues.  I was concerned if it was likely that the library normalization step would be compromised if expression was too different between the tissues.  In general, should I be putting different kinds of tissue together, or should be keeping them apart?  Is there some metric I should be examining to know when the tissue types are okay to keep together?

This is a FAQ in the vignette. We have some recommendations for how to make the decision