Hi to all,
Simple question for DiffBind that isn't clear to me despite a lot of searching and (over) thinking.
How exactly is the count data treated when one invokes
db.analyze(method=DESeq2) using default parameters in current release?
My understanding is that some years ago the devs set
bFullLibrarySize=TRUE as default to guard against violation of the assumption that the majority of enrichment regions were similar between groups. This has not been updated in the technical notes section 7.3 of the user guide. Therefore, currently, does
db.analyze(method=DESeq2) only scale on depth of library, or is there a subsequent TMM normalization on these scaled counts?
Thank you for any responses!
I realize there's no hard rules and you would need data to make a real suggestion, but is simply scaling counts by library-size be considered sufficient normalization or would you recommend TMM / RPKM required in addition (or instead of)?
All my work in DiffBind uses the default scoring system