Entering edit mode
Given an OTU table (rows are microbiota and columns are samples) and I want to compare treatment vs placebo sample groups using DESeq2. It is not clear to me what counts I should supply to DESeq2: raw counts or the percentages?
Note, this is not about gene expression, I understand I would use raw counts for gene expression. This question is specifically about microbiome data (I have 16S data as well as other tables from whole genome shotgun metagenomics).
Thank you Michael. I was hoping to use DESeq2 and compare the results (the hits) to those achieved with other methods (ALDEx2, ANCOM-II). A bit like they did in this highly cited paper (PMID: 3503952). But for DESeq2, I do not even understand what I would input instead of gene counts: total or proportions. corncob also looks exciting, I will look into it!