sorry for bringing this up again, but this issue left me a little confused: while working on the worklow I started wondering how to normalize between samples. Since we will have different sequencing depths and different community compositions between samples, I think this is an important factor to consider. On this behalf, I found the discussions here: Metaphlan normalization and Humann normalization.
Can I use DEseq2 like in differential RNA epxression analysis? Or may this conflict with the way Metaphlan calculates abundances within a given sample? I also thought of using a spike-in control, but here we have the problem of contamination and we loose covarage.
Using total sum scaling like in Humann could be an option? Could this be performed with the option -t marker_ab_table see here?
Thanks a lot,