@gibsonnews @MehrbodEstaki agree with everything @gibsonnews says here, I tried to lay it out in this review https://t.co/AIU0msJHPz . Also may be interested in this empirical comparison of "count models" vs "compositional models": https://t.co/a46tbd3rqh
1,891 followers
941 followers
@MehrbodEstaki @jjminich That said, lib size is only meaningful (IMHO) if you follow equivolumetric protocols, for a given range of quantification. I don't have refs opposing the concept of compositionality, but there's evidence that non-CoDa methods perfo
493 followers
@aqua_jamie Exactly. Here are two recent benchmarking papers on differential abundance testing but again they don't show much concordance in the conclusions. https://t.co/IrW6VhC8r0
3,022 followers
RT @gavin_m_douglas: Interesting comparison of microbiome DA methods in new @GenomeBiology article. No tool clearly reigns supreme and ther…