I have a question about the sub-hypothesis testing of ANCOM.
Based on the documentation and Qiime2 implementation of ANCOM, now it seems there is only
scipy.stats.f_oneway option for the hypothesis test.
However, I’ve checked there were options for non-parametric test before here.
I’m wondering why non-parametric options were disregarded, and how normality assumption of alr transformed data can be made with zero-inflated sparse microbiome data.
Thanks in advance!