ANCOM-BC global test and multiple comparisons about ancombc

[FrederickHuangLin]

Hi @tenguzame,

Thanks for your inquiry!

As for your questions, the ANCOM-BC global test aims to identify taxa that are (significantly) differentially abundant in at least one group (if there are multiple groups). Therefore, no reference group is needed.

You do not have to normalize the count table before doing any differential abundance analysis using ANCOM-BC, since the normalization (or we called it bias-correction) step is embedded in the ANCOM-BC procedure).

Visualization using CLR-normalized data is a great idea! Since CLR transformation is compositionally robust, there is no need to normalize data either.

Best,
Huang