I was searching for an explanation/tutorial of the ancombc2 command that would help me to analyze differentially abundant ASVs along a time series of experiments. It seems that ancombc2 has not (yet) been included in the major q2 tutorials (moving pictures, Atacama soil etc.).
I used –p-fixed-effects-formula groups to define the column containing the different groups, and used the –p-reference-levels groups::control to define the control group for pairwise testings. How can I instruct ancombc2 to test for differential abundance of ASVs within each timepoint for each timepoint?
That's a great question. Currently, the implementation of ancombc2 within q2-composition only supports cross-sectional studies. However, the ancombc2 R package does support longitudinal analyses - here's a section of their docs that goes through an example longitudinal dataset:
Thank you for your reply, I was aware about this tutorial outside qiime2.
Let me take the chance to include @hulfred in this reply, as I already found a note on the future tentatively ancombc3. I assume he/she is one of the developers. Would be great to have the longitudinal option also available within qiime2.
Indeed—ANCOM-BC3 is under active development! The major upcoming update will be the addition of differential prevalence analysis (presence/absence patterns) alongside differential abundance analysis.
