Tools for functional differential analysis

Library Support
,
1

Hi everyone!

I have 16S amplicons from plant rhizosphere and ran those through QIIME2. Then ran PICRUST2 on the feature table to predict community functions.
I already ran ANCOM-BC on the feature table I got from QIIME2, but to explore the data further I want to run a differential abundance analysis on the PICRUSt2 output. Would it be adequate to use ANCOM-BC for this?
As far as I understand, ANCOM-BC relies on certain assumptions, and I don't know if the PICRUSt2 output data follows those.

Also while searching info about this I found the ggpicrust2 R package that explicitly mentions it excludes ANCOM and ANCOM-BC, but doesn't explain why.

If ANCOM-BC is not suitable for this could someone recommend me an alternative?

2

Hi! Welcome to the forum!

Have you read the ANCOM-BC paper that the tool is based off of? It will have those assumptions as well as example data that might even look enough like yours. :slight_smile: The Moving Pictures Tutorial also has a section on ANCOM-BC.

There's also ANCOM-BC2, which has its own paper and it used in the Gut-to-Soil Axis Tutorial. You can see the options for the Qiime2 Composition tool in the library. I can't suggest other tools/methods because I'm also new to qiime, this kind of analysis, and even have my own questions on the applicability of ANCOM. :sweat_smile: