Assigning biomarkers to variables with groups using ANCOM


I am using ANCOM to identify deferentially abundant taxa between 4 different groups: A, B, C and D. I have an output showing taxa that have been identified but how do I know whether they are assigned to A, B, C or D in my groups?

My output consists of:

  • a volcano plot of clr against W.
  • a table of ANCOM statistical results with the taxa, W and the TRUE or FALSE column for the hypothesis.
  • a percentile abundances of feature tables table.

I have looked at all of these but can’t identify which biomarkers are associated with which groups.

I am assuming I am missing something here but after searching through the forum and trying to interpret the data i am no closer to understanding it.

Is it possible to assign biomarkers to A, B, C and D here?


Hi @LVC,

ANCOM will not “assign” these to any group. When you have more than 2 groups, it can be difficult to interpret which group is associated with any differentially abundant features, and you have two options:

  1. look at the abundances of those features in each group to “eyeball” which group(s) most likely have different abundances of that feature
  2. if you really must, run a post-hoc test (which in this case would be running ANCOM on each pair of groups)

The percentile abundance table is sort of inconvenient but should give you an idea which group(s) that feature is more abundant in. Another (more useful?) option: filter your feature table to contain only the significant features and plot as a heatmap with qiime feature-table heatmap.

Good luck!

Thanks @Nicholas_Bokulich. That makes sense, just wanted to make sure I hadn’t missed something.

1 Like

This topic was automatically closed 31 days after the last reply. New replies are no longer allowed.