Abundance overlap microbiome

General Discussion
1

Hi everyone,

I am working with a dataset consisting of sample triplets (Sets of 3: Sample A, Sample B, and Sample C).

  • Samples A and B are "neighbors" (potential sources).
  • Sample C is a "sink" that can conceptually be composed of A, B, and many other unknown sources.

To test this, I ran FEAST (Fast Expectation-Maximization for Microbial Source Tracking). The results confirmed my hypothesis: Samples A and B are indeed contributors to the microbiome of Sample C.

Now that I know the overall community-level contributions, I want to zoom in on the specific bacteria driving this relationship. Specifically, I want to:

  1. Identify the exact bacterial taxa (ASVs) that overlap between A, B, and C.
  2. Determine the "importance" or differential abundance of these overlapping taxa to understand which ones are truly characteristic of this source-sink transition.

What are the best workflows or packages to extract and visualize these specific overlapping taxa?
How should I calculate and rank their "importance"?

I am primarily working in QIIME2/R studio, but with limited expierience. Any advice, tool recommendations, or papers with similar workflows would be greatly appreciated!

Thanks in advance!

2

Hi @ppm,

I dont have any direct answers for you because this is a non-trivial question. You might look into the strain sharing and maternal transmission spaces. You might consider looking at paired sample techniques and ask questions like whether the sample is more similar (and how) to the sources or sinks. If you can define an appropriate gradient, gemelli might be a tool to consider. It, of course, depends on your sample size. I love an rPCA or modified rPCA for feature ranking. There are also rank-based techniques around.

Best,
Justine

2 Likes
3

Hi Justine,

Thank you for your help, much appreciated. I will definitely try that.

Best,
Peter