Analysing Paired Data Taken at One Timepoint

User Support
3

Hi @macrobiome, I'm going to answer this question and the post where you refer to this problem here.

What might make sense is for you to use mantel here, as I think you were getting at here.

It sounds like you have samples from the same subjects at multiple body sites. If you want to know if the distances between subjects are correlated across body sites (i.e., if two individuals differ in their fecal microbiome, do they also differ in their intestinal microbiome), mantel will get you that information.

You can do this as follows.

  1. Use qiime feature-table split to generate on feature table per body site.
  2. Use qiime feature-table group to group samples in each of the "body site feature tables" (i.e., those resulting from step 1) by subject. In effect, this just renames the samples in each feature table to the subject identifiers.
  3. Run qiime diversity beta on each body site feature table to generate a distance matrix between the samples in that table.
  4. Run qiime diversity mantel on each pair of distance matrices resulting from step 3.
3 Likes