Statistics - relative abundance


I have 16 subjects (5 samples for each subject over a 21 day period), and I have calculated the mean relative abundance for each time point at phylum level. 8 of these subjects have a disease, 8 and of these are healthy controls.

Basically, I want to test the statistical significance of any shifts in relative abundance, based on whether a subject has the disease or not.

Any advice on statistical methods would be much appreciated.



Metastats sounds like your best option.

Thank you for your suggestion.

I have considered using metastats, but this is complicated by the fact that my file contains data from each animal at multiple time points, and I can only input “Diseased or Healthy” into my design file and this would not take time point into account.

I may be missing something really obvious, though…



You’d probably want to use an ANOVA/linear models approach in R with the data from the shared file.

Great, thank you!