The bioBakery help forum

MaAsLin2: possible with 2x4 samples?

Hello everyone,
I have a general statistics question and would be grateful if anyone had an answer! We have a shotgun metagenomics dataset of 2x4 samples and conducted MaAsLin2 analysis. However, we are really insecure whether this is valid for so few replicates. Can we use the results?
Thank you very much in advance and best regards,
Kat

Hi Kat -

Apologies for the late response. While the sample size is low and power is likely off for the statistical analysis running these data through MaAsLin while recognizing those factors is not unrealistic. That being said I would make sure that the product of the models is biologically plausible - which is good practice for any differential abundance test.

Best,
Kelsey