Hi, I’m trying to perform alpha diversity analysis on my samples and I’ve come across some papers criticizing the practice of rarefying. I’m wondering if there is a way to do the alpha diversity analysis in a different way that does not involve rarefaction?
Hi @ygao2,
Qiime2 currently doesn’t have any alternatives to rarefying but will likely sometime in the future. If you wanted to perform some other normalization method such as CSS or DESeq you would have to do this outside of qiime2. See this discussion for more details. A note worth mentioning that tools such as ANCOM and gneiss in qiime2 work with relative abundances and so don’t require rarefying. The uneven sampling depth issue is more of a concern with alpha and beta diversity analyses.
3 Likes
This topic was automatically closed 31 days after the last reply. New replies are no longer allowed.