Seeding the Rarefaction Algorithm

There was a good discussion about reproducibility at the Phoenix Qiime2 workshop, and I had a thought as we were working through the tutorial. Would it be possible to set a seed for the Rarefaction when doing diversity analysis? This would allow for the direct reproduction of an analysis.

Is this even possible within the current Qiime2 implementation?

Hi @Droush,
Thanks a lot for your interest in QIIME 2 and for your suggestion. I created an issue here to discuss this. It’s not something that’s currently possible in QIIME 2 (or QIIME 1, for that matter), though technically it would be possible. Are there specific scenarios where you would want to exactly replicate rarefaction? I wonder how useful this would be, relative to just re-using the rarefied table that had already been computed.

Thanks again for the post!