I’m working with the precomputed PICRUSt KEGG/GG tables and I am interested in doing some hierarchical clustering of the GG features based on their functional profiles.
To do that, I would need some kind of distance, but because this data is presence-absence, I am unable to use
qiime diversity beta without lying about what type this data is. It would be convenient if there was a
beta-qualitative method which accepted presence-absence tables and only exposed the qualitative metrics (much like
I imagine a similar thing could be done for alpha diversity as well.
I’m curious what other’s thoughts are on this? Is it perhaps too niche a use-case? At the moment I can only think of Jaccard and Hamming being relevant, but are there others?