Dear all,
I don’t understand what is best: taxonomy summaries and heatmaps should be generated from the complete feature table or on the rarefied one? Or is enough to use relative frequencies instead of rarefying?
Thanks a lot in advance for any recommendation about this
From the complete one --- rarefaction is only necessary for some diversity metrics that are sensitive to sampling depth, otherwise, use the full feature table.
These are two different things --- relative frequencies are computed by determining the per-sample relative abundance for each feature in said sample, which can be done with or without the table having been rarefied.