Consider using taxa instead of feature

[TuomasBorman]

I think we use both "feature" and "taxa" when we are talking about rows. "Feature" is more generic, but we could use more "taxa" as most of the readers are working with them as @nuorenarra pointed out.

[antagomir]

Consistency is good. I would say that consistency is even more important.

It could be taxa but on the other hand, we may also deal with functional predictions (e.g. from HUMAnN) or other omics (e.g. metabolomics) using the same methods, and this is now done on a regular basis. Therefore referring to "taxa" can be confusing as well. It might depend on the context.

Perhaps "taxonomic features"? And if there are other types of examples, then functional or metabolomic features.

It might be good, anyway, to explain the relation between these in the introductory chapter where the data structure is introduced.