How to handle transferred and split EC numbers in humann outputs

Microbial community profiling HUMAnN
1

Hello,

I am analyzing HUMAnN gene family results that were renamed to EC numbers using humann_rename_table with the EC mapping. I noticed that some EC entries in my tables are EC numbers that have been officially transferred, split, or deleted by the EC nomenclature, yet they still appear with non-zero abundances in HUMAnN outputs.

This raises a conceptual problem for interpretation.

My questions are:

1. When an EC has been transferred to a single new EC, is it scientifically valid to sum the abundance of the old EC into the new EC for downstream analysis?

2. When an EC has been transferred into multiple new ECs, what is the correct quantitative treatment of the HUMAnN signal? Should it be split, duplicated, or kept as an ambiguous legacy EC?

Thank you in advance,

Liat

2

If an EC X has been renamed or regrouped into an EC Y, then it should be reasonable to transfer (add) X’s abundance to Y. If X was split into Y and Z, then it’s ambiguous what to do with X’s abundance, since it could arguably be all Y, all Z, or (probably more likely) some Y+Z mix.