PICRUST2 superpathways


I am currently working with PICRUST2 and we ran into an discussion about removing superpathways. We read some publications where these were removed but without any obvious rationality mentioned.

Talking to other researchers working with whole genome shotgun data also gave different results. One says why should you remove superpathways - the other says they need to be removed.

Is there a reason to remove those? I now that for databases with a limited documentation (e.g. biopsie samples) it can be tricky but we are currently working with stool samples where the databses are quiet good.

I would really appreciate your help, experciences and ideas!