biobakery_workflows 3.1 with MetaPhlAn 3 vs 4 — what are the differences?

Hi,

I’m setting up bioBakery workflows on an HPC cluster. When installing biobakery_workflows via conda, the solver installs MetaPhlAn 3.0.13, while on my local machine with a newer conda version it installs MetaPhlAn 4.0.6, both with biobakery_workflows 3.1. The HPC cluster has conda 4.10.1 and I am unable to update it as it is a shared system managed by the administrators.

Since MetaPhlAn 3 and 4 use different databases, I’m concerned about how this affects results and workflow compatibility.

My questions:

  1. What are the key differences in taxonomic profiling results between MetaPhlAn 3 and 4?

  2. For mouse gut microbiome shotgun metagenomics with downstream HUMAnN3 functional profiling, would you recommend one version over the other?

  3. Is it recommended to separately install MetaPhlAn 4 from bioconda after installing biobakery_workflows via the biobakery channel, given that the biobakery channel only provides MetaPhlAn 3?

Thanks!