Differences in HUMAnN VS MetaPhlAn only relative abundances


I have previously run MetaPhlAn (v3) on a metagenomic sample from stool using default parameters to obtain relative abundances (relative abundance profile samp1_S10_L001_metaphlan_rel_ab_from_metaphlanOnly.tsv attached). I am now running HUMAnN on the same sample and compared the MetaPhlAn relative abundances that is outputted during running of HUMAnN (this profile also attached samp1_S10_L001_metaphlan_rel_ab_from_humann.tsv). I notice the same clades are detected from each run, but the relative abundances of some of these clades are slightly different. I also notice some differences in the bowtie2 mappings as well.

On the surface it looks like the same MetaPhlAn command is being performed in both scenarios, so why the difference in the relative abundances and bowtie2 mappings?

Thank you for the help,
Zach W
samp1_S10_L001_metaphlan_rel_ab_from_humann.tsv (37.1 KB)
samp1_S10_L001_metaphlan_rel_ab_from_metaphlanOnly.tsv (37.1 KB)