MetaPhlAn and HUMAnN for murine oral and gingival samples: High unclassified proportions

Microbial community profiling MetaPhlAn
1

Can MetaPhlAn and HUMAnN reliably profile murine oral and gingival microbiomes?

We are analyzing samples from mouse colon, oral, and gingival sites. Initial MetaPhlAn run, approximately 65% of gingival samples were reported as 100% unclassified, compared with approximately 10% of samples from each of the other two sites.

We are unsure whether this reflects biological differences or a technical issue. DNA amount was relatively similar in the FASTQ files from all sites.

Beyond removing host reads using a Mus musculus reference:

  1. Are there recommended preprocessing steps, settings, or databases for profiling oral and gingival samples with MetaPhlAn and HUMAnN?

  2. Could limited database representation of murine oral microbes contribute to this pattern?

  3. Do you recommend an QC checks to distinguish low microbial signal vs. limited database coverage or a technical pipeline issue?

    Thank you for making these tools available!