Abundance table output annotates surprisingly few taxa even after changing parameters


I am trying to run MetaPhlAn on a set of data involving mice, however, the obtained output is yielding a surprisingly low number of annotations. The filtered FASTQ file being used as input has approximately 10 million sequences. I have checked some of the reads that are not annotated via BLAST, and they appear to be aligning against commensal taxa within the mouse microbiome such as the Muribaculaceae.

In trying to increase the number of annotations observed, I have changed the following parameters: --min_alignment_len, --pres_th, --min_mapq_val, and --bt2_ps BowTie2 presets. The resulting merged_abundance_table.txt file, however, stays the same, showing no changes in either direction even after making dramatic changes in the parameters. Is there is a reason why the table is staying the same, and is there a way to resolve this problem?

Thank you for any help you can provide!


Dear colleagues,

I am working on shotgun metagenomics with hamster samples and facing a challenge similar to what Rachael described. I have a substantial number of unclassified reads and a significant proportion classified only at the kingdom level (Bacteria).

I am exploring two scenarios: parameter optimization and the impact of the Metaphlan version used.

Any hints or insights to help resolve this issue would be greatly appreciated.

Thank you!