I have simulated a metagenome sample from known species. When running humann3 for this metagenome sample, during metaphlan stage some reads are matching to a species that does not exist as an input. I bowtie2’ed the simulated reads against the species that matahplan output gives and there is 60-70% read match. I wonder if I can increase the percent identity threshold for metaphlan option when running humann3 so that the metagenomic reads will not match with other species.
Thank you for your help,