asahaman
commented
3 years ago I used the option refseq_masher matches. I think with the option of refseq_masher contains you get many more hits (like contamination). It may be interesting to compare outputs of matches vs contains to see if we can flag contamination. I have read about Kraken too but haven't explored it.
emarinier
When performing lineage assignment, flag and inform the user of possible contamination. The assemble command currently reports a probability of different species assignment, but doesn't consider contamination.
This would ideally be at the contig level, but creating rules for handling contamination at the assembly level might be a good starting point.