Do any existing qiime2 tools allow classification against the NCBI 16S DB? I’m still debating about turning my experimental code into a qiime2 plugin so it is helpful to understand the issues with exsting tools as seen by current users. For custom assignments will most of these tools accept an arbitrary collection of known sequences? In this case could a qiime2 user run the unknowns against chloroplast or 18S or is BLAST’ing about the only option?
Thanks. Also if you ware just interested in speculative opinions on unassignable reads, I would be very curious to see what I come up with for fish gut. Are these fish anywhere near hydrothermal vents? lol. I’ve seen extremeophiles come up some times in unexpected places. Thanks.
All taxonomic classification methods used in current QIIME 2 plugins (that I am aware of) can take any reference database, so long as it is formatted correctly. So the NCBI 16S DB could be downloaded, formatted, and imported for use with existing methods.
What does not exist — and could be pretty cool — is a method to remotely query NCBI BLAST and report those results in a single QIIME artifact. We have BLAST-based methods, but none that directly query NCBI without downloading a database.
Yes
Yes. It would be possible to filter out sequences that fail to classify and reclassify against a second database, or use q2-quality-filter to split a sequence file into "hits" and "misses" above/below a specified similarity threshold to custom reference sequences (e.g., to remove chloroplast).
I hope that helps!
1 Like
This topic was automatically closed 31 days after the last reply. New replies are no longer allowed.