Hi mothur forum,
may I have a reply on how to improve on unclassified
in the taxonomy file, if possible? For example, can
output from megablast become a taxonomy file?
Sigmund
blast is a pretty horrible way to do classification.
To improve the number of unclassified you can try…
- Change your database - use greengenes, silva, and rdp and see which works best
- Trim the database to the region you are interested in
- Get longer reads
- Pick a different region of the gene that will result in better classification of the taxa you are most interested in
- Get over the need to have names that are as specific as you want.
We tend to focus on #5 with a dash of #1 and #2. 