You should be able to just use SILVA 138 to classify, then if you do want to use SEPP, you can get the SILVA 128 SEPP reference database from the bottom of the data-resources page.
I also reached out to another moderator who made some resource suggestions and additional steps if you want to build your own SEPP database:
- Trouble generating SeppReferenceDatabase qza
- Build a SILVA tree using FastTree
- Then simply place the various files into a directory like so, and import using the
SeppReferenceDatabasetype, with the following files:aligned-dna-sequences.fastatree.nwkraxml-info.txt
Neither of us have ever actually needed to run this workflow, so let us know if you run into any difficulties, but that should get you headed in the right direction!