steineggerlab
Add hit descriptions to output
Expected Behavior
Include a brief description of the database hits in the html and tabular output files from a search.
Current Behavior
The Foldseek web server includes some functional descriptions, but the command line version does not. It would be great if the html (and tabular) output from the command line program was more similar to the public web server, with hit descriptions and database hyperlinks.
Context
When using Foldseek for functional annotation, particularly in automated functional annotation pipelines, including information about the function of database hits would be extremely helpful. Now that tools like AlphaFold2 and ESMFold are fast enough to predict structures for entire genomes, it is feasible to use Foldseek in structure-based genome annotation pipelines, for example by searching the predicted structure of every protein in the genome against the AF2-SwissProt database and transferring annotations from the top database hits to the new sequence.
One solution would be to keep a separate database of annotations for all the sequences in the Foldseek database and look up all the hits after the search. I guess that may be what you do on the backend of the Foldseek webserver. It would be great if you made command line tools available for that, or integrated support for hit descriptions/functional annotations into Foldseek itself.
I think I found approximately the code that does this for the public webserver. I guess I can probably figure out how to adapt that to a command line program: https://github.com/soedinglab/MMseqs2-App/blob/da8747a513b06927541a755c45026f8ba2999cb5/frontend/Result.vue#L304
The convertalis flags are here if in the code:
https://github.com/soedinglab/MMseqs2-App/blob/cbb25067fe237aaf46c1fcb0acf552d14947681e/backend/worker.go#L169
The Foldseek html output does indeed lack behind. We'll have to update that.
