Custom genetic database – Deepmind/Alphafold

It is possible, but only with a code change in data/pipeline.py:

• If the database is a FASTA file, you could add a new Jackhmmer searcher for that database. You can take a look at the jackhmmer_uniref90_runner and basically follow the same logic for your database.
• If the database is a HH Suite database, you could add a new HHBlits searcher for that database. You can take a look at the hhblits_bfd_uniclust_runner for inspiration in this case.

Alternatively, if this is a custom FASTA file, you could try passing that instead of one of the existing databases, e.g. small_BFD or MGnify. Note this might lead to worse results as you would be replacing one database with another.