Prepare reference databases

For a minimal reference database:

metapathways build_db -d /data/MPDB --func swissprot -a fast

SILVA and supporting enzyme/taxonomy files are built alongside functional references. Select multiple functional databases with --func swissprot cazy uniref50; uniref90 is also supported and substantially larger. -a blast builds BLAST indexes instead of FAST indexes. Use --dryrun to inspect the planned downloads and indexing jobs first.

The builder downloads public references from their configured upstream locations. These are not all version-pinned. eggNOG requires a local FASTA at MPDB/functional/eggnog; the old builder had no working eggNOG acquisition rule. Licensed MetaCyc reference acquisition is separate from this public builder. Existing compatible MPDB installations can be used directly with -d.

Choose annotation databases actually present in your MPDB. The FAST/BLAST choice in run must match its indexes. Keep reference release records and checksums; changing a database can change biological results.