MetaEuk doesn't work with databases created using the commands metaeuk createdb or mmseqs createdb

[amizeranschi]

The current pipeline version (3.2.1) doesn't support running MetaEuk when supplying a database (via ) created with the commands metaeuk createdb or mmseqs createdb ran on a manually downloaded FASTA with protein sequences. The MetaEuk module is searching for a file called <prefix>.version, which isn't created automatically by the previous commands.

Relevant discussion on Slack: https://nfcore.slack.com/archives/CE9MS66BS/p1733150568097039?thread_ts=1733144901.017309&cid=CE9MS66BS (on #mag, 2nd Dec)

[jfy133]

Ugh this looks like it's related to mmseqs which I really don't like atm for similar reasons 😬 , I will investigate ASAP

[jfy133]

Reading through the slack thread, this seems to be more like a mmseqs2 issue more than mag, but I will test it otherwise to see if I can reproduce:

1. ✅ Running test_adapterremoval which includes auto-download of metaEuk databases (Giving a FASTA file of protein sequences to --metaeuk_db )
2. ✅ Downloading an mmseqs2 database with mmseqs2 databases Kalamari kalamari/kalamari tmp, and giving that to --metaeuk_db -> Downloading results in Cannot open kalamari/kalamari.source for writing, i hate mmseqs so much... will finish it tomorrow -> OK this worked for me no issues: nextflow run ../main.nf -profile test_adapterremoval,docker --outdir ./results-test_adapterremoval-downloaddb --metaeuk_db ~/cache/databases/mmseqs/kalamari/ -resume
3. Specifying the database name with --metaeuk_mmseqs_db -> nextflow run ../main.nf -profile test_adapterremoval,docker --outdir ./results-test_adapterremoval-downloaddb --metaeuk_mmseqs_db Kalamari -resume --metaeuk_db false

All of the tests above worked. Admittedly I've used the smallest database (Kalamari) offered MMSeqs and the Yeast genome, however all database variants worked fine.

So I suspect this isss something problematic with mmseqs download and/or UniRef100 and not something to do with the pipeline.

[jfy133]

Final test completed wiht ✅ - so unable to replicate.

If it happens again please reopen @amizeranschi