CAT prepare doesn't consider existing diamond database

[nikolasbasler]

Hi,

I have just installed CAT 5.3 and I'm trying to prepare a databse. I already have a diamond database nr.dmnd but CAT doesn't seem to find it (although it should, according to #66). Instead it will make a new one and place it into a db subfolder. Since the database is 300GB large, I don't want to have it twice.

This is the command I'm using:
CAT prepare --db_fasta nr.faa --names names.dmp --nodes nodes.dmp --acc2tax prot.accession2taxid.FULL.gz --db_dir .

A few questions:

1. Is it possible to point CAT prepare to the existing database so it doesn't create a new one?
2. Can I delete CAT's diamond database and move the other generated files into the folder of my diamond database and then refer to that as the database folder when running e.g. CAT contigs -d <folder>?
3. In that case, do I have to observe any naming conventions or will CAT find the files in the database folder by their endings (.dmnd, .fastaid2LCAtaxid and .taxids_with_multiple_offspring)?

Thanks and kind regards,
Nikolas

[harrytchild]

Hi,

I second this request. Could you also comment on the use of pre-existing GTDB-tk databases to avoid using too much memory for databases?

Kind regards,

Harry