Hi, I’m running into an issue when running the nf-core/metatdenovo pipeline (v1.3.0) on a cluster. I’m annotating metatranscriptomic data with eggNOG, and the eggNOG database is already available on the cluster. I’ve specified the database path in my nf-params.json file using: "eggnog_dbpath": "/sw/data/eggNOG_data/5.0.0/rackham" However, the pipeline seems to ignore this parameter and instead tries to download the database from http://eggnogdb.embl.de/download/emapperdb-5.0.2/, which fails with a 404 error and interrupts the run. Has anyone encountered this before, or does anyone have suggestions for how to force the pipeline to use the local eggNOG database instead of attempting a download?

Hi! the param

--eggnog_dbpath

should avoid the download of a new db. I have few questions to check everything is fine: • do you have access to the eggnog database? • eggnog recognise specific files in the directory so make sure your path is pointing to the right one • where the pipeline is trying to download the database?

Hi! I forgot to mention that I'm working on PDC/Dardel. Yes, I should have access to the database, these are the files in the path I'm pointing to:

-rw-rw-r--+   1 mdahlo sinstall 41370988544 Mar  2  2021 eggnog.db
-rw-rw-r--+   1 mdahlo sinstall  9285439161 Mar  2  2021 eggnog_proteins.dmnd
-rw-r--r--+   1 mdahlo sinstall   278003712 Nov 11  2020 eggnog.taxa.db
-rw-r--r--+   1 mdahlo sinstall     6628719 Nov 11  2020 eggnog.taxa.db.traverse.pkl
-rw-rw-r--+   1 mdahlo sinstall       25895 Dec 23  2022 hist
drwxrwsr-x+ 381 mdahlo sinstall       40960 Jan 15  2023 hmmer
drwxrwsr-x+   2 mdahlo sinstall        4096 Dec 15  2022 mmseqs
-rw-rw-r--+   1 mdahlo sinstall  1291693577 Jul 19  2022 novel_fams.dmnd
drwxrwsr-x+   2 mdahlo sinstall        4096 Dec 15  2022 pfam
-rwxrwxr-x+   1 mdahlo sinstall        2051 Jan 19  2023 process_taxa-list.sh
-rw-rw-r--+   1 mdahlo sinstall       51772 Dec 15  2022 taxa-list.txt
-rw-rw-r--+   1 mdahlo sinstall       22579 Jan 15  2023 taxa-table.txt

I think the pipeline tries to download the database to my work directory but the problem is that the link doesn't work/is outdated. Still, the

--eggnog_dbpath

parameter is not recognised by the pipeline.

The files look fine. could you share more information? where the pipeline is storing the downloaded database from eggnog? by default, it should download them in a new directory called "eggnog"

I think this is because the pipeline also expects a

versions.yml

file. Add one looking like the below:

"NFCORE_METATDENOVO:METATDENOVO:EGGNOG:EGGNOG_DOWNLOAD":
    eggnog: 2.1.9

Thank you for your input! Yes, the problem was that the pipeline could not find a version.yml file in the eggnog db-folder on the cluster and the default link to the eggnog database is no longer working.