emapper.py has multiple problems

[drdna]

1. The instructions state that basic usage is as follows: emapper.py -i FASTA_FILE_PROTEINS -o test This is not correct if you download the databases to a directory other than the eggnog-mapper directory.

2. The instructions imply that eggnog-mapper will query the EGGNOG_DATA_DIR environment variable. It does not. (base) [user@mcc-login001 user]$ echo $EGGNOG_DATA_DIR /scratch/user/tmp

(base) [user@mcc-login001 farman]$ ls $EGGNOG_DATA_DIR/ e5.proteomes.faa e5.taxid_info.tsv eggnog.db eggnog_proteins.dmnd eggnog.taxa.db eggnog.taxa.db.traverse.pkl fungi.dmnd

(base) [user@mcc-login001 user]$ singularity run --app eggnogmapper2112 /share/singularity/images/ccs/conda/amd-conda22-rocky9.sinf emapper.py -i SSproteins.fasta -o MyProteinAnnotations

DIAMOND database /usr/local/Miniconda3/envs/eggnog-mapper-2.1.12/lib/python3.11/site-packages/data/eggnog_proteins.dmnd not present. Use download_eggnog_database.py to fetch it

3. If you point explicitly to the to the diamond database that is to be used, it will throw another error: (base) [user@mcc-login001 farman]$ singularity run --app eggnogmapper2112 /share/singularity/images/ccs/conda/amd-conda22-rocky9.sinf emapper.py -i SSproteins.fasta --dmnd_db tmp/fungi.dmnd -o MyProteinAnnotation

Annotation database data/eggnog.db not present. Use download_eggnog_database.py to fetch it

4. There doesn't appear to be any option flag that allows one to specify the path to the annotation file

5. The process does not die on failure to detect a database. SLURM jobs keep on running until timeout.

6. Jobs run on the login node run forever, open empty files and empty directories and never write to them. No error messages are produced.

7. Jobs run in slurm fail to start any processes, keep on running forever, and throw no error messages.

[RilRis]

My custom DIAMOND database is also in a nonstandard , and I fixed the "DIAMOND database [name].dmnd not present. Use download_eggnog_database.py to fetch it" issue by changing a single line (line 49) in the eggnogmapper/search/search_modes.py file.

Under def get_eggnog_dmnd_db, change this:

if dmnd_db is not None:
        ret_dmnd_db = dmnd_db

to this:

 if dmnd_db is not None:
        ret_dmnd_db = pjoin(data_path, dmnd_db)

From there, running emapper.py -i [protein fasta] --output_dir [output directory] -m diamond --dbname [path to eggnog db dir] --dmnd_db fungi.dmnd appears to work