DRAM database installation problem; The subcommand mmseqs createindex DRAM_data/uniref90.20240308.mmsdb DRAM_data/tmp --threads 10 experienced an error

[camillaln]

Hi,

I am having problems installing the DRAM database. I get the same error on two different servers. I installed DRAM following the instructions:

git clone https://github.com/WrightonLabCSU/DRAM.git
cd DRAM
conda create --name dram_env -f environment.yaml
conda activate dram_env

Install pip

conda install pip3
pip3 install ./

Here is the error:

2024-03-08 18:16:35,301 - Processing uniref

2024-03-08 18:52:57,067 - The subcommand ['mmseqs', 'createindex', 'DRAM_data/uniref90.20240308.mmsdb', 'DRAM_data/tmp', '--threads', '10'] experienced an error: Killed

Traceback (most recent call last):
File "/gpfs/fs3c/aafc/aafc_shp/can0002/miniconda3/envs/dram_env/bin/DRAM-setup.py", line 187, in
args.func(**args_dict)
File "/gpfs/fs3c/aafc/aafc_shp/can0002/miniconda3/envs/dram_env/lib/python3.10/site-packages/mag_annotator/database_processing.py", line 615, in prepare_databases
processed_locs = process_functions[i](locs[i], output_dir, LOGGER,
File "/gpfs/fs3c/aafc/aafc_shp/can0002/miniconda3/envs/dram_env/lib/python3.10/site-packages/mag_annotator/database_processing.py", line 322, in process_uniref
make_mmseqs_db(uniref_fasta_zipped, uniref_mmseqs_db, logger, create_index=True, threads=threads, verbose=verbose)
File "/gpfs/fs3c/aafc/aafc_shp/can0002/miniconda3/envs/dram_env/lib/python3.10/site-packages/mag_annotator/utils.py", line 98, in make_mmseqs_db
run_process(['mmseqs', 'createindex', output_loc, tmp_dir, '--threads', str(threads)], logger, verbose=verbose)
File "/gpfs/fs3c/aafc/aafc_shp/can0002/miniconda3/envs/dram_env/lib/python3.10/site-packages/mag_annotator/utils.py", line 71, in run_process
raise subprocess.SubprocessError(f"The subcommand {' '.join(command)} experienced an error, see the log for more info.")
subprocess.SubprocessError: The subcommand mmseqs createindex DRAM_data/uniref90.20240308.mmsdb DRAM_data/tmp --threads 10 experienced an error, see the log for more info.

[crane90]

i have the same problem when setting up the databases on our Linux cluster.
I used this code:

<DRAM-setup.py prepare_databases --output_dir DRAM_data>

and get this error in the log file:

2024-04-02 14:26:39,079 - Downloading function_heatmap_form
2024-04-02 14:26:39,294 - Downloading amg_database
2024-04-02 14:26:39,524 - Downloading etc_module_database
2024-04-02 14:26:39,749 - All raw data files were downloaded successfully
2024-04-02 14:26:39,749 - Processing kofam_hmm
2024-04-02 14:30:15,372 - KOfam database processed
2024-04-02 14:30:16,640 - Moved kofam_hmm to final destination, configuration updated
2024-04-02 14:30:16,640 - Processing kofam_ko_list
2024-04-02 14:30:16,723 - KOfam ko list processed
2024-04-02 14:30:16,726 - Moved kofam_ko_list to final destination, configuration updated
2024-04-02 14:30:16,726 - Processing uniref
2024-04-02 15:19:14,992 - The subcommand ['mmseqs', 'createindex', 'db/DRAM_new/uniref90.20240402.mmsdb', 'db/DRAM_new/tmp', '--threads', '10'] experienced an error: db/DRAM_new/tmp/11412079875784970340/createindex.sh: line 56: 412650 Killed "$MMSEQS" $INDEXER "$INPUT" "$INPUT" ${INDEX_PAR}

[crane90]

in my case it seems to have run out of working memory, now that i have allocated more working memory for the command it worked fine

[BioRRW]

Thank you for choosing to use DRAM. Unfortunately, at this time, we do not have the capacity to provide support for DRAM. We appreciate your understanding and encourage you to stay tuned for the release of DRAM2, expected in 2025.

Best regards, DRAM dev team

[WhatChan123]

@crane90 Hi, Can you please show me how to allocate more working memory for the command? I also have the same problem! Thank you!