ValueError: Could not find HHBlits database path/uniclust30_2018_08

[ArsLeicholt]

Hey guys

Great work! I am trying to get OpenFold running on our cluster and having problems with hhblits on uniclust30. I adapted the longer paths to just "path" for privacy reasons. Help would be much appreciated. Tried out reinstalling the uniclust30 database already and triple-checked all the paths.

ERROR Message

INFO:/path/run_pretrained_openfold.py:Generating alignments for goddard... ERROR:root:Could not find HHBlits database path/uniclust30_2018_08 Traceback (most recent call last): File "path/run_pretrained_openfold.py", line 451, in <module> main(args) File "path/run_pretrained_openfold.py", line 284, in main precompute_alignments(tags, seqs, alignment_dir, args) File "path/run_pretrained_openfold.py", line 66, in precompute_alignments alignment_runner = data_pipeline.AlignmentRunner( File "path/data_pipeline.py", line 377, in __init__ self.hhblits_bfd_uniclust_runner = hhblits.HHBlits( File "pathhhblits.py", line 86, in __init__ raise ValueError( ValueError: Could not find HHBlits database path/uniclust30_2018_08

The uniclust30_2018_08 directory contains the 13 files from the hhblits_hmm

uniclust30_2018_08.cs219 uniclust30_2018_08_a3m.ffdata uniclust30_2018_08_a3m_db uniclust30_2018_08_cs219.ffdata uniclust30_2018_08_hhm.ffdata uniclust30_2018_08_hhm_db uniclust30_2018_08_md5sum uniclust30_2018_08.cs219.sizes uniclust30_2018_08_a3m.ffindex uniclust30_2018_08_a3m_db.index uniclust30_2018_08_cs219.ffindex uniclust30_2018_08_hhm.ffindex uniclust30_2018_08_hhm_db.index

Complete slurm script

`#!/bin/bash

#SBATCH --nodes=1 #SBATCH --ntasks 4 #SBATCH --cpus-per-task 8 #SBATCH --mem=80G #SBATCH --partition=gpu2080 #SBATCH --gres=gpu:1 #SBATCH --time=24:00:00 #SBATCH --job-name=openfold_test #SBATCH --mail-type=ALL

module --force purge

ml palma/2021a ml gompi/2021a ml OpenFold/1.0.0-CUDA-11.3.1

wait

run_pretrained_openfold.py
/path/fasta/
/path/pdb_mmcif/mmcif_files
--uniref90_database_path /path/uniref90.fasta
--mgnify_database_path /path/small_bfd/bfd-first_non_consensus_sequences.fasta
--pdb70_database_path /path/data/alphafold/pdb70
--uniclust30_database_path /path/uniclust30_2018_08
--output_dir path/results/
--bfd_database_path=/path/bfd/bfd_metaclust_clu_complete_id30_c90_final_seq.sorted_opt
--model_device "cuda:0"
--jackhmmer_binary_path /path/jackhmmer.py
--hhblits_binary_path /path/hhblits.py
--hhsearch_binary_path /path/hhsearch.py
--kalign_binary_path /path/kalign.py
--config_preset "model_1_ptm"
--openfold_checkpoint_path /path/finetuning_ptm_2.pt

#ENV`

[tsbartle]

I am receiving the same error in a similar environment.

[vetmax7]

Hello!

Could you find solutions?

[1999caozhigang]

please!!!

[vetmax7]

I could only run my script if I removed ''--uniref30_database_path". Also I write full path for "--bfd_database_path": "bfd/bfd_metaclust_clu_complete_id30_c90_final_seq.sorted_opt".

[1999caozhigang]

我只有删除“--uniref30_database_path”才能运行我的脚本。此外，我还为“--bfd_database_path”写入了完整路径：“bfd/bfd_metaclust_clu_complete_id30_c90_final_seq.sorted_opt”。

Can you tell me what this problem is? Is it a problem with the database address pointing to?