set Diamond load threads parameter

Author: oschwengers

bakta/expert/protein_sequences.py

@@ -28,6 +28,7 @@ def search(cdss: Sequence[dict], cds_fasta_path: Path, expert_system: str, db_pa
         '--max-target-seqs', '1',  # single best output
         '--outfmt', '6', 'qseqid', 'sseqid', 'qlen', 'slen', 'length', 'pident', 'evalue', 'bitscore', 'stitle',
         '--threads', str(cfg.threads),
+        '--load-threads', str(min(cfg.threads, 8)),  # limit load threads to max 8
         '--tmpdir', str(cfg.tmp_path),  # use tmp folder
         '--block-size', '4',  # increase block size for faster executions
         '--index-chunks', '1',  # set index chunks to 1 for faster executions

bakta/features/cds.py

@@ -600,6 +600,7 @@ def predict_pseudo_candidates(hypotheticals: Sequence[dict]) -> Sequence[dict]:
         '--max-target-seqs', '1',  # single best output
         '--outfmt', '6', 'qseqid', 'sseqid', 'qlen', 'slen', 'length', 'pident', 'evalue', 'bitscore', 'qstart', 'qend', 'sstart', 'send', 'full_sseq',
         '--threads', str(cfg.threads),
+        '--load-threads', str(min(cfg.threads, 8)),  # limit load threads to max 8
         '--tmpdir', str(cfg.tmp_path),
         '--block-size', '3',  # slightly increase block size for faster executions
         '--fast'

bakta/features/s_orf.py

@@ -331,6 +331,7 @@ def search(sorfs: Sequence[dict], cluster_type: str):
         '--evalue', '1',
         '--outfmt', '6', 'qseqid', 'sseqid', 'qlen', 'slen', 'length', 'pident', 'evalue', 'bitscore',
         '--threads', str(cfg.threads),
+        '--load-threads', str(min(cfg.threads, 8)),  # limit load threads to max 8
         '--tmpdir', str(cfg.tmp_path),  # use tmp folder
         '--block-size', '3',  # slightly increase block size for faster executions
         '--ultra-sensitive',

bakta/psc.py

@@ -45,6 +45,7 @@ def search(cdss: Sequence[dict]) -> Tuple[Sequence[dict], Sequence[dict], Sequen
         '--max-target-seqs', '1',  # single best output
         '--outfmt', '6', 'qseqid', 'sseqid', 'qlen', 'slen', 'length', 'pident', 'evalue', 'bitscore',
         '--threads', str(cfg.threads),
+        '--load-threads', str(min(cfg.threads, 8)),  # limit load threads to max 8
         '--tmpdir', str(cfg.tmp_path),  # use tmp folder
         '--block-size', '3',  # slightly increase block size for faster executions
         '--fast'

bakta/pscc.py

@@ -38,6 +38,7 @@ def search(cdss: Sequence[dict]) -> Tuple[Sequence[dict], Sequence[dict], Sequen
         '--max-target-seqs', '1',  # single best output
         '--outfmt', '6', 'qseqid', 'sseqid', 'qlen', 'slen', 'length', 'pident', 'evalue', 'bitscore',
         '--threads', str(cfg.threads),
+        '--load-threads', str(min(cfg.threads, 8)),  # limit load threads to max 8
         '--tmpdir', str(cfg.tmp_path),  # use tmp folder
         '--block-size', '3',  # slightly increase block size for faster executions
         '--fast'

db-scripts/diamond.nf

@@ -46,6 +46,7 @@ process diamond {
         --max-target-seqs 1 \
         -b4 \
         --threads ${task.cpus} \
+        --load-threads ${task.cpus} \
         --out diamond.tsv \
         --outfmt 6 qseqid sseqid stitle length pident qlen slen evalue \
         --fast