Hmmscan

# 2) Use curated gathering thresholds from the database
hmmscan \
  --cut_ga \
  --domtblout domains.tbl \
  Pfam-A.hmm \
  proteins.fa

# 3) Produce smaller text output while keeping parseable tables
hmmscan \
  --noali \
  --tblout hits.tbl \
  --domtblout domains.tbl \
  Pfam-A.hmm \
  proteins.fa

Recommended Workflow

1. Press the HMM database once with hmmpress if you will reuse it.
2. Scan the sequence FASTA and save both sequence-level and domain-level tables.
3. Use curated thresholds when available; otherwise set E-value thresholds explicitly.
4. Review domain architecture, not just best-hit name, before functional labeling.

Guardrails

• Positional argument order matters: HMM database first, query sequence file second.
• Use -h for help; --help and --version are not valid here.
• --cut_ga, --cut_tc, and --cut_nc only make sense if the HMMs actually carry curated thresholds.
• Use --domtblout when domain boundaries matter; --tblout alone is not enough.