SUBSEQ.BIO
DOCS-ESMFOLD

ESMFold

Predict protein structures quickly from amino-acid sequence without MSAs or templates.

Overview

  • Rapid foldability checks for variants, designs, orthologs, and metagenomic sequences.
  • First-pass structural annotation when homolog context is limited.
  • Batch single-chain protein structure generation for visualization or filtering.

Modes

ModeInput shapeWhen to use it
structure_prediction
Structure Prediction Default		 Consumes a folder or output set; useful for batches and pipeline handoffs. Predict structures for all FASTA inputs in the selected source.

Canonical Job Configuration

These are the fields exposed by the default job configuration for esmfold. They are also returned by GET /api/v1/program/params?program=esmfold and submitted as the params JSON object to POST /api/v1/job/submit.

No additional fields in this group.

Advanced configuration fields
ParameterTypeModesWhat it does
prediction_depth
Prediction Depth		 Text All modes Fast uses fewer recycling passes; Thorough spends more runtime on recycling.
Default: Standard; Options: Fast, Standard, Thorough
split_fasta_chainbreaks
Split FASTA Chain Breaks		 Yes/no All modes When FASTA records contain ':' chain breaks, submit them as separate FASTA records. Disable to preserve upstream colon handling.
Default: true

Outputs And Metrics

  • One PDB structure per predicted sequence.
  • Per-residue pLDDT stored in the PDB B-factor field.
  • Mean pLDDT summarizes local confidence, but inspect low-confidence regions separately.

Common Examples

  • Single protein FASTA for a quick structure check.
  • Batch independent designs in one FASTA file.
  • Colon-separated FASTA segments split into separate predictions when chain-break splitting is enabled.

Example API params

{
  "mode": "structure_prediction",
  "prediction_depth": "Standard"
}

Caveats

  • ESMFold may underperform MSA/template-based predictors when evolutionary or structural context is important.
  • This workflow does not explicitly model ligands, nucleic acids, cofactors, or modifications.
  • High pLDDT supports local structure confidence, not biological state or binding.

Advanced Submit

Advanced submit is still available for direct program arguments through POST /api/v1/job/submit-advanced. Prefer canonical configuration unless you need exact low-level arguments or are reproducing a known command line.

  • Advanced submit accepts direct ESMFold folding arguments for recycle count and FASTA handling.
  • Use complex-capable predictors for interfaces or ligand/nucleic-acid context.
curl -X POST https://subseq.bio/api/v1/job/submit \
  -H "Authorization: Bearer <api_key>" \
  -F program=esmfold \
  -F 'params={"mode":"structure_prediction","prediction_depth":"Standard"}'

Further Reading