AlphaFold 2

Overview

Protein monomer structure prediction from amino-acid FASTA.
Protein-only homo- or heteromeric complex modeling when stoichiometry is known.
Comparing homologs, mutants, truncations, or construct boundaries.
Inspecting pLDDT, PAE, pTM, and ipTM before downstream design or simulation.

Modes

Mode	Input shape	When to use it
`monomer` Monomer Default	Consumes a folder or output set; useful for batches and pipeline handoffs.	Predict each FASTA input as a single-chain target.
`multimer` Multimer	Consumes a folder or output set; useful for batches and pipeline handoffs.	Predict each multi-sequence FASTA input as one complex.

Canonical Job Configuration

These are the fields exposed by the default job configuration for alphafold2. They are also returned by GET /api/v1/program/params?program=alphafold2 and submitted as the params JSON object to POST /api/v1/job/submit.

Parameter	Type	Modes	What it does
`template_cutoff` Template Cutoff Date	Text	All modes	Only structural templates released on or before this date are considered; sequence databases are not date-limited. Leave blank to use the UTC submission date. Format: YYYY-MM-DD
`msa_search` MSA Search	Text	All modes	Fast uses the reduced database set; Full searches the full database set and is slower. Default: Fast; Options: Fast, Full
`pairwise_confidence` Pairwise Confidence	Yes/no	Monomer	Use the monomer pTM model to produce PAE/pTM confidence outputs. Default: false
`complex_sampling` Complex Sampling	Text	Multimer	Standard runs the usual five seeds per multimer model; Fast runs one seed per model. Default: Standard; Options: Standard, Fast

Advanced configuration fields

Parameter	Type	Modes	What it does
`relaxation` Relaxation	Text	All modes	Relax the top-ranked model, skip relaxation, or relax every model. Default: Best; Options: Best, None, All

Outputs And Metrics

Ranked PDB structures, with the top model usually named like ranked_0.pdb.
Relaxed and unrelaxed structures when relaxation is enabled.
Confidence files and model result files containing pLDDT, PAE, pTM, and multimer interface scores.
pLDDT is per-residue local confidence on a 0-100 scale; above 90 is very high, 70-90 is usually confident, 50-70 is low, and below 50 is often unreliable or flexible.
PAE is predicted aligned error in angstroms; low PAE between domains or chains supports confident relative placement.
pTM summarizes global fold/domain-packing confidence; ipTM focuses on multimer interfaces.

Common Examples

Single enzyme: monomer mode, Fast MSA search, pairwise confidence off, relaxation Best.
Multi-domain protein: monomer mode with pairwise confidence on; inspect PAE before trusting domain arrangement.
Homotrimer: multimer FASTA with the same sequence repeated three times; use Standard sampling for final review.

Example API params

{
  "mode": "monomer",
  "msa_search": "Fast",
  "relaxation": "Best"
}

Caveats

AlphaFold 2 does not explicitly model ligands, nucleic acids, cofactors, metals, post-translational modifications, or solvent.
High pLDDT within each chain does not guarantee correct domain-domain or chain-chain orientation.
Template cutoff controls templates, not a complete historical sequence database cutoff.

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 AlphaFold command-line flags such as FASTA paths, model preset, template date, MSA/database preset, and relaxation settings.
Use advanced submit for unusual AlphaFold flags or when reproducing an older command line; use canonical configuration for routine jobs.

curl -X POST https://subseq.bio/api/v1/job/submit \
  -H "Authorization: Bearer <api_key>" \
  -F program=alphafold2 \
  -F 'params={"mode":"monomer","msa_search":"Fast","relaxation":"Best"}'