Protein complex structure prediction

What is protein complex structure prediction?

Protein complex structure prediction is the computational process of building the three-dimensional arrangement and interfaces of two or more protein chains from their sequences and specified composition. It differs from docking because the model may predict partner folds and assembly jointly rather than starting from fixed monomer structures.

Complex prediction depends on more than individual fold quality. Researchers must specify the biologically relevant chains and stoichiometry, then judge whether the predicted interface is supported by interface confidence, model recurrence, known residues, and cellular context.

A plausible high-confidence interface is not proof that two proteins interact in vivo. Incorrect paralogs, missing cofactors, alternative oligomeric states, and training-set familiarity can all produce misleadingly neat assemblies.

When to use protein complex structure prediction

• Suitable protein complex structure prediction question. Oligomeric assemblies and protein interaction hypotheses when partner sequences are known
• Inputs and evidence are available. Correct partner sequences, chain multiplicity, and intended stoichiometry
• A validation plan is in place. Use interface-specific confidence and orthogonal interaction evidence. Validate decisive contacts with mutagenesis, crosslinking, binding measurements, or experimental structure determination.

Benefits of protein complex structure prediction

• Predicts folds and interfaces jointly.
• Supports alternative-model comparison.
• Generates explicit interaction hypotheses.

Primary limitations

• Stoichiometry must be specified correctly.
• Confident interfaces can be biologically irrelevant.
• Large assemblies may exceed model or compute limits.

Protein complex structure prediction methods and interpretation

Multimer predictors use learned geometric and evolutionary information to place chains jointly. Chain-pair PAE, ipTM-like scores, and interface residue confidence are generally more relevant than monomer-wide confidence when evaluating an assembly.

Compare alternative stoichiometries only when biologically plausible. A larger assembly can gain contacts simply because more chains were supplied, so conclusions need orthogonal evidence such as co-purification, crosslinking, mutagenesis, or cryo-EM.

How to run protein complex structure prediction online

The ProteinIQ workflow keeps the inputs, actual tool runs, method-native files, and comparison outputs together. Follow these steps while preserving the scientific boundary described above.

1. Define partners and stoichiometry. Create one FASTA record per chain and duplicate records only when the proposed stoichiometry requires it.
2. Prepare chain sequences. Run AlphaFold2 Multimer, ESMFold2, Boltz-2, and Protenix from the same input.
3. Run multimer models. Retain ranked models and every chain-aware confidence output.
4. Inspect interface confidence. Inspect interfaces, clashes, buried surface, PAE blocks, and agreement across model families.
5. Compare biological support. Prioritize assemblies that agree with biochemical and cellular evidence, then export models and provenance.

How protein complex structure prediction works

Run four multimer-capable predictors from the same multi-record FASTA and compare their interface confidence and assembly geometry.

1. Define partners and stoichiometry. Create one FASTA record per chain and duplicate records only when the proposed stoichiometry requires it.
2. Prepare chain sequences. Run AlphaFold2 Multimer, ESMFold2, Boltz-2, and Protenix from the same input.
3. Run multimer models. Retain ranked models and every chain-aware confidence output.
4. Inspect interface confidence. Inspect interfaces, clashes, buried surface, PAE blocks, and agreement across model families.
5. Compare biological support. Prioritize assemblies that agree with biochemical and cellular evidence, then export models and provenance.

Inputs and outputs

Check formats before running, then inspect and download the result from every workflow step.

Inputs

• Research input. FASTA PDB mmCIF A multi-record protein FASTA containing each intended chain with the correct multiplicity.

Outputs

• Prediction and review outputs. PDB mmCIF CSV JSON Ranked complex models, chain-aware confidence metrics, PAE or interface scores, and downloadable structures.