Protein engineering
Single-sequence protein structure prediction
Predict three-dimensional protein structures directly from one sequence without supplying alignments, templates, or homologous sequences.
What is single-sequence protein structure prediction?
Single-sequence protein structure prediction is the computational process of building a three-dimensional protein model from one amino-acid sequence without requiring a supplied multiple sequence alignment or template. Modern methods obtain structural signal from protein language models, but accuracy remains sequence- and region-dependent.
These models trade explicit user-supplied evolutionary input for representations learned from large sequence collections. They reduce setup time and are useful when MSA construction is slow or homologs are sparse.
No supplied MSA does not make a model evidence-free: training data still informs the representation. Compare predictors, retain native confidence outputs, and inspect domain packing, termini, and flexible regions carefully.
When to use single-sequence protein structure prediction
- Only one sequence is available. Use it for fast prediction and orphan proteins.
- MSA search latency is impractical. Language-model predictors can support high-throughput triage.
- Independent models can be reviewed. Compare native confidence and structural agreement before choosing a model.
Benefits of single-sequence protein structure prediction
- Only one sequence is required. No user-supplied alignment is needed.
- Prediction can be fast. It avoids a separate MSA-search setup.
- Many sequences can be triaged. Outputs support rapid structural hypothesis generation.
Primary limitations
- Domain orientations can be uncertain. A confident local region does not guarantee a reliable global arrangement.
- Disordered regions can look over-structured. Inspect native confidence and context.
- Confidence is not experimental validation. It does not establish a biological hypothesis.
How to run single-sequence protein structure prediction online
- Provide an ungapped sequence. Confirm the biological construct boundaries.
- Run the predictor panel. Submit it to ESMFold, ESMFold2, and MiniFold.
- Retain native outputs. Download every model and returned confidence values.
- Compare structural agreement. Inspect domain packing, loops, termini, and low-confidence regions.
- Record the selection. Choose a model or ensemble for the downstream question and document why.
How single-sequence protein structure prediction works
Submit one FASTA sequence to ESMFold, ESMFold2, and MiniFold and compare their structures and confidence.
- Clean the sequence. Provide one ungapped amino-acid sequence with the correct biological boundaries.
- Run language-model predictors. Run ESMFold, ESMFold2, and MiniFold from the shared workflow input.
- Collect native confidence. Download every model and retain pLDDT, PAE, or other native confidence outputs when returned.
- Compare model agreement. Align the predictions and inspect domain packing, termini, loops, and low-confidence segments.
- Select regions for validation. Choose a model or ensemble for the downstream question and record the selection rationale.
Inputs and outputs
Check formats before running, then inspect and download the result from every workflow step.
Inputs
Research input
FASTAPDBmmCIFOne protein sequence in FASTA or plain-text format.
Outputs
Prediction and review outputs
PDBmmCIFCSVJSONPredicted PDB or mmCIF structures with model-native confidence and downloadable files.
Tools for single-sequence protein structure prediction
Use these methods to prepare inputs, run the core analysis, inspect outputs, and validate the evidence described in this workflow.
ESMfold
Fast single-sequence structure prediction
ESMFold2
Language-model folding with mmCIF and confidence outputs
MiniFold
Rapid lightweight single-sequence prediction
IntelliFold 2
Controllable biomolecular prediction
AlphaFold2
Compare with free single-sequence or MSA-assisted modes
OpenFold-3
Alternative all-atom structure prediction
Boltz-2
Alternative biomolecular structure prediction
PDBFixer
Repair selected PDB models
MolProbity
Review stereochemical quality
Ramachandran plot
Inspect backbone torsions
USAlign
Measure agreement between predictions
DSSP
Assign coordinate-derived secondary structure
Other protein engineering workflows
Compare related approaches based on the molecular system, available evidence, required inputs, and decision you need to support.
Frequently asked questions
Single-sequence protein structure prediction is the computational process of building a three-dimensional protein model from one amino-acid sequence without requiring a supplied multiple sequence alignment or template. Modern methods obtain structural signal from protein language models, but accuracy remains sequence- and region-dependent.
Fast prediction, orphan proteins, sparse homologs, and high-throughput sequence triage. The required starting evidence is one ungapped protein sequence.
Keep each tool’s native confidence definition. Confidence estimates expected model error or consistency; it is not a probability that a biological hypothesis is true.
Treat low-confidence regions, inter-domain orientations, and long insertions as hypotheses. Cross-check conserved motifs and test important structural claims experimentally.
Public managed prediction services range from free noncommercial AlphaFold Server access to published core rates such as $125 for a first UW–Madison prediction. ProteinIQ Plus is $29 per month and Pro is $99 per month. Compute-heavy predictions consume credits according to the selected model and sequence length; expert project support is scoped separately.
Start with a workflow you can inspect and edit
Add your inputs, review the settings, and keep every structure, score, table, and file connected to the step that produced it.