ESMFold — single-sequence fold
Paste a FASTA, get a predicted structure with pLDDT. ESM-2 language-model fold, monomer-only — ~30 s per run.
What it is for
Pick ESMFold when you need the fastest possible monomer fold - no MSA, no multimer, single-sequence ESM-2 language-model prediction. Pair with ColabFold (D3) for multimers or AF2 standalone (D2) for full MSA-backed accuracy.
ESMFold (Lin et al., Science 2023). Single-sequence monomer structure prediction from the ESM-2 language model. No MSA, no multimer support — fastest fold available when an MSA is unavailable or unhelpful.
When it fits:
- You need a monomer fold in well under a minute.
- Your sequence has no detectable homologs (orphan or designed).
- You're triaging a large set of sequences and need throughput.
Inputs
You will need:
- Single FASTA sequence (monomer only).
- Sequence length under ~600 residues for best accuracy.
Each run uses a preset that sets the scale and scope:
- Standalone - your FASTA
- Paste a single-chain FASTA (10-400 aa monomer) and get pLDDT + predicted structure. ~30 s on A100-40GB once the 3B model is warm. No MSA, no multimer - pair with ColabFold (D3) or AF2 (D2) when you need those.
- Batch - many monomers
- Fold many monomer sequences (FASTA records or one per line) in a single job, up to 500 records. Each fold is shipped back through the partial-results contract; the results table renders per-design pLDDT, PDB download and NGL viewer as folds complete. Cost scales linearly with batch size.
Parameters you set on the form:
- Sequence
- Single-chain FASTA. Multimers and non-canonical residues are not supported — use ColabFold or AF2 instead.
Typical runtime:
- standalone
- ~30 s
How to read the results
Predicted PDB with per-residue pLDDT. No PAE (single-sequence prediction has no inter-domain signal). Use as a fast self-consistency check on designed sequences.
Where a tool reports them, the scores mean:
- ipTM
- Predicted confidence in the binder to target interface. Higher is better. Aim above roughly 0.7 on a tractable target.
- pLDDT
- Per-residue confidence in the predicted fold. Higher means the model is more sure of that part of the structure.
- i_pAE and pAE
- Predicted alignment error, at the interface (i_pAE) or across the whole structure (pAE). Lower is better.
Try these examples
One-click sample inputs that load straight into the run form. Edit any field before submitting.
- Ubiquitin (76 aa)
- Tiny monomer benchmark. ~30 s on the ESM-2 3B model; fastest possible feedback loop.
- Top7 de novo design (93 aa)
- Canonical de novo designed protein. Shows the ESM-2 single-sequence fold on a designed protein.
References
Lin et al., Science 2023