OmegaFold
Single-sequence protein folding with no MSA search, from a geometry-inspired language model.
About
OmegaFold (Wu et al., bioRxiv 2022) was one of the first models to fold proteins at near-AlphaFold2 accuracy from a single primary sequence, with no multiple sequence alignment. HeliXon's team built a geometry-inspired transformer around a protein language model, so the method works on orphan proteins and designed sequences where MSA-based folders have nothing to align.
The authors report accuracy on par with AlphaFold2 on recently released structures and stronger results on antibody loops and orphan proteins, though those comparisons are provider-claimed with no independent blind replication. The release is genuinely easy to run: a pip install, one FASTA-in/PDB-out command, and sharded execution to trade speed for GPU memory. The repository has been quiet since around 2023, and newer single-sequence folders have largely taken its place, but the code and Apache-2.0 weights still work.
Specs
Context window
Pricing
Free quota
API access
Weights
Released
Benchmarks
Author-reported (Wu et al., bioRxiv 2022): outperformed RoseTTAFold and matched AlphaFold2 on recently released structures; stronger than AlphaFold2 on antibody loops and orphan proteins (provider-claimed); no independent blind-benchmark replication published
Provider-reported unless a source says otherwise. Benchmarks are marketing-adjacent by nature — treat them as directional.
Use cases
Trust score — 7/10
Scored 7/10 (metrics tier, metrics-v1). Evidence: repo github.com/HeliXonProtein/OmegaFold (629 stars, Apache 2.0) and the bioRxiv paper confirm the single-sequence method and the reported accuracy; the one-command pip workflow is verified in the README. Strengths: genuinely MSA-free, fast, simple to run, monomer-friendly. Docked for: (1) no meaningful repo updates since ~2023; (2) accuracy claims are author-reported only; (3) newer single-sequence folders (ESMFold, Chai-1) have largely superseded it. License: Apache 2.0.
How the trust score is built
Full methodology →- Adoption · stars/downloads, log-scaled vs peers
- 7.0/10
- Maintenance · recency, releases, issue load
- 3.0/10
- Openness · license + docs
- 9.0/10
- Manual review · sanity read + red-flag check
- 7.0/10
- · Early MSA-free single-sequence folder; pip-installable with one-command inference.
- · No meaningful repo updates since ~2023.
- · Accuracy claims are author-reported only; superseded by newer single-sequence folders.