Boltz-2
Co-folding with built-in binding affinity prediction
Boltz-2 co-folds proteins with ligands, nucleic acids and cofactors while predicting micromolar binding affinities alongside structures. Trained on PDB complexes with affinity labels, it returns docked poses, confidence scores and estimated IC50 values. Useful for rapid triage of binder designs and virtual screening hitlists.
Boltz-2 is live on FoldForge — paste a sequence and press Fold.
At a glance
- Input
- Sequences and ligands (FASTA, SDF)
- Output
- Poses with affinity (mmCIF)
- Developed by
- MIT
- Published
- Passaro et al., 2025 · 2025
#co-folding#affinity#diffusion
Related tools
More methods in Structure Prediction & Folding
