BoltzGen — structure + affinity design

Boltz-2 binder design. Generates mini-protein, nanobody, antibody, or peptide backbones against a target, refolds each candidate, and scores affinity via ipTM and pLDDT.

What it is for

Pick BoltzGen when you want one model that can design mini-proteins, nanobodies, antibodies, or peptides against the same target, or when your target involves glycans, post-translational modifications, or non-canonical residues.

BoltzGen (Wohlwend et al., MIT 2024). Boltz-2 binder design — jointly generates a binder backbone against a target, refolds each candidate end-to-end, and scores affinity via ipTM and pLDDT. Ships four design protocols (mini-protein, nanobody, antibody, peptide) and handles glycans, post-translational modifications, and non-canonical residues natively.

When it fits:

  • You want one model that can target the same epitope with mini-proteins, nanobodies, antibodies, or peptides.
  • Your target has glycans, PTMs, modified residues, or non-canonical chemistry.
  • You want refolding RMSD as a self-consistency signal alongside ipTM and pLDDT.
  • You need ~5 to 60 min per run and a budget-tunable number of candidates.

Inputs

You will need:

  • Target structure (.pdb / .cif).
  • Chain ID of the target.
  • At least one hotspot residue.

Each run uses a preset that sets the scale and scope:

Pilot — your target, ~30 min
Real BoltzGen run against your uploaded target. Up to 24 final candidates with refolding RMSD + ipTM scores; results emailed when complete (~15-60 min on A100-40GB).

Parameters you set on the form:

Protocol
Boltz-2 design protocol. protein-anything for general mini-protein binders, nanobody-anything for VHH scaffolds, antibody-anything for antibody scaffolds, peptide-anything for short cyclic or linear peptides.
Hotspot residues
Comma-separated target-chain residue indices the binder should contact. Click residues in the 3D viewer to toggle.
Binder length (min/max)
Residue-count window for the generated binder. Typical starting ranges: mini-protein 50–100, nanobody 110–130, antibody 110–200, peptide 5–30.
Budget (designs)
Number of designs Boltz-2 generates and ranks. Higher budgets cost more and run longer.

Typical runtime:

pilot
15–60 min

How to read the results

Ranked candidate binders with ipTM, pLDDT, refolding RMSD, and downloadable PDBs. Refolding RMSD < 2 Å on the top design typically signals self-consistent binding.

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.

SARS-CoV-2 RBD (6m0j chain E)
Boltz-2 binder design against the spike RBD. Same hotspots as the BindCraft / RFdiffusion examples for easy comparison.

References

Wohlwend et al., MIT (2024)

Open the BoltzGen — structure + affinity design form All guides