RFantibody

Pick RFantibody when you need a VHH (nanobody) scaffold against a target PDB. For de novo non-antibody binders, use BindCraft. For designs involving modified residues or glycans, use BoltzGen.

RFantibody (Bennett et al., bioRxiv 2024). RoseTTAFold-derived diffusion model that generates VHH (single-domain heavy-chain antibody) scaffolds against a target. Outputs are scored with AF2 re-prediction (pAE, pLDDT, ipAE).

When it fits:

• You want a VHH (nanobody) scaffold rather than a de novo mini-protein.
• Your downstream validation uses yeast display, mammalian display, or hybridoma workflows.
• Your target is a standard protein epitope without heavy glycosylation.

You will need:

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

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

Your target, ~30 min start to first results

Real RFantibody design against your uploaded target PDB. Pick 1 to 1000 final VHH candidates. Start with a small batch (4 designs, ~30 to 60 min) to confirm your target and hotspots, then scale to 100+ once outputs look real. Results emailed when run completes; A100-80GB.

Ranked VHH candidates with pAE, pLDDT, ipAE, and downloadable PDBs. Filter at pAE ≤ 5 / ipAE ≤ 6 for downstream wet-lab work.

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.

Bennett et al., bioRxiv 2024