How to Use BindCraft2

BindCraft2 expands the BindCraft family into a campaign-based binder design suite with multiple binder modalities. It supports de novo miniproteins, larger binders, peptides, cyclic peptides, homo-oligomers, multidomain binders, VHHs, ankyrin repeat proteins, scFvs, Fabs, and conformational objectives such as induced-fit or fold-switching binders.

On Neurosnap, researchers upload a Target Structure, optionally restrict Target Chains, and can define Hotspots or Coldspots using uploaded PDB numbering. Binder Format, length bounds, final-design count, and optional biological properties configure the campaign while runtime-sensitive settings and output paths remain managed by Neurosnap.

BindCraft2 combines AlphaFold2-guided sequence optimization, ProteinMPNN redesign, held-out AlphaFold validation, and structural filtering. Campaign outputs are organized by stage: trajectories record gradient-design attempts, refolded candidates capture redesigned sequences and failed filters, and ranked results contain accepted designs sorted for follow-up review.

The Neurosnap form separates target setup, binder modality, biological design properties, and shared advanced options. Properties such as Focused Epitope, Humanize Binder, Protease Stable, Disulfide Staple, Mixed Topology, Nearby Termini, and Accessible Termini map to BindCraft2 presets that alter objectives or acceptance criteria. Maximum Trajectories can cap exploration, while Campaign Seed supports reproducibility.

Results should be treated as computational design evidence, not experimental validation. ranked_designs.csv summarizes accepted binders, optional candidate and trajectory tables explain how the campaign searched, and predicted complex structures support manual inspection of interface geometry, target accessibility, developability risks, and candidate diversity before synthesis or assay testing.

• Broad modality coverage: BindCraft2 handles de novo binders, peptides, cyclic peptides, oligomers, scaffolded binders, and antibody-like formats in one campaign model.
• Epitope-aware inputs: Target chains, hotspots, coldspots, focused epitope targeting, and termini objectives let researchers align the search with biological constraints.
• Integrated validation: ProteinMPNN redesign, held-out structure prediction, and structural filters help prioritize candidates beyond raw sequence generation.
• Readable campaign outputs: Ranked designs, refolded candidates, trajectory summaries, and predicted structures make it easier to triage binders for expression, biophysical screening, and downstream optimization.