Protein binder design

What is protein binder design?

Protein binder design is the process of creating proteins intended to recognize a specified molecular target or surface. Structure-based methods generate a binder backbone in the context of a target, optimize its sequence, and assess whether the candidate is predicted to fold and maintain the designed interface. The computational result is a prioritized binder hypothesis, not evidence of affinity, specificity, or biological activity.

Target preparation and epitope choice strongly influence the search. Missing loops, glycans, cofactors, oligomeric partners, or an incorrect conformational state can make a technically successful design irrelevant to the biological target.

Generate structural and sequence diversity, then inspect interface contacts, buried surface, clashes, binder self-folding, aggregation risk, and alternatives across the target surface. Prospective experiments must measure affinity and specificity rather than relying on model confidence.

When to use protein binder design

• Best fit. Research binders, affinity reagents, and therapeutic starting points against a known target surface
• Required starting evidence. A reviewed target structure, intended epitope, binder constraints, and binding assays

Benefits of protein binder design

• Focused search. Targets a specified surface
• Connected evidence. Generates new binding scaffolds
• Testable candidates. Connects backbone and sequence design

Primary limitations

• Model scope. Sensitive to target preparation
• Score uncertainty. Model confidence is not affinity
• Experimental requirement. Off-target binding needs experiments

Protein binder design methods

Binder generators optimize a candidate in the context of a target structure and often incorporate predicted complex confidence. Hotspot constraints can focus the interface, but overly narrow constraints may reduce diversity or favor strained solutions.

Sequence redesign and independent refolding test whether a generated interface can be supported by a plausible binder sequence. They do not establish that the target and binder will associate under experimental conditions.

How to run protein binder design online

Use the workflow as an inspectable computational funnel. Preserve the native output of each method, apply explicit acceptance gates, and keep the evidence behind every selected and rejected candidate.

1. Prepare target. Prepare the biologically relevant target structure, assembly, cofactors, glycans, and conformational state.
2. Define epitope. Define the intended epitope or hotspot residues and binder size, topology, and chemistry constraints.
3. Generate binders. Generate diverse target-conditioned backbones and multiple sequence variants for accepted designs.
4. Validate models. Review interface geometry, refold agreement, clashes, buried surface, solubility, and aggregation risk.
5. Measure binding. Express selected candidates and measure affinity, specificity, competition, and biological function.

How to evaluate protein binder design results

Inspect target relevance, epitope coverage, interface contacts, buried surface, clashes, shape complementarity, binder self-confidence, refold agreement, sequence diversity, solubility, and aggregation risk.

Experiments should include expression and monodispersity checks, quantitative affinity, specificity against related and unrelated proteins, competition or epitope mapping, and a function-relevant assay.

Experimental validation and handoff

Keep the exact target structure, assembly, hotspot definition, generation settings, sequence variants, and rejected candidates. Measure affinity, specificity, and function experimentally.

Export structures, sequences, settings, scores, logs, and selection criteria together. A reproducible handoff makes computational assumptions visible to the team planning synthesis, expression, biophysical characterization, and functional assays.

How protein binder design works

Generate target-conditioned binders with BindCraft, redesign sequences with ProteinMPNN, refold them, and inspect geometry.

1. Prepare target. Prepare the biologically relevant target structure, assembly, cofactors, glycans, and conformational state.
2. Define epitope. Define the intended epitope or hotspot residues and binder size, topology, and chemistry constraints.
3. Generate binders. Generate diverse target-conditioned backbones and multiple sequence variants for accepted designs.
4. Validate models. Review interface geometry, refold agreement, clashes, buried surface, solubility, and aggregation risk.
5. Measure binding. Express selected candidates and measure affinity, specificity, competition, and biological function.

Inputs and outputs

Check formats before running, then inspect and download the result from every workflow step.

Inputs

• Design input. PDB FASTA JSON TXT A target protein PDB with the intended biological assembly and optional hotspot residues.

Outputs

• Design and review outputs. PDB FASTA CSV JSON Binder backbone PDB files, designed FASTA sequences, refolded models, interface scores, and review results.