Protein Binder Design

Protein binder design aims to create proteins—such as antibodies, nanobodies, and peptides—that attach to chosen targets, potentially enabling new therapeutics and research tools. Methods include diffusion and flow models that jointly generate sequences and structures, language-model and structure-predictor-guided design, and optimization or ranking of candidates for target fit, affinity, and selectivity. Benchmarks increasingly test unseen targets and epitopes, structural diversity, and computational cost, while experiments assess whether predicted binders actually bind. A key challenge is that structure-based confidence scores do not reliably predict experimental affinity or generalize across target families, so efficient design still needs robust validation and better measures of function.