Overcoming complex therapeutic challenges requires next-gen precision and targeting for maximum efficacy, something that is often addressed by pursuing de novo antibody design. However, in silico antibody discovery models need accurate, high-resolution empirical structural data to correctly predict effective epitope/paratope interactions, especially for de novo antibody design. Here, we introduce an innovative approach that utilizes high-throughput, high-resolution analysis of higher-order structure (HOS) through radical footprinting mass spectrometry, enabling rapid, accurate characterization of epitopes/paratopes, conformational changes, and higher-order structure to generate a bespoke database of large libraries of antibody/antigen interactions. 

We then use these data to inform selection of binders with the greatest therapeutic potential. Through a combination of high throughput structural screening and empirically-constrained computational modeling, we systematically engineered antibodies that bind to a critical epitope on Human Fibroblast Activating Protein (hFAP), a historically challenging target, with exceptional affinity and specificity.