We used Zernike3D descriptors to identify crystallographic antibodies that experimentally bind to epitopes with surface geometry similar to the Nipah virus epitope. This screening was performed using Ab-Set, a curated dataset comprising over 800,000 antibody structures and their corresponding molecular descriptors, including both experimentally determined and in silico–generated antibody–antigen complexes.

After identifying a prototype antibody with suitable geometric similarity, we submitted it to Ab-SELDON, a modular and easily customizable antibody design pipeline. Ab-SELDON enables iterative optimization of the antibody–antigen interaction through five modification stages, including CDR grafting, framework grafting, and targeted mutagenesis. The optimization process is guided by diversity data derived from millions of publicly available human antibody sequences.