Abstract
Enhancing antibody affinity is a critical goal in antibody design, as it improves therapeutic efficacy, specificity, and safety while reducing dosage requirements. Traditional methods, such as single-point mutations or combinatorial mutagenesis, are limited by the impracticality of exhaustively exploring the vast mutational space. To address this challenge, we developed a novel computational pipeline that integrates evolutionary constraints, antibody-antigen-specific statistical potentials, molecular dynamics simulations, metadynamics, and a suite of deep learning models to identify affinity-enhancing mutations. Our deep learning framework includes MicroMutate, which predicts microenvironment-specific amino acid mutations, and graph-based models that evaluate postmutation antigen-antibody-binding probabilities. Using this approach, we screened 12 single-point mutant antibodies targeting the hemagglutinin of the H7N9 avian influenza virus, starting from antibodies with initial affinities in the subnanomolar range, with one showing a 4.62-fold improvement. To demonstrate the generalizability of our method, we applied it to engineer an antibody against death receptor 5 with initial affinities in the subnanomolar range, successfully identifying a mutant with a 2.07-fold increase in affinity. Our work underscores the transformative potential of integrating deep learning and computational methods for rapidly and precisely discovering affinity-enhancing mutations while preserving immunogenicity and expression. This approach offers a powerful and universal platform for advancing antibody therapeutics.