Attentive geometric learning improves antibody design accuracy and affinity
AbGaze: Attentive Geometric Representation Learning for End-to-End Antibody Design
Artificial Intelligence
Summary
Designing antibodies that fit well with antigens is important for medicine, but it requires understanding complex 3D shapes and interactions. The authors created AbGaze, a method that looks closely at distances, directions, and surface shapes to better capture these details. This helps design antibodies more accurately, predict their structures, and improve how strongly they bind. AbGaze works better than earlier methods in several tests involving antibody parts, docking, and strength of binding.
What this means in practice
- •For biotechnology engineers: Design antibodies with improved accuracy and binding strength by capturing detailed geometric interactions at the antigen-antibody interface.$Commercial implications: Enables development of better therapeutic antibodies with enhanced target binding, creating commercially valuable biologic drugs.
- •For computational drug designers: Predict antibody structures and optimize sequences more reliably to accelerate drug discovery workflows involving protein interactions.
Authors
Jiashuo Wang, Siqi Fan, Yizhen Luo, Zaiqing Nie
Abstract
Computational antibody design requires representations that capture the geometric patterns underlying antigen--antibody interactions, yet existing approaches often rely on scalar distances or surface-intrinsic features, leaving cross-molecular geometry largely implicit. We present AbGaze, an end-to-end antibody design framework based on attentive geometric representation learning, which encodes distance, spatial direction, and surface-normal orientation of antigen surfaces relative to antibody-residue local frames, and adaptively aggregates these geometric interactions according to their interfacial context. The learned interaction representation is shared across multi-CDR co-design, complex structure prediction, and affinity optimization, with local-frame geometric supervision further constraining the representation. AbGaze outperforms prior methods across all three tasks: relative to the second-best method, it improves amino-acid recovery by 7.1% and reduces structural error by 14.9% on average over the six CDRs, improves interface docking quality (DockQ) by 6.6%, and raises the affinity improvement rate (IMP) by 32.5%.