Efficient Antibody Structure Refinement Using Energy-Guided SE(3) Flow Matching

TL;DR

A novel antibody structure refinement method termed FlowAB is developed based on energy-guided flow matching that achieves new state-of-the-art performance on the antibody structure prediction task when used in conjunction with an appropriate prior model while incurring only marginal computational overhead.

Abstract

Antibodies are proteins produced by the immune system that recognize and bind to specific antigens, and their 3D structures are crucial for understanding their binding mechanism and designing therapeutic interventions. The specificity of antibody-antigen binding predominantly depends on the complementarity-determining regions (CDR) within antibodies. Despite recent advancements in antibody structure prediction, the quality of predicted CDRs remains suboptimal. In this paper, we develop a novel antibody structure refinement method termed FlowAB based on energy-guided flow matching. FlowAB adopts the powerful deep generative method SE(3) flow matching and simultaneously incorporates important physical prior knowledge into the flow model to guide the generation process. The extensive experiments demonstrate that FlowAB can significantly improve the antibody CDR structures. It achieves new state-of-the-art performance on the antibody structure prediction task when used in conjunction with an appropriate prior model while incurring only marginal computational overhead. This advantage makes FlowAB a practical tool in antibody engineering.

Figures (7)

• Figure 1: Illustration of the six Complementarity-Determining Regions (CDRs) of an Antibody. The regions labeled are CDR-L1, L2, L3 on the light chain, and CDR-H1, H2, H3 on the heavy chain. These CDRs, being highly variable and playing a critical role in antigen recognition, are the focus of structural refinement in this work.
• Figure 2: Illustration of the parametrization of the antibody backbone.
• Figure 3: Flow matching with an energy guidance term $\nabla \mathcal{E}(x)$. The energy function gives guidance for the flow to reach a more physically plausible sample.
• Figure 4: Illustration of inter-atomic potentials: a) the improper torsion potential; b) the bond potential; c) the angle potential.
• Figure 5: The model architecture of FlowAB. The model takes in the residue embedding $\mathbf{h}_0$, pairwise embedding $\mathbf{z}_0$, SO(3) vector $\mathbf{r}_t$ and $\mathbb{R}^3$ vector $\mathbf{x}_t$ at time $t$.