AbFlowNet: Optimizing Antibody-Antigen Binding Energy via Diffusion-GFlowNet Fusion

2025-05-18

AI Summary

Existing CDR design methods predominantly rely on reconstruction loss, neglecting binding affinity—a critical efficacy metric—while affinity optimization typically demands costly and unstable online reinforcement learning. This work introduces AbFlowNet, the first generative framework integrating GFlowNet with denoising diffusion models, where diffusion steps serve as state nodes and geometric reconstruction and true binding affinity are jointly optimized end-to-end in a single flow. AbFlowNet incorporates a physics-informed energy scoring function, enabling direct steering of generation by binding affinity signals—eliminating reliance on pseudo-labels or online RL paradigms. Experiments demonstrate significant improvements: amino acid recovery rate, RMSD, and binding affinity improvement rate increase by 3.06%, 20.40%, and 3.60%, respectively; top-1 total energy and binding affinity prediction errors decrease by 24.8% and 38.1%.

Technology Category

• Search and Optimization: Sampling/Simulation-based Search
• Machine Learning: Deep Generative Models & Autoencoders
• Natural Language Processing: Generation

Application Category

• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Graph Algorithms and Modeling for the Web: Representation, reconstruction, and subgraph or motif discovery in Web-related graphs
• Economics, Online Markets and Human Computation: Economic ramifications for generative AI infrastructure and applications

Abstract

Complementarity Determining Regions (CDRs) are critical segments of an antibody that facilitate binding to specific antigens. Current computational methods for CDR design utilize reconstruction losses and do not jointly optimize binding energy, a crucial metric for antibody efficacy. Rather, binding energy optimization is done through computationally expensive Online Reinforcement Learning (RL) pipelines rely heavily on unreliable binding energy estimators. In this paper, we propose AbFlowNet, a novel generative framework that integrates GFlowNet with Diffusion models. By framing each diffusion step as a state in the GFlowNet framework, AbFlowNet jointly optimizes standard diffusion losses and binding energy by directly incorporating energy signals into the training process, thereby unifying diffusion and reward optimization in a single procedure. Experimental results show that AbFlowNet outperforms the base diffusion model by 3.06% in amino acid recovery, 20.40% in geometric reconstruction (RMSD), and 3.60% in binding energy improvement ratio. ABFlowNet also decreases Top-1 total energy and binding energy errors by 24.8% and 38.1% without pseudo-labeling the test dataset or using computationally expensive online RL regimes.

Problem

• Optimizing antibody-antigen binding energy efficiently
• Improving CDR design without costly RL pipelines
• Unifying diffusion models and energy optimization jointly

Innovation

• Integrates GFlowNet with Diffusion models
• Optimizes diffusion losses and binding energy jointly
• Directly incorporates energy signals into training