Empower Structure-Based Molecule Optimization with Gradient Guidance

2024-11-20

AI Summary

Structure-based molecular optimization (SBMO) requires simultaneous optimization of 3D atomic coordinates and discrete atom types, yet the discreteness renders gradients undefined and often causes inter-modal inconsistency. This paper proposes MolJO, a novel framework that employs Bayesian latent-space modeling to construct a continuous, differentiable molecular representation while preserving SE(3)-equivariance—enabling joint gradient-guided optimization of geometry and topology. To balance exploration and exploitation, MolJO introduces a sliding-window backward correction mechanism. It further supports multi-objective optimization, R-group replacement, and scaffold hopping. On CrossDocked2020, MolJO achieves a success rate of 51.3%—four times that of gradient-based baselines—with AutoDock Vina score −9.05, synthetic accessibility (SA) score 0.78, and a Me-Better ratio twice that of state-of-the-art 3D-based methods.

Technology Category

• Search and Optimization: Mixed Discrete/Continuous Search
• Machine Learning: Optimization
• Reasoning under Uncertainty: Stochastic Optimization

Application Category

• Graph Algorithms and Modeling for the Web: Representation, reconstruction, and subgraph or motif discovery in Web-related graphs
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Systems and Infrastructure for Web, Mobile and WoT: Applied ML and AI for Web-based mobile applications

Abstract

Structure-Based molecule optimization (SBMO) aims to optimize molecules with both continuous coordinates and discrete types against protein targets. A promising direction is to exert gradient guidance on generative models given its remarkable success in images, but it is challenging to guide discrete data and risks inconsistencies between modalities. To this end, we leverage a continuous and differentiable space derived through Bayesian inference, presenting Molecule Joint Optimization (MolJO), the gradient-based SBMO framework that facilitates joint guidance signals across different modalities while preserving SE(3)-equivariance. We introduce a novel backward correction strategy that optimizes within a sliding window of the past histories, allowing for a seamless trade-off between explore-and-exploit during optimization. MolJO achieves state-of-the-art performance on CrossDocked2020 benchmark (Success Rate 51.3%, Vina Dock -9.05 and SA 0.78), more than 4x improvement in Success Rate compared to the gradient-based counterpart, and 2x"Me-Better"Ratio as much as 3D baselines. Furthermore, we extend MolJO to a wide range of optimization settings, including multi-objective optimization and challenging tasks in drug design such as R-group optimization and scaffold hopping, further underscoring its versatility.

Problem

• Optimize molecules with continuous and discrete properties for protein targets
• Address inconsistencies between modalities in gradient-guided generative models
• Enable versatile drug design tasks like multi-objective optimization and scaffold hopping

Innovation

• Continuous differentiable space via Bayesian inference
• Sliding window backward correction strategy
• SE(3)-equivariant joint gradient guidance framework