Empower Structure-Based Molecule Optimization with Gradient Guided Bayesian Flow Networks
This work addresses structure-based molecule optimization for drug design, offering a versatile framework that extends to multi-objective optimization and challenging tasks like R-group optimization and scaffold hopping, though it appears incremental as it builds on gradient guidance methods from other domains.
The paper tackled the problem of optimizing molecules with both continuous coordinates and discrete types against protein targets by introducing a gradient-based framework that facilitates joint guidance signals across modalities while preserving SE(3)-equivariance. It achieved state-of-the-art performance on the CrossDocked2020 benchmark with a Success Rate of 51.3%, Vina Dock score of -9.05, and SA of 0.78, showing more than 4x improvement in Success Rate compared to gradient-based counterparts and 2x 'Me-Better' Ratio over 3D baselines.
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. Code is available at https://github.com/AlgoMole/MolCRAFT.