Empower Structure-Based Molecule Optimization with Gradient Guided Bayesian Flow Networks
Keyue Qiu, Yuxuan Song, Jie Yu, Hongbo Ma, Ziyao Cao, Zhilong Zhang, Yushuai Wu, Mingyue Zheng, Hao Zhou, Wei-Ying Ma
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 their 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, allowing for a seamless trade-off between exploration and exploitation 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 4× improvement in Success Rate compared to the gradient-based counterpart, and 2× "Me-Better" Ratio as high as 3D baselines. Furthermore, we extend MolJO to a wide range of 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.
Ask Lune
Know this paper in minutes, not an afternoon.
Lune can read every page of it and the top-tier papers that cite it, and gives you its claims, its evidence and its limits, each one cited to the paper it comes from. No skimming, no AI slop.
1 top-tier paper has cited it since. Know what they found before your reviewers do.
Lune can draw on
the full paper · the top-tier papers that cite it · related work
Free to start. No credit card required.
Lineage
- Towards Docking-oriented De Novo Ligand Design via Gradient Inversion cites Pocket2Mol: Efficient Molecular Sampling Based on 3D Protein Pockets
- Towards Docking-oriented De Novo Ligand Design via Gradient Inversion cites MARS: Markov Molecular Sampling for Multi-objective Drug Discovery
- Towards Docking-oriented De Novo Ligand Design via Gradient Inversion cites Flow Network based Generative Models for Non-Iterative Diverse Candidate Generation
- Diffusion Self-Guidance for Controllable Image Generation cites Diffusion Models Beat GANs on Image Synthesis
- Flow Network based Generative Models for Non-Iterative Diverse Candidate Generation cites MARS: Markov Molecular Sampling for Multi-objective Drug Discovery
Cited by top-tier papers1
Builds on25
- Diffusion Models Beat GANs on Image SynthesisPrafulla Dhariwal, Alexander Quinn NicholNeurIPS 2021 · 13,211 citations
- Flow Network based Generative Models for Non-Iterative Diverse Candidate GenerationEmmanuel Bengio, Moksh Jain, Maksym Korablyov, Doina Precup et al.NeurIPS 2021 · 565 citations
- Diffusion Self-Guidance for Controllable Image GenerationDave Epstein, Allan Jabri, Ben Poole, Alexei A. Efros et al.NeurIPS 2023 · 411 citations
- Pocket2Mol: Efficient Molecular Sampling Based on 3D Protein PocketsXingang Peng, Shitong Luo, Jiaqi Guan, Qi Xie et al.ICML 2022 · 291 citations
- MARS: Markov Molecular Sampling for Multi-objective Drug DiscoveryYutong Xie, Chence Shi, Hao Zhou, Yuwei Yang et al.ICLR 2021 · 186 citations
Related papers
- Piloting Structure-Based Drug Design via Modality-Specific Optimal ScheduleKeyue Qiu, Yuxuan Song, Zhehuan Fan, Peidong Liu et al.ICML 2025
- MolCRAFT: Structure-Based Drug Design in Continuous Parameter SpaceYanru Qu, Keyue Qiu, Yuxuan Song, Jingjing Gong et al.ICML 2024 · 57 citations
- Prior-Guided Flow Matching for Target-Aware Molecule Design with Learnable Atom NumberJingyuan Zhou, Hao Qian, Shikui Tu, Lei XuNeurIPS 2025 · 11 citations
- Compositional Flows for 3D Molecule and Synthesis Pathway Co-designTony Shen, Seonghwan Seo, Ross Irwin, Kieran Didi et al.ICML 2025
- EvoEGF-Mol: Evolving Exponential Geodesic Flow for Structure-based Drug DesignYaowei Jin, Junjie Wang, Cheng Cao, Penglei Wang et al.ICML 2026
