MolEditRL: Structure-Preserving Molecular Editing via Discrete Diffusion and Reinforcement Learning

2025/05/26 by Yuanxin Zhuang, Zhuang, Yuanxin, Dazhong Shen +3 · 1 citation
Engineering · Medicine · #Innovative Microfluidic and Catalytic Techniques Innovation #Nanofabrication and Lithography Techniques #Monoclonal and Polyclonal Antibodies Research

paper · pdf · doi:10.48550/arxiv.2505.20131

Abstract

Molecular editing aims to modify a given molecule to optimize desired chemical properties while preserving structural similarity. However, current approaches typically rely on string-based or continuous representations, which fail to adequately capture the discrete, graph-structured nature of molecules, resulting in limited structural fidelity and poor controllability. In this paper, we propose MolEditRL, a molecular editing framework that explicitly integrates structural constraints with precise property optimization. Specifically, MolEditRL consists of two stages: (1) a discrete graph diffusion model pretrained to reconstruct target molecules conditioned on source structures and natural language instructions; (2) an editing-aware reinforcement learning fine-tuning stage that further enhances property alignment and structural preservation by explicitly optimizing editing decisions under graph constraints. For comprehensive evaluation, we construct MolEdit-Instruct, the largest and most property-rich molecular editing dataset, comprising 3 million diverse examples spanning single- and multi-property tasks across 10 chemical attributes. Experimental results demonstrate that MolEditRL significantly outperforms state-of-the-art methods in both property optimization accuracy and structural fidelity, achieving a 74% improvement in editing success rate while using 98% fewer parameters.

Citations

• GeLLMO: Generalizing Large Language Models for Multi-property Molecule Optimization
• Text-guided multi-property molecular optimization with a diffusion language model
• Instruction-Based Molecular Graph Generation with Unified Text-Graph Diffusion Model
• Rational Molecular Editing: A New Paradigm in Drug Discovery
• DrugAssist: A Large Language Model for Molecule Optimization
• Gotta be SAFE: A New Framework for Molecular Design
• BioT5: Enriching Cross-modal Integration in Biology with Chemical Knowledge and Natural Language Associations
• Training Diffusion Models with Reinforcement Learning
• Domain-Agnostic Molecular Generation with Chemical Feedback
• Multi-modal Molecule Structure-text Model for Text-based Retrieval and Editing
• Structured Denoising Diffusion Models in Discrete State-Spaces
• Therapeutics Data Commons: Machine Learning Datasets and Tasks for Drug Discovery and Development
• MIMOSA: Multi-constraint Molecule Sampling for Molecule Optimization
• Hierarchical Generation of Molecular Graphs using Structural Motifs
• RoBERTa: A Robustly Optimized BERT Pretraining Approach
• Randomized SMILES strings improve the quality of molecular generative models
• Graph Convolutional Policy Network for Goal-Directed Molecular Graph Generation
• Fréchet ChemNet Distance: A metric for generative models for molecules in drug discovery
• Junction Tree Variational Autoencoder for Molecular Graph Generation
• Molecular De Novo Design through Deep Reinforcement Learning
• Grammar Variational Autoencoder

Cited by

• Discrete Diffusion in Large Language and Multimodal Models: A Survey

Related

• MolEdit: Knowledge Editing for Multimodal Molecule Language Models
• MolAct: An Agentic RL Framework for Molecular Editing and Property Optimization
• MoleCode unlocks structural intelligence in large language models
• MolLangBench: A Comprehensive Benchmark for Language-Prompted Molecular Structure Recognition, Editing, and Generation
• Coder as Editor: Code-driven Interpretable Molecular Optimization