Saved in:
Bibliographic Details
Main Authors: Zhu, Wenyu, Li, Chengzhu, Tian, Xiaohe, Wang, Yifan, Jia, Yinjun, Wang, Jianhui, Gao, Bowen, Zhang, Ya-Qin, Ma, Wei-Ying, Lan, Yanyan
Format: Preprint
Published: 2025
Subjects:
Online Access:https://arxiv.org/abs/2510.14455
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866918161724473344
author Zhu, Wenyu
Li, Chengzhu
Tian, Xiaohe
Wang, Yifan
Jia, Yinjun
Wang, Jianhui
Gao, Bowen
Zhang, Ya-Qin
Ma, Wei-Ying
Lan, Yanyan
author_facet Zhu, Wenyu
Li, Chengzhu
Tian, Xiaohe
Wang, Yifan
Jia, Yinjun
Wang, Jianhui
Gao, Bowen
Zhang, Ya-Qin
Ma, Wei-Ying
Lan, Yanyan
contents Molecular optimization is a central task in drug discovery that requires precise structural reasoning and domain knowledge. While large language models (LLMs) have shown promise in generating high-level editing intentions in natural language, they often struggle to faithfully execute these modifications-particularly when operating on non-intuitive representations like SMILES. We introduce MECo, a framework that bridges reasoning and execution by translating editing actions into executable code. MECo reformulates molecular optimization for LLMs as a cascaded framework: generating human-interpretable editing intentions from a molecule and property goal, followed by translating those intentions into executable structural edits via code generation. Our approach achieves over 98% accuracy in reproducing held-out realistic edits derived from chemical reactions and target-specific compound pairs. On downstream optimization benchmarks spanning physicochemical properties and target activities, MECo substantially improves consistency by 38-86 percentage points to 90%+ and achieves higher success rates over SMILES-based baselines while preserving structural similarity. By aligning intention with execution, MECo enables consistent, controllable and interpretable molecular design, laying the foundation for high-fidelity feedback loops and collaborative human-AI workflows in drug discovery.
format Preprint
id arxiv_https___arxiv_org_abs_2510_14455
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Coder as Editor: Code-driven Interpretable Molecular Optimization
Zhu, Wenyu
Li, Chengzhu
Tian, Xiaohe
Wang, Yifan
Jia, Yinjun
Wang, Jianhui
Gao, Bowen
Zhang, Ya-Qin
Ma, Wei-Ying
Lan, Yanyan
Machine Learning
Biomolecules
Molecular optimization is a central task in drug discovery that requires precise structural reasoning and domain knowledge. While large language models (LLMs) have shown promise in generating high-level editing intentions in natural language, they often struggle to faithfully execute these modifications-particularly when operating on non-intuitive representations like SMILES. We introduce MECo, a framework that bridges reasoning and execution by translating editing actions into executable code. MECo reformulates molecular optimization for LLMs as a cascaded framework: generating human-interpretable editing intentions from a molecule and property goal, followed by translating those intentions into executable structural edits via code generation. Our approach achieves over 98% accuracy in reproducing held-out realistic edits derived from chemical reactions and target-specific compound pairs. On downstream optimization benchmarks spanning physicochemical properties and target activities, MECo substantially improves consistency by 38-86 percentage points to 90%+ and achieves higher success rates over SMILES-based baselines while preserving structural similarity. By aligning intention with execution, MECo enables consistent, controllable and interpretable molecular design, laying the foundation for high-fidelity feedback loops and collaborative human-AI workflows in drug discovery.
title Coder as Editor: Code-driven Interpretable Molecular Optimization
topic Machine Learning
Biomolecules
url https://arxiv.org/abs/2510.14455