MolecularCanvas: LLM-assisted Small-Molecule Drug Discovery via Structure-Guided Constraints

2026-07-31

AI Summary

This work addresses the limitations of existing generative AI tools in supporting precise, transparent, and evaluable iterative optimization of molecular structures during drug discovery. The authors propose an interactive system that, for the first time, enables structure-guided context construction by integrating high-level design objectives, structural annotations, property constraints, and reference preferences to steer large language models toward generating small-molecule candidates aligned with expert intent. The system seamlessly incorporates computational evaluation tools into a closed-loop workflow and combines interactive molecular editing, property prediction, and evidence tracing to significantly enhance the interpretability and practical utility of AI-generated suggestions. User studies demonstrate that the system effectively assists medicinal chemists in efficiently refining candidate molecules, thereby improving both design efficiency and decision confidence in real-world drug discovery scenarios.

Abstract

Small-molecule drug discovery relies on iterative molecular optimization, where chemists repeatedly modify candidate compounds to balance multiple competing properties such as efficacy, toxicity, and solubility. Recent advances in generative AI (GenAI) have shown promise in accelerating this process by automatically proposing new molecular structures or targeted modifications. However, existing GenAI-based molecular design tools remain poorly aligned with experts' real-world workflows. Specifically, they offer limited support for specifying structure-level modification intents on molecules, provide insufficient transparency into model-generated modifications, and lack integrated support for downstream property evaluation with external computational tools. To address these challenges, we introduce MolecularCanvas, an interactive system that enables users to iteratively construct an optimization context by integrating high-level goals, structure-level annotations, property constraints, and reference-based preferences. This context guides the generation of candidate molecules across diverse molecular structures. MolecularCanvas further enhances transparency by providing evidence for AI-generated suggestions and streamlines molecular evaluation by integrating commonly used computational tools for property assessment into a unified interface. Finally, a user study with 12 participants demonstrates the usefulness and effectiveness of MolecularCanvas in helping users optimize candidate molecules.

Problem

• small-molecule drug discovery
• generative AI
• molecular optimization
• structure-guided constraints
• interactive molecular design

Innovation

• LLM-assisted drug discovery
• structure-guided molecular generation
• interactive molecular design
• transparent AI in chemistry
• integrated property evaluation