Synthesizable by Design: A Retrosynthesis-Guided Framework for Molecular Analog Generation

2025-07-03

AI Summary

AI-generated molecules often exhibit desirable properties but suffer from low synthetic feasibility, hindering drug and materials discovery. To address this, we propose SynTwins, a retrosynthesis-guided molecular generation framework that uniquely integrates retrosynthetic analysis, similarity-based building-block retrieval, and end-to-end virtual synthesis validation to strictly enforce synthetic accessibility while preserving target properties. SynTwins jointly optimizes the “property–structure–synthetic pathway” triad by coupling a high-accuracy retrosynthetic prediction model, reaction-rule-aware similarity search, and rigorous in silico synthesis verification. On multiple benchmark datasets, SynTwins significantly outperforms existing synthesis-aware generative models: it achieves a 32% improvement in synthetic accessibility (SAscore) while maintaining high structural similarity to reference molecules (Tanimoto > 0.8), and matches the property optimization performance of unconstrained generative models.

Technology Category

• Natural Language Processing: Code Generation / Program Synthesis from Natural Language
• Search and Optimization: Sampling/Simulation-based Search
• Cognitive Modeling & Cognitive Systems: Computational Creativity

Application Category

• Search and Retrieval-Augmented AI: Agentic search
• Graph Algorithms and Modeling for the Web: Foundation models and LLMs for Web-related graphs
• Web Mining and Content Analysis: Web data generation and simulation

Abstract

The disconnect between AI-generated molecules with desirable properties and their synthetic feasibility remains a critical bottleneck in computational drug and material discovery. While generative AI has accelerated the proposal of candidate molecules, many of these structures prove challenging or impossible to synthesize using established chemical reactions. Here, we introduce SynTwins, a novel retrosynthesis-guided molecular analog design framework that designs synthetically accessible molecular analogs by emulating expert chemist strategies through a three-step process: retrosynthesis, similar building block searching, and virtual synthesis. In comparative evaluations, SynTwins demonstrates superior performance in generating synthetically accessible analogs compared to state-of-the-art machine learning models while maintaining high structural similarity to original target molecules. Furthermore, when integrated with existing molecule optimization frameworks, our hybrid approach produces synthetically feasible molecules with property profiles comparable to unconstrained molecule generators, yet its synthesizability ensured. Our comprehensive benchmarking across diverse molecular datasets demonstrates that SynTwins effectively bridges the gap between computational design and experimental synthesis, providing a practical solution for accelerating the discovery of synthesizable molecules with desired properties for a wide range of applications.

Problem

• Bridges AI-generated molecules and synthetic feasibility
• Ensures synthesizability while maintaining molecular properties
• Guides molecular analog design via retrosynthesis strategies

Innovation

• Retrosynthesis-guided molecular analog design
• Emulates expert chemist strategies
• Ensures synthesizability while maintaining properties