RetroAgent: Harnessing LLMs to Search Over Structured Memory for Agentic Retrosynthesis Planning

2026-07-15

AI Summary

This work addresses the challenge of multi-step retrosynthetic planning, where the vast combinatorial search space renders traditional methods ineffective at evaluating the feasibility of complete synthetic routes. The authors propose a large language model (LLM) agent that integrates symbolic search with neural reasoning and introduces, for the first time, a structured memory mechanism. This mechanism dynamically records explored pathways, candidate transformations, and physicochemical properties of intermediates, while leveraging cheminformatics tools to enable global pathway awareness and seamless integration of domain knowledge. Evaluated on both in-distribution and out-of-distribution benchmarks, the proposed approach significantly outperforms existing methods, demonstrating enhanced accuracy and generalization in multi-step retrosynthetic route planning.

Abstract

Multi-step retrosynthesis planning seeks to decompose a target molecule into commercially available building blocks through a sequence of feasible reactions. The vast combinatorial search space makes this task challenging even for expert chemists. Traditional methods combine tree search with offline-trained value networks that score candidates in isolation, without reasoning about complete multi-step routes. Recent work leverages Large Language Models (LLMs) for this task, but relies on simple interfaces that limit exploration of the full search space. We introduce RetroAgent, an LLM agent that bridges symbolic search and neural reasoning through a harness with structured memory. Through memory and chemistry tools, the agent observes the full search state, including explored routes, available alternatives, and properties of intermediates, enabling informed decisions grounded in both global progress and domain knowledge. Experiments on in-distribution and out-of-distribution benchmarks demonstrate that RetroAgent delivers strong performance and generalization.

Problem

• retrosynthesis planning
• combinatorial search space
• multi-step synthesis
• molecular decomposition

Innovation

• RetroAgent
• Large Language Models
• structured memory
• retrosynthesis planning
• agentic reasoning