发表机构
Université Paris-Saclay; INRAE; AgroParisTech（巴黎萨克雷大学; 法国国家农业食品与环境研究院; 巴黎高科农业学院）
机构由 AI 辅助整理，请以论文原文为准。AI 中文总结
该研究提出基于学习前沿选择的智能体逆生物合成框架，经微调的Qwen2.5-7B策略在多个基准上优于MCTS等方法，可提升预算内逆合成搜索性能。
AI 中文摘要
大型语言模型越来越多地被用作多步逆合成的智能体，由此产生了一个问题：在多大程度上其搜索策略独立于底层反应模型发挥作用。我们在生物背景下通过基于规则的逆生物合成研究该问题：确定性生化引擎为所有方法生成相同的验证过的转化，搜索终止于大肠杆菌底盘可用代谢物的路线，而策略仅选择接下来要扩展的前沿分子。经提示和LoRA微调的Qwen2.5-7B策略使用仅包含严格选择的接口。在LASER基准上，微调策略在10次扩展时达到65±1%的解决率，而MCTS为59%；在200次扩展时，在LASER基准上达到78±1%，在RetroPath RL黄金基准上达到88±3%，在BioNavi-NP基准上达到63±2%，而对应MCTS分别为75%、80%、45%。微调策略也始终优于直接提示策略。这些结果表明，路线监督的前沿选择可在不改变生化生成的情况下提升预算内搜索的性能，尽管性能仍依赖于前沿构建和反应排序。
英文摘要
Large language models are increasingly used as agents for multistep retrosynthesis, raising the question of how much their search policy contributes independently of the underlying reaction model. We investigate this question in a biological setting through rule-based retrobiosynthesis: a deterministic biochemical engine generates the same validated transitions for every method, searching for routes that terminate in metabolites available to an \emph{Escherichia coli} chassis, while the policy only selects which frontier molecule to expand next. Prompted and LoRA-tuned Qwen2.5-7B policies use a strict choice-only interface. The fine-tuned policy reaches $65\pm1$\% solve rate at 10 expansions on LASER versus 59\% for MCTS, and at 200 expansions reaches $78\pm1$\% versus 75\% on LASER, $88\pm3$\% versus 80\% on the RetroPath RL Golden benchmark, and $63\pm2$\% versus 45\% on the BioNavi-NP benchmark. Fine-tuning also consistently outperforms direct prompting. These results show that route-supervised frontier selection can improve budgeted search without altering biochemical generation, although performance remains dependent on frontier construction and reaction ranking.
