Pyrite 基准：LLM 智能体在蛋白-配体构象排序上超越专家
Pyrite: Building and Benchmarking Protein-Ligand Pose Ranking Agents
搞药物发现或计算化学的可以看看这篇：专家靠 3D 工具和物理打分排序分子构象几乎等于瞎猜，而两个 LLM 智能体分别做到 67% 和 74%。用代码生成特征做回归这套路子，比直接让模型推理更有效，开源模型经 GRPO 微调后还能追平闭源模型。
论文介绍 Pyrite，一个蛋白-配体结合构象（pose）排序基准，包含 1,500 个 pose 对，覆盖多样的蛋白-配体复合物。结果显示，配备 3D 可视化工具和基于物理打分的专家化学家表现接近随机水平，说明现有化学信号难以直接解读。
作者提出 PyriteRep，一种结构化表示分子几何、蛋白-配体相互作用与理化分数的格式，并构建两个 LLM 智能体：PyriteReasoner 直接通过文本推理比较构象，最高达到 67% 准确率；PyriteFeaturizer 生成可执行代码构建特征用于逻辑回归，最高达到 74%，超过专家表现。研究还发现 LLM 引导的特征发现整体优于直接推理，GRPO（Group Relative Policy Optimization）训练能让开源权重模型匹配或超过闭源前沿模型。
对做计算化学或药物发现的团队来说，这条工作流提供了一种用 LLM 解读和组合计算化学工具输出的思路，Pyrite 也成为衡量 pose 排序能力的新测试标准。
Pyrite: Building and Benchmarking Protein-Ligand Pose Ranking Agents
Predicting how a ligand binds to its target is central to small-molecule drug discovery. Modern cofolding models can generate diverse candidate poses, but reliably ranking these candidates remains challenging. We introduce Pyrite, a pairwise pose-ranking benchmark comprising 1,500 pose pairs across diverse protein-ligand complexes. Expert chemists equipped with 3D visualization tools and physics-based scores perform near chance, highlighting the difficulty of interpreting available chemical signals. To investigate automated ranking, we introduce PyriteRep, a structured representation of molecular geometry, protein-ligand interactions, and physicochemical scores, and develop two LLM-based agents: PyriteReasoner, which directly compares poses through text-based reasoning, and PyriteFeaturizer, which generates executable code to construct features for logistic regression, optionally incorporating features derived from LLM subcalls. These agents achieve up to 67% and 74% accuracy, respectively, exceeding observed expert performance. LLM-guided feature discovery generally outperforms direct reasoning, while Group Relative Policy Optimization (GRPO) improves open-weight models, enabling them to match or surpass closed-source frontier models in our evaluations. These results establish Pyrite as a challenging test of pose-ranking ability and highlight LLM-guided feature discovery as a promising approach to interpret and combine the outputs of computational chemistry tools.
This report is categorized under AITOP's "Papers" desk. Within our cross-regional analytical framework, it directly maps to AI developer toolchains, coding agents, and engineering workflows. Key entities and technical topics under continuous tracking include Pyrite, LLM智能体, 药物发现, GRPO.
Source attribution: curated from arXiv cs.LG (@Megha Srivastava, Michael H. Lim, Allison M. Keys, Pranav Murugan, Nate Gruver).