English static mirror for SEO/GEO · AI-assisted translation · Read Chinese original

Pyrite: Building and Benchmarking Protein-Ligand Pose Ranking Agents

Forum topic · 小凯 · 2026-11-12

Summary

Pyrite is a pairwise protein-ligand pose-ranking benchmark of 1,500 pose pairs across diverse protein-ligand complexes, introduced to address a key bottleneck in small-molecule drug discovery: reliably ranking candidate poses generated by modern cofolding models. Notably, expert chemists equipped with 3D visualization tools and physics-based scores perform near chance on the benchmark, underscoring how hard it is to interpret existing chemical signals. To enable automated ranking, the authors propose PyriteRep, a structured representation covering molecular geometry, protein-ligand interactions, and physicochemical scores, and build two LLM-based agents: PyriteReasoner, which compares poses via text-based reasoning, and PyriteFeaturizer, which generates executable code to construct features for logistic regression, optionally incorporating features from LLM sub-calls. The agents reach accuracies of 67% and 74%, exceeding human experts. LLM-guided feature discovery generally outperforms direct reasoning, and GRPO training lifts open-weight models to match or surpass closed-source frontier models. The work (arXiv:2611.14511) establishes Pyrite as a challenging testbed for pose-ranking capability.

Paper Overview

  • Field: Machine Learning
  • Authors: Megha Srivastava, Michael H. Lim, Allison M. Keys, Pranav Murugan, Nate Gruver
  • Released: 2026-11-10
  • arXiv: 2611.14511
  • Abstract

    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. The authors 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, they introduce PyriteRep, a structured representation of molecular geometry, protein-ligand interactions, and physicochemical scores, and develop two LLM-based agents:

  • PyriteReasoner: directly compares poses through text-based reasoning
  • PyriteFeaturizer: generates executable code to construct features for logistic regression, with the option to fuse features produced by LLM sub-calls
  • Key Findings

  • The two agents achieve accuracies of 67% and 74%, surpassing human experts on the benchmark.
  • LLM-guided feature discovery generally outperforms direct reasoning over structured representations.
  • GRPO training improves open-weight models, enabling them to match or exceed closed-source frontier models on pose ranking.

Significance

The results establish Pyrite as a challenging test of pose-ranking capability and highlight LLM-guided feature discovery as a promising direction for understanding and combining the outputs of computational chemistry tools.

---

*Auto-collected on 2026-11-12*

Tags

#machine-learning#drug-discovery#llm-agents#benchmarks#computational-chemistry#protein-ligand#reinforcement-learning

This page is an English static mirror generated for search and AI citation. It may be a full translation or structured summary of the Chinese original. Canonical interactive discussion lives on the Chinese page: https://zhichai.net/topic/178637065