Errington David, Schneider Constantin, Bouysset Cédric, Dreyer Frédéric A
Recursion, Oxford, UK.
Exscientia, Oxford Science Park, Oxford, OX4 4GE, UK.
J Cheminform. 2025 May 19;17(1):76. doi: 10.1186/s13321-025-01011-6.
The field of protein-ligand pose prediction has seen significant advances in recent years, with machine learning-based methods now being commonly used in lieu of classical docking methods or even to predict all-atom protein-ligand complex structures. Most contemporary studies focus on the accuracy and physical plausibility of ligand placement to determine pose quality, often neglecting a direct assessment of the interactions observed with the protein. In this work, we demonstrate that ignoring protein-ligand interaction fingerprints can lead to overestimation of model performance, most notably in recent protein-ligand cofolding models which often fail to recapitulate key interactions.Scientific Contribution The interaction analysis used in this study is provided as a python package at https://github.com/Exscientia/plif_validity .
近年来，蛋白质-配体构象预测领域取得了显著进展，基于机器学习的方法如今已普遍用于替代传统对接方法，甚至用于预测全原子蛋白质-配体复合物结构。大多数当代研究专注于配体放置的准确性和物理合理性以确定构象质量，常常忽略对与蛋白质相互作用的直接评估。在这项工作中，我们证明忽略蛋白质-配体相互作用指纹可能导致对模型性能的高估，最明显的是在最近的蛋白质-配体共折叠模型中，这些模型常常无法重现关键相互作用。科学贡献 本研究中使用的相互作用分析作为一个Python包提供，网址为https://github.com/Exscientia/plif_validity 。
