Biocenter Kuopio (BCK) and Department of Pharmaceutical Chemistry, University of Kuopio, P.O. Box 1627, 70211 Kuopio, Finland.
J Chem Inf Model. 2009 Dec;49(12):2742-8. doi: 10.1021/ci900364w.
As tautomerism and ionization may significantly change the interaction possibilities between a ligand and a target protein, these phenomena could have an effect on structure-based virtual screening. Tautomeric- and protonation-state enumeration ensures that the state with optimal interaction possibilities is included in the screening process, as the predicted state may not always be the optimal binder. However, there is very little information published if tautomer and protomer enumeration actually improves the enrichment of active molecules compared to the alternative of using a predicted form of each molecule. In this study, a retrospective virtual screening was performed using AutoDock on 19 drug targets with a publicly available data set. It is proposed that tautomer and protomer prediction can significantly save computing resources and can yield similar results to enumeration.
由于互变异构和离子化可能会显著改变配体与靶蛋白之间的相互作用可能性，因此这些现象可能会对基于结构的虚拟筛选产生影响。互变异构体和质子化态枚举可确保将具有最佳相互作用可能性的状态包含在筛选过程中，因为预测的状态不一定总是最佳配体。然而，如果与使用每种分子的预测形式相比，互变异构体和前体枚举实际上可以提高活性分子的富集度，那么发表的相关信息却很少。在这项研究中，使用 AutoDock 对具有公开数据集的 19 个药物靶标进行了回顾性虚拟筛选。本文提出，互变异构体和前体预测可以显著节省计算资源，并产生与枚举相似的结果。
