Dataset Information

Correcting the impact of docking pose generation error on binding affinity prediction.

ABSTRACT: Pose generation error is usually quantified as the difference between the geometry of the pose generated by the docking software and that of the same molecule co-crystallised with the considered protein. Surprisingly, the impact of this error on binding affinity prediction is yet to be systematically analysed across diverse protein-ligand complexes.Against commonly-held views, we have found that pose generation error has generally a small impact on the accuracy of binding affinity prediction. This is also true for large pose generation errors and it is not only observed with machine-learning scoring functions, but also with classical scoring functions such as AutoDock Vina. Furthermore, we propose a procedure to correct a substantial part of this error which consists of calibrating the scoring functions with re-docked, rather than co-crystallised, poses. In this way, the relationship between Vina-generated protein-ligand poses and their binding affinities is directly learned. As a result, test set performance after this error-correcting procedure is much closer to that of predicting the binding affinity in the absence of pose generation error (i.e. on crystal structures). We evaluated several strategies, obtaining better results for those using a single docked pose per ligand than those using multiple docked poses per ligand.Binding affinity prediction is often carried out on the docked pose of a known binder rather than its co-crystallised pose. Our results suggest than pose generation error is in general far less damaging for binding affinity prediction than it is currently believed. Another contribution of our study is the proposal of a procedure that largely corrects for this error. The resulting machine-learning scoring function is freely available at http://istar.cse.cuhk.edu.hk/rf-score-4.tgz and http://ballester.marseille.inserm.fr/rf-score-4.tgz .

SUBMITTER: Li H

PROVIDER: S-EPMC5046193 | biostudies-literature | 2016 Sep

REPOSITORIES: biostudies-literature

ACCESS DATA

Publications

Correcting the impact of docking pose generation error on binding affinity prediction.

Li Hongjian H Leung Kwong-Sak KS Wong Man-Hon MH Ballester Pedro J PJ

BMC bioinformatics 20160922 Suppl 11

<h4>Background</h4>Pose generation error is usually quantified as the difference between the geometry of the pose generated by the docking software and that of the same molecule co-crystallised with the considered protein. Surprisingly, the impact of this error on binding affinity prediction is yet to be systematically analysed across diverse protein-ligand complexes.<h4>Results</h4>Against commonly-held views, we have found that pose generation error has generally a small impact on the accuracy ...[more]

PMID: 28185549

Similar Datasets

Project description:Protein-ligand interactions are increasingly profiled at high-throughput, playing a vital role in lead compound discovery and drug optimization. Accurate prediction of binding pose and binding affinity constitutes a pivotal challenge in advancing our computational understanding of protein-ligand interactions. However, inherent limitations still exist, including high computational cost for conformational search sampling in traditional molecular docking tools, and the unsatisfactory molecular representation learning and intermolecular interaction modeling in deep learning-based methods. Here we propose a geometry-aware attention-based deep learning model, GAABind, which effectively predicts the pocket-ligand binding pose and binding affinity within a multi-task learning framework. Specifically, GAABind comprehensively captures the geometric and topological properties of both binding pockets and ligands, and employs expressive molecular representation learning to model intramolecular interactions. Moreover, GAABind proficiently learns the intermolecular many-body interactions and simulates the dynamic conformational adaptations of the ligand during its interaction with the protein through meticulously designed networks. We trained GAABind on the PDBbindv2020 and evaluated it on the CASF2016 dataset; the results indicate that GAABind achieves state-of-the-art performance in binding pose prediction and shows comparable binding affinity prediction performance. Notably, GAABind achieves a success rate of 82.8% in binding pose prediction, and the Pearson correlation between predicted and experimental binding affinities reaches up to 0.803. Additionally, we assessed GAABind's performance on the severe acute respiratory syndrome coronavirus 2 main protease cross-docking dataset. In this evaluation, GAABind demonstrates a notable success rate of 76.5% in binding pose prediction and achieves the highest Pearson correlation coefficient in binding affinity prediction compared with all baseline methods.

Project description:The protein targets for general anesthetics remain unclear. A tool to predict anesthetic binding for potential binding targets is needed. In this study, we explored whether a computational method, AutoDock, could serve as such a tool.High-resolution crystal data of water-soluble proteins (cytochrome C, apoferritin, and human serum albumin), and a membrane protein (a pentameric ligand-gated ion channel from Gloeobacter violaceus [GLIC]) were used. Isothermal titration calorimetry (ITC) experiments were performed to determine anesthetic affinity in solution conditions for apoferritin. Docking calculations were performed using DockingServer with the Lamarckian genetic algorithm and the Solis and Wets local search method (http://www.dockingserver.com/web). Twenty general anesthetics were docked into apoferritin. The predicted binding constants were compared with those obtained from ITC experiments for potential correlations. In the case of apoferritin, details of the binding site and their interactions were compared with recent cocrystallization data. Docking calculations for 6 general anesthetics currently used in clinical settings (isoflurane, sevoflurane, desflurane, halothane, propofol, and etomidate) with known 50% effective concentration (EC(50)) values were also performed in all tested proteins. The binding constants derived from docking experiments were compared with known EC(50) values and octanol/water partition coefficients for the 6 general anesthetics.All 20 general anesthetics docked unambiguously into the anesthetic binding site identified in the crystal structure of apoferritin. The binding constants for 20 anesthetics obtained from the docking calculations correlate significantly with those obtained from ITC experiments (P = 0.04). In the case of GLIC, the identified anesthetic binding sites in the crystal structure are among the docking predicted binding sites, but not the top ranked site. Docking calculations suggest a most probable binding site located in the extracellular domain of GLIC. The predicted affinities correlated significantly with the known EC(50) values for the 6 frequently used anesthetics in GLIC for the site identified in the experimental crystal data (P = 0.006). However, predicted affinities in apoferritin, human serum albumin, and cytochrome C did not correlate with these 6 anesthetics' known experimental EC(50) values. A weak correlation between the predicted affinities and the octanol/water partition coefficients was observed for the sites in GLIC.We demonstrated that anesthetic binding sites and relative affinities can be predicted using docking calculations in an automatic docking server (AutoDock) for both water-soluble and membrane proteins. Correlation of predicted affinity and EC(50) for 6 frequently used general anesthetics was only observed in GLIC, a member of a protein family relevant to anesthetic mechanism.

Dataset Information

Correcting the impact of docking pose generation error on binding affinity prediction.

Publications

Correcting the impact of docking pose generation error on binding affinity prediction.

Similar Datasets

OmicsDI is part of the ELIXIR infrastructure

Tweets