ProSelection: A Novel Algorithm to Select Proper Protein Structure Subsets for in Silico Target Identification and Drug Discovery Research

Abstract

Molecular docking is widely applied to computer-aided drug design and has become relatively mature in the recent decades. Application of docking in modeling varies from single lead compound optimization to large-scale virtual screening. The performance of molecular docking is highly dependent on the protein structures selected. It is especially challenging for large-scale target prediction research when multiple structures are available for a single target. Therefore, we have established ProSelection, a docking preferred-protein selection algorithm, in order to generate the proper structure subset(s). By the ProSelection algorithm, protein structures of “weak selectors” are filtered out whereas structures of “strong selectors” are kept. Specifically, the structure which has a good statistical performance of distinguishing active ligands from inactive ligands is defined as a strong selector. In this study, 249 protein structures of 14 autophagy-related targets are investigated. Surflex-dock was used as the docking engine to distinguish active and inactive compounds against these protein structures. Both t test and Mann–Whitney U test were used to distinguish the strong from the weak selectors based on the normality of the docking score distribution. The suggested docking score threshold for active ligands (SDA) was generated for each strong selector structure according to the receiver operating characteristic (ROC) curve. The performance of ProSelection was further validated by predicting the potential off-targets of 43 U.S. Federal Drug Administration approved small molecule antineoplastic drugs. Overall, ProSelection will accelerate the computational work in protein structure selection and could be a useful tool for molecular docking, target prediction, and protein–chemical database establishment research.

Selection of protein conformations for structure-based polypharmacology studies

“…Indeed, several computational approaches have been successfully applied to identify potential activities of molecules on other targets, e.g., on the basis of their structural similarity with already reported modulators [21,22]. At present, a variety of structure and ligand-based methods are available to analyze the wealth of information provided by public databases [23,24], and to guide the selection of the most suitable protein conformations for structure-based drug design [13,25].…”

“…A variety of protein-centered approaches might represent efficient tools to identify suitable protein conformations for polypharmacology studies [9,18,25,45]. In fact, the shape and properties of a binding site are mostly dependent on amino acid rearrangements, switches between active and inactive protein states, and the presence and chemical nature of the bound ligand.…”

“…Molecular docking can be considered as well a useful tool for the selection of the most suitable protein conformations for structure-based VSs. In this regard, a recent study of Wang et al reported how integration of molecular docking and statistical analysis could help in identifying the most suitable protein conformations for structure-based drug design [25]. In particular, the authors investigated 249 protein structures of 14 autophagy-related targets by using Surflex-dock [79].…”

ALADDIN: Docking Approach Augmented by Machine Learning for Protein Structure Selection Yields Superior Virtual Screening Performance

“…explored three knowledge‐based strategies to generate ensembles of protein structures yielding maximum virtual screening performance. More recently, an approach for the pre‐selection of protein structures for docking (“ProSelection”) was introduced, which identifies protein structures as “strong selectors” or “weak selectors” based on the distribution of docking scores among the inactive and active compounds …”

“…More recently, an approach for the pre-selection of protein structures for docking ("ProSelection") was introduced, which identifies protein structures as "strong selectors" or "weak selectors" based on the distribution of docking scores among the inactive and active compounds. [28] In this work, we develop and test a new docking strategy that integrates machine learning to select, individually for each compound of interest, from an ensemble of protein structures, the single most suitable protein structure for docking. We refer to this method as the integrated approach for mAchine Learning AnD DockINg (ALADDIN).…”