Title: Developing target-specific scoring using black-box optimisation

Authors: Elham Shamsara; Jamal Shamsara

Addresses: School of Mathematical Sciences, Department of Applied Mathematics, Ferdowsi University of Mashhad (FUM), Mashhad, 91775-1365, Iran ' Pharmaceutical Research Center, School of Pharmacy, Mashhad University of Medical Sciences, Mashhad, 91775-1365, Iran

Abstract: In this study, the screening power of a AutoDock Vina scoring function was considered as an optimisation problem. It was hypothesised that the screening power of the AutoDock Vina scoring function can be optimised by a black-box optimiser. The weights of the energy terms of the AutoDock Vina scoring function were considered as input parameters for the black-box optimiser. This was implemented in Python. The study was designed to develop target-specific weights for six protein targets using active/decoy datasets retrieved from a database of useful (docking) decoys (DUD-E). The results demonstrated some improvements in the area under the curve (AUC) of the ROC curve and in the enrichment factor of both the training and test sets.

Keywords: AutoDock Vina; black box optimisation; molecular docking; DUD-E; machine learning; Python; virtual screening; target-specific scoring; protein targets; natural evolution strategy; NES.

DOI: 10.1504/IJCBDD.2017.082806

International Journal of Computational Biology and Drug Design, 2017 Vol.10 No.1, pp.12 - 23

Received: 15 Mar 2016
Accepted: 16 Aug 2016

Published online: 07 Mar 2017 *

Full-text access for editors Access for subscribers Purchase this article Comment on this article