Machine Learning Boosted Docking (HASTEN): An Open-Source Tool To Accelerate Structurebased Virtual Screening Campaigns

Abstract

The software macHine leArning booSTed dockiNg (HASTEN) was developed to acceleratestructure-based virtual screening using machine learning models. It has been validated usingdatasets both from literature (12 datasets, each containing three million molecules dockedwith FRED) and in-house sources (one dataset of four million compounds docked withGlide). HASTEN showed reasonable performance by having the mean recall value of 0.78 ofthe top one percent scoring molecules after docking 10 % of the dataset for the literature data,whereas excellent recall value of 0.95 was achieved for the in-house data. The program can beused with any docking- and machine learning methodology, and is freely available fromhttps://github.com/TuomoKalliokoski/HASTEN.