Lean-Docking: Exploiting Ligands’ Predicted Docking Scores to Accelerate Molecular Docking

Abstract

In structure-based virtual screening (SBVS), a binding site on a protein structure is used to search for ligands with favorable nonbonded interactions. Because it is computationally difficult, docking is time-consuming and any docking user will eventually encounter a chemical library that is too big to dock. This problem might arise because there is not enough computing power or because preparing and storing so many three-dimensional (3D) ligands requires too much space. In this study, however, we show that quality regressors can be trained to predict docking scores from molecular fingerprints. Although typical docking has a screening rate of less than one ligand per second on one CPU core, our regressors can predict about 5800 docking scores per second. This approach allows us to focus docking on the portion of a database that is predicted to have docking scores below a user-chosen threshold. Herein, usage examples are shown, where only 25% of a ligand database is docked, without any significant virtual screening performance loss. We call this method “lean-docking”. To validate lean-docking, a massive docking campaign using several state-of-the-art docking software packages was undertaken on an unbiased data set, with only wet-lab tested active and inactive molecules. Although regressors allow the screening of a larger chemical space, even at a constant docking power, it is also clear that significant progress in the virtual screening power of docking scores is desirable.

Efficient Exploration of Chemical Space with Docking and Deep Learning

“…Hierarchical workflows are similarly unattractive, because they compromise scoring function accuracy for the method that screens all compounds in order to achieve acceptable throughput. Others have modified the docking cascade to improve efficiency on large datasets. , Accordingly, there is a need for docking workflows that can find, with minimal loss of accuracy, the best scoring ligands in multibillion molecule libraries that exceed our capacities to screen or perhaps even build them explicitly, through combinatorial enumeration.…”

Automated discovery of noncovalent inhibitors of SARS-CoV-2 main protease by consensus Deep Docking of 40 billion small molecules

“…In order to speed up and automate conventional docking (without significant loss of valuable information), we recently developed Deep Docking (DD) – an AI-driven platform that can work in conjunction with any docking program and provides economical yet reliable access to billion-sized chemical libraries for SBVS. 19 Recently, similar approaches to DD that utilize AI to predict docking outcomes have emerged, including but not limited to MolPAL (molecular pool-based active learning), 20 lean-docking, 21 and AutoQSAR/DeepChem models. 22 Importantly, the high-throughput nature of DD can be utilized for docking ultra-large libraries but also enables the simultaneous use of multiple docking programs, facilitating the deployment of stringent consensus protocols, as advocated by the best practices of computer-aided drug design (CADD).…”

V-Dock: Fast Generation of Novel Drug-like Molecules Using Machine-Learning-Based Docking Score and Molecular Optimization

“…To overcome this limitation, computational approaches that use ligand docking scores as approximations of experimental data have been suggested. Recently, virtual screening [ 13 , 14 , 15 , 16 , 17 ] algorithms combining ML algorithms with docking scores have been suggested. Gentile et al [ 13 ] suggested a deep-docking approach that uses a docking score prediction model to reduce the resources required for docking computations.…”

“…Using this model, it was possible to increase screening enrichment by approximating the docking results of 13 million molecules of ZINC15 without performing actual docking calculations. Similarly, Berenger et al suggested a lean-docking approach that used a linear support vector regressor trained with docking scores of 10,000 molecules [ 14 ]. These results showed that efficient massive virtual screening is possible using a docking score prediction model without deteriorating the screening power.…”