Ultra-large library docking for discovering new chemotypes

Abstract

Despite intense interest in expanding chemical space, libraries of hundreds-of-millions to billions of diverse molecules have remained inaccessible. Here, we investigate structure-based docking of 170 million make-on-demand compounds from 130 well-characterized reactions. The resulting library is diverse, representing over 10.7 million scaffolds otherwise unavailable. The library was docked against AmpC β-lactamase and the D 4 dopamine receptor. From the top-ranking molecules, 44 and 549 were synthesized and tested, respectively. This revealed an unprecedented phenolate inhibitor of AmpC, which was optimized to 77 nM, the most potent non-covalent AmpC inhibitor known. Crystal structures of this and other new AmpC inhibitors confirmed the docking predictions. Against D 4 , hit rates fell monotonically with docking score, and a hit-rate vs. score curve predicted 453,000 D 4 ligands in the library. Of 81 new chemotypes discovered, 30 were sub-micromolar, including a 180 pM sub-type selective agonist.

Rapid Identification of Potential Inhibitors of SARS‐CoV‐2 Main Protease by Deep Docking of 1.3 Billion Compounds

“…Collectively, our results strongly support the use of docking the largest available compound library for identifying novel potent scaffolds or chemicals, as concluded by Lyu et al. [31]…”

“…Second, ROC AUC value for Glide SP used to dock 81 Mpro inhibitors and~4,000 decoys was 0.72, similarly to the more computationally expensive Glide XP protocol (Figure 1b), and 0.74 when active molecules were diluted in 1 million random compounds extracted from ZINC15 (Figure S1 in supplementary material). Thus, in light of recent studies advocating for extending virtual screening to large chemical libraries when docking works well at smaller scales, [31] we decided to use Glide SP as DD docking program to screen ZINC15 against SARS-CoV-2 Mpro. DD relies on a deep neural network trained with docking scores of small random samples of molecules extracted from a large database to predict the scores of remaining molecules and, therefore, discard low scoring molecules without investing time and resources to dock them.…”

“…[6,[28][29][30] The main reason for that is that conventional docking is too computationally expensive and slow, while the libraries of available chemicals are growing exponentially. [31] To address this general challenge, we have recently developed a novel deep learning-based approach for accelerated screening of large chemical libraries, consisting of billions of entities. This Deep Docking (DD) platform utilizes quantitative structure-activity relationship (QSAR) models trained on docking scores of database subsets to approximate in an iterative manner the docking outcome of the remaining entries.…”

The Symbiotic Relationship Between Drug Discovery and Organic Chemistry

“…[241] Recently,t his ultralarge virtuall ibrary was used to discover novel, highly potent AmpC b-lactamase inhibitors and D 4 dopamine receptor ligands. [251]…”

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

“…[5] The largest docking study reported to date includes the virtual screening of a total of 170 million make-on-demand compounds against AmpC β-lactamase and the D 4 dopamine receptor, as a result of which several novel and, in part, highly potent inhibitors of these proteins were identified. [6] Despite these successes, the ability of scoring functions to estimate in particular absolute ligand binding affinities remains clearly limited, [7,8] which is related to the inadequate consideration of protein flexibility, [9,10] solvation effects, [10,11] and entropy. [12] The computational costs involved in sampling the relevant conformational states of biomacromolecules are often prohibitive to the consideration of protein flexibility and solvation in docking, in particular in the context of virtual screening.…”