Machine Learning on DNA-Encoded Libraries: A New Paradigm for Hit Finding.

Autor:	McCloskey K; Google Research Applied Science, Mountain View, California 94043, United States., Sigel EA; X-Chem, Waltham, Massachusetts 02453, United States., Kearnes S; Google Research Applied Science, Mountain View, California 94043, United States., Xue L; X-Chem, Waltham, Massachusetts 02453, United States., Tian X; X-Chem, Waltham, Massachusetts 02453, United States., Moccia D; X-Chem, Waltham, Massachusetts 02453, United States.; Cognitive Dataworks, Amesbury, Massachusetts 01913, United States., Gikunju D; X-Chem, Waltham, Massachusetts 02453, United States., Bazzaz S; X-Chem, Waltham, Massachusetts 02453, United States., Chan B; X-Chem, Waltham, Massachusetts 02453, United States., Clark MA; X-Chem, Waltham, Massachusetts 02453, United States., Cuozzo JW; X-Chem, Waltham, Massachusetts 02453, United States., Guié MA; X-Chem, Waltham, Massachusetts 02453, United States., Guilinger JP; X-Chem, Waltham, Massachusetts 02453, United States., Huguet C; X-Chem, Waltham, Massachusetts 02453, United States., Hupp CD; X-Chem, Waltham, Massachusetts 02453, United States., Keefe AD; X-Chem, Waltham, Massachusetts 02453, United States., Mulhern CJ; X-Chem, Waltham, Massachusetts 02453, United States., Zhang Y; X-Chem, Waltham, Massachusetts 02453, United States., Riley P; Google Research Applied Science, Mountain View, California 94043, United States.
Jazyk:	angličtina
Zdroj:	Journal of medicinal chemistry [J Med Chem] 2020 Aug 27; Vol. 63 (16), pp. 8857-8866. Date of Electronic Publication: 2020 Jun 11.
DOI:	10.1021/acs.jmedchem.0c00452
Abstrakt:	DNA-encoded small molecule libraries (DELs) have enabled discovery of novel inhibitors for many distinct protein targets of therapeutic value. We demonstrate a new approach applying machine learning to DEL selection data by identifying active molecules from large libraries of commercial and easily synthesizable compounds. We train models using only DEL selection data and apply automated or automatable filters to the predictions. We perform a large prospective study (∼2000 compounds) across three diverse protein targets: sEH (a hydrolase), ERα (a nuclear receptor), and c-KIT (a kinase). The approach is effective, with an overall hit rate of ∼30% at 30 μM and discovery of potent compounds (IC 50 < 10 nM) for every target. The system makes useful predictions even for molecules dissimilar to the original DEL, and the compounds identified are diverse, predominantly drug-like, and different from known ligands. This work demonstrates a powerful new approach to hit-finding.
Databáze:	MEDLINE
Externí odkaz:	Zobrazit plný text záznamu