Ertl, Peter (2020) Identification of bioisosteric analogs by deep neural network. Journal of Chemical Information and Modeling, 60 (7). pp. 3369-3375. ISSN 1549-95961549-960X

Abstract

Bioisosteric design is a classical technique used in medicinal chemistry to improve potency, drug-like properties or the synthetic accessibility of a compound or to find similar potent compounds that exist in novel chemical space. Bioisosteric design involves replacing part of a molecule by another part that has similar properties. The replacements may be identified by applying medicinal chemistry knowledge, by mining chemical databases or by choosing analogs similar in molecular physicochemical properties. In this article a novel approach to identify bioisosteric analogs is described where the suggestions are made by a deep neural network trained on data collected from a large corpus of medicinal chemistry literature. Thanks to this the trained network is able to mimic the decision making of experienced medicinal chemists and identify standard as well as non-classical bioisosteric analogs even for structures outside the training set. Examples of the results are provided and application possibilities discussed.

Item Type:	Article
Date Deposited:	12 Aug 2020 00:45
Last Modified:	12 Aug 2020 00:45
URI:	https://oak.novartis.com/id/eprint/42524