Lewis, Richard, Isert, Clemens, Kromann, Jimmy, Stiefl, Nikolaus and Schneider, Gisbert (2023) Machine intelligence models for fast, quantum mechanics-based approximation of drug lipophilicity. ACS Omega. ISSN 2470-13432470-1343

Abstract

Lipophilicity, as measured by the partition coefficient between octanol and water (log P), is a key parameter in early drug discovery research. However, measuring log P experimentally is difficult for specific compounds and log P ranges. The resulting lack of reliable experimental data impedes development of accurate in silico models for such compounds. In discovery projects at Novartis focused on such compounds, a quantum mechanics (QM) based tool for log P estimation has emerged as a useful supplement to experimental measurements and as a preferred alternative to existing empirical models. However, this QM-based approach incurs substantial computational cost, limiting its applicability to smaller series and prohibiting quick, interactive ideation. This work explores a set of machine intelligence models (Random Forest, Lasso, XGBoost, Chemprop, and Chemprop3D) to learn calculated log P values on both a public and an in-house dataset to obtain a computationally affordable, QM-based estimation of drug lipophilicity. Chemprop emerges as the best-performing model with mean absolute errors of 0.44 and 0.33 log units for scaffold split test sets of the public and in-house dataset, respectively. Analysis of learning curves suggests that a further decrease in test set error can be achieved by increasing the training set size. We discuss advantages of using synthetic data and the impact of dataset splitting strategy and gain insights into model failure modes. Potential use cases for the presented models include pre-screening of large compound collections and prioritization of compounds for full QM calculations.

Item Type:	Article
Date Deposited:	19 Jan 2023 00:45
Last Modified:	19 Jan 2023 00:45
URI:	https://oak.novartis.com/id/eprint/47882