Novel In Silico Models for ADMET Using Rapid Estimates of Quantum Chemical Descri

DESCRIPTION (provided by applicant): New empirical descriptors to replace ab initio calculations will be deployed for incorporation into new and improved Ultra High Throughput (UHT, gt200,000 compounds/hour) in silico models for Absorption, Distribution, Metabolism, Excretion and Toxicity (ADMET) properties of molecules, with only their 2D structures as input. In Phase I, we developed partial atomic charges including s and p components, atomic and molecular reactivities, and related descriptors from a dataset of almost 700 molecules. In Phase II we will augment the dataset with additional elements and complete the development of the charge model. Improved descriptor selection algorithms will be developed to capture the maximum information content of the new descriptors in model building. New and improved models for ADMET properties will be built using in-house and publicly available data sources. The models and new descriptors will be embodied in our commercially available software already in use by organizations worldwide to provide rapid and accurate estimation of ADMET properties to help prioritizing drug candidates for synthesis and screening. PUBLIC HEALTH RELEVANCE: Quantum level chemical descriptors have been demonstrated to provide improved predictions of ADMET (Absorption, Distribution, Metabolism, Excretion, and Toxicity) properties in our Phase I effort. While traditional methods for calculating these descriptors from 3-dimensional molecular structures perform at a rate of only about one molecule per day, our Phase I approach enables calculation of these descriptors at the rate of hundreds of thousands of molecules per hour. The refined quantum level descriptors and predictive models to be developed under this Phase II effort are expected to model important charge and reactivity effects that are crucial to predicting such properties as metabolism and toxicity, including carcinogenicity.