The dissociation of a proton from a heteroatom has a significant influence on the charge distribution and interactions of a molecule. These influence many important molecular properties, including binding to target and off-target proteins, absorption, distribution, metabolism and excretion (ADME) and pharmacokinetic (PK) properties such as solubility, tissue or cellular distribution and permeability. Therefore, the ability to predict the propensity of a molecule to lose or gain a proton in water is crucial for the development of new chemical entities with desirable PK, ADME and binding properties. We describe a method for prediction of the acid dissociation constant (pK<sub>a</sub>) of a heteroatom that combines quantum-mechanical (QM) descriptors, calculated using the semi-empirical AM1 method, with machine learning to generate an accurate quantitative structure-activity relationship (QSAR) model. The QM descriptors capture the geometric and electronic properties of the environment of site of (de)protonation, in the context of the whole molecule. This provides greater discrimination between potentially acidic or basic sites and greater transferability than simple fragment descriptors. The resulting model achieves a coefficient of determination of greater than 0.9 and a root-mean-square error less than one log unit on an external test set containing both mono- and multi-protic compounds. We also present results and comparisons with other methods for published benchmarking sets.

Predicting PKA using a combination of quantum and machine learning methods

ACS Fall 2021

The official archive and repository for all ACS Meetings & Expositions abstracts and select posters or presentations.