pKa Prediction using Quantum Mechanics and Machine Learning

The dissociation of a proton from a heteroatom has a significant influence on the charge distribution and interactions of a molecule, influencing many important molecular properties which include pharmacokinetic (PK) attributes such as solubility, tissue or cellular distribution and permeability. The ability to predict the likelihood 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.

In this webinar, presented by Peter Hunt, Tamsin Mansley and Nick Foster, we describe a new method for the prediction of the acid dissociation constant (pK_a) of a heteroatom that combines quantum-mechanical and machine learning methods to generate an accurate quantitative structure-activity relationship (QSAR) model. The model achieves an R² > 0.9 and an RMSE < 1 log unit on an external test set containing both mono- and multi-protic compounds.

_{To hear the narration, please increase your speaker volume.}