Litcius/Paper detail

Epik: p <i>K</i> <sub>a</sub> and Protonation State Prediction through Machine Learning

Ryne C. Johnston, Kun Yao, Zachary Kaplan, Monica Chelliah, Karl Leswing, Sean Seekins, Shawn Watts, David R. Calkins, Jackson Chief Elk, Steven V. Jerome, Matthew P. Repasky, John C. Shelley

2023Journal of Chemical Theory and Computation320 citationsDOI

Abstract

Epik version 7 is a software program that uses machine learning for predicting the p K a values and protonation state distribution of complex, druglike molecules. Using an ensemble of atomic graph convolutional neural networks (GCNNs) trained on over 42,000 p K a values across broad chemical space from both experimental and computed origins, the model predicts p K a values with 0.42 and 0.72 p K a unit median absolute and root mean square errors, respectively, across seven test sets. Epik version 7 also generates protonation states and recovers 95% of the most populated protonation states compared to previous versions. Requiring on average only 47 ms per ligand, Epik version 7 is rapid and accurate enough to evaluate protonation states for crucial molecules and prepare ultra-large libraries of compounds to explore vast regions of chemical space. The simplicity and time required for the training allow for the generation of highly accurate models customized to a program’s specific chemistry.

Topics & Concepts

ProtonationComputer scienceChemical spaceMoleculeConvolutional neural networkArtificial intelligenceComputational chemistryChemistryMachine learningDrug discoveryIonBiochemistryOrganic chemistryComputational Drug Discovery MethodsMachine Learning in Materials ScienceFree Radicals and Antioxidants
Epik: p <i>K</i> <sub>a</sub> and Protonation State Prediction through Machine Learning | Litcius