The role of structural, pharmacokinetic and energy properties in the high-throughput prediction of redox potentials for organic molecules with experimental calibration

Reduction/Oxidation (redox) potential of small organic compounds is a key property that drives innumerable chemical and biological electron transfer reactions. However, experimental measurement of redox potential is time-consuming and expensive, yielding few and small experimental measured datasets. Computational methods have previously been applied to create redox predictors applicable only to a specific dataset. In this work, we investigate the effectiveness of various descriptors, including structural and functional properties, molecular energies, and drug-like properties, to predict redox potential. We use Gaussian Process Regression (GPR) as a model, as it is suitable for fitting small datasets and has shown promise in predicting redox potential. We train and test our redox predictor on three organic molecule datasets. We demonstrate that a GPR-based redox predictor using a combination of molecular descriptors, DFT energies, and pharmacokinetic properties works well across the datasets. Finally, we test the trained model against an experimental dataset of quinones to show that the model makes predictions well correlated with experimental data.