Prediction of inhibition of the sodium ion - Proton antiporter by benzoylguanidine derivatives from molecular structure

The use of quantitative structure-activity relationships to predict IC50 va lues of 113 potential Na+/H+ antiporter inhibitors is reported. Multiple li near regression and computational neural networks (CNNs) are used to develo p models using a set of information-rich descriptors. The descriptors encod e information about topology, geometry, electronics, and combination hybrid s. A five-descriptor CNN model with root-mean-square (rms) errors of 0.278 log units for the training set and 0.377 log units for the prediction set w as developed. Examination of data set subclasses showed that systematic str uctural variations were also well-encoded resulting in 100% accuracy of pre diction trends. An experiment involving a committee of five CNNs was also p erformed to examine the effect of network output averaging. This showed imp roved results decreasing the training and cross-validation set rms error to 0.228 log units and the prediction set rms error to 0.296 log units.