QSAR study of a series of peptidomimetic derivatives towards MERS-CoV inhibitors

ABSTRACT

In this study, we report the quantitative structure activity relationships (QSAR) investigation to determine the relationship between the anti-MERS-CoV activity and a set of chemical descriptors computed using ChemSketch, MarvinSketch and ChemOffice software. Herein, the principal components analysis (PCA), multiple linear regression (MLR) and multiple non-linear regression (MNLR) methods were used with the intention to obtain a reliable QSAR model with good predictive capacity. The original data set of 43 peptidomimetic compounds was randomly divided into training and test set of 35 and 8 compounds, respectively. The values obtained by MLR and MNLR for the determination coefficient are 0.777 and 0.813, respectively. The predictive ability of the MLR model was assessed by external validation using the eight compounds of the test set with predicted determination coefficients R2test of 0.655.