Autocorrelation of Molecular Electrostatic Potential Surface Properties Combined with Partial Least Squares Analysis as New Strategy for the Prediction of the Activity of Human A<INF>3</INF> Adenosine Receptor Antagonists

The combination of molecular electrostatic potential (MEP) surface properties (autocorrelation vectors) with the conventional partial least squares (PLS) analysis has been used for the prediction of the human A<INF>3</INF> receptor antagonist activities. Three-hundred-fifty-eight structurally diverse human A<INF>3</INF> receptor antagonists have been utilized to generate a novel ligand-based three-dimensional structure−activity relationship. Remarkably, our chemical library includes all 21 important chemical classes of human A<INF>3</INF> antagonists currently discovered, and it represents the largest molecular collection used to generate a general human A<INF>3</INF> antagonist structure−activity relationship. A robust quantitative model has been obtained as described by both cross-validated correlation coefficient (r<INF>cv</INF> = 0.81) and prediction capability (r<INF>pred</INF> = 0.82). The proposed MEP/PLS approach can be considered as an alternative hit identification tool in virtual screening applications.