Three-Dimensional Quantitative Structure−Activity Relationships from Tuned Molecular Quantum Similarity Measures: Prediction of the Corticosteroid-Binding Globulin Binding Affinity for a Steroid Family
David Robert, Lluı́s Amat, Ramon Carbó‐Dorca · Journal of Chemical Information and Computer Sciences · 1999
Predictive models based on tuned molecular quantum similarity measures and their application to obtain quantitative structure-activity relationships (QSAR) are described. In the present paper, the corticosteroid-binding globulin binding affinity of a 31 steroid family is studied by means of a multilinear regression using molecular descriptors derived from mixed steric-electrostatic quantum similarity matrixes as parameters, obtaining satisfactory predictions. A systematic procedure to treat outliers by using triple-density quantum similarity measures is also presented. This method depicts an alternative to the grid-based QSAR techniques, providing a consistent approach that avoids problematic result dependency on the grid parameters.