Comparative Occupancy Analysis (CoOAn) - A Straightforward and Directly Applicable 3D-QSAR Formalism to Extract Molecular Features Obligatory for Designing Potent Leads.

Abstract

A simple and directly applicable 3D-QSAR method, termed Comparative Occupancy Analysis (CoOAn), has been developed. The method is based on the comparison of local occupancies of fragments of an aligned set of molecules in a 3D-grid space. The formalism commendably extracts the crucial position-specific molecular features and correlates them quantitatively to their biological endpoints. The method has been effectively applied and efficaciously validated on three large and diverse datasetsthrombin, glycogen phosphorylase b (GPB), and thermolysin inhibitors. Several robust and statistically significant predictive 3D-QSAR models were developed while simultaneously considering the influence of grid spacing on the accuracy of the results. The models, generated by the G/PLS chemometric method, not only unswervingly identified the obligatory chemical features but advantageously detected those that are unfavourable or detrimental for the molecular activity. The CoOAn models can profitably be used to optimize existing molecules as well as to design new leads with more desirable (and/or less detrimental) features. The activity-modulating features (together with their distance-constraints) extracted by the methodology can also be incorporated into a pharmacophore-type query to search a chemical database for novel leads.