| It has long been the aim of the pharmaceutical industry to speed the process of new drug development, and many strategies, including Three-Dimensional Structure-Activity Relationships (3D QSAR), have been developed toward this goal. In general these techniques equate a description of the 3D chemical structure and properties of molecules with some measure of biological activity. Distance Atomic Physicochemical Parameter Energy Relationships (DAPPER) is a novel technique for the development of 3D QSAR models. Included are the complete description of the SLOGP measure of atomic physicochemical parameters and MeMAD, an indication of the validity of the technique.; DAPPER describes molecule properties in terms of atomic contributions to octanol water partition coefficient, molar refractivity and charge, and 3D structural information is encoded using the set of all intramolecular atom pair distances. A new approach to measuring atom pair distances in flexible molecules is also described. The property and distance data are evaluated on a grid of Gaussians in eight property dimensions to generate a list of QSAR descriptors. Biological activity data are treated as intervals to accommodate any errors in measurement and the molecular descriptors are fit to this data using a modified partial least squares algorithm.; The method is validated on several benchmark data sets such as dihydrofolate reductase inhibitors, and angiotensin converting enzyme inhibitors with positive results. Direct comparison with existing methods is infeasible due to the specific process of model development. A graphical representation of model features is also generated. |
