DYNAMIC QSAR - A NEW SEARCH FOR ACTIVE CONFORMATIONS AND SIGNIFICANT STEREOELECTRONIC INDEXES

A new approach called ''dynamic'' QSAR is introduced to enhance the ex ploration of active chemicals and relevant molecular descriptors. In c ontrast to conventional QSAR methods where chemical structure is descr ibed by a single, low energy conformer, ''dynamic'' QSAR simulates the multiplicity of 3-D molecular shapes that a molecule can assume in co mplex reaction environments. The core of the new methodology is the co upling of the 3DGEN algorithm which exhaustively generates conformers and a rule-based system to rapidly screen conformers for desired prope rties. Hypotheses regarding receptor shape and interaction mechanisms are conveniently incorporated into the screening algorithm. A full arr ay of stereoelectronic parameters available to OASIS can be combined w ith conventional topological and physicochemical indices for all confo rmations and explored using a variety of mathematical and visualizatio n techniques. The ''dynamic'' QSAR method is illustrated by modeling t he acute toxicity of a series of unsaturated alcohols in fish.