干旱区生态环境知识资源中心

A new approach called ’’dynamic’’ QSAR is introduced to enhance the exploration of active chemicals and relevant molecular descriptors. In contrast to conventional QSAR methods where chemical structure is described by a single, low energy conformer, ’’dynamic’’ QSAR simulates the multiplicity of 3-D molecular shapes that a molecule can assume in complex reaction environments. The core of the new methodology is the coupling of the 3DGEN algorithm which exhaustively generates conformers and a rule-based system to rapidly screen conformers for desired properties. Hypotheses regarding receptor shape and interaction mechanisms are conveniently incorporated into the screening algorithm. A full array of stereoelectronic parameters available to OASIS can be combined with conventional topological and physicochemical indices for all conformations and explored using a variety of mathematical and visualization techniques. The ’’dynamic’’ QSAR method is illustrated by modeling the acute toxicity of a series of unsaturated alcohols in fish.