Grant: Discovery Grants Program - Individual

The proposed research draws on four inter-related areas: Chemistry: Reaction mechanisms, both in gas phase and in solution, are studied using high levels of theory. A long-term goal is to design and develop simple models to help chemists to a priori predict and explain reactivity and selectivity. Algorithms: The long-term aim is to develop highly optimized algorithms for all aspects of computational chemistry. Special consideration is given to algorithms that can take advantage of modern computer architectures, e.g., parallel processing. Theory: Novel wavefunction theories are being developed which can account for electron correlation and that lend themselves well to Divide-and-Conquer methods and to parallel processing. In addition, a Simulated Electronic Structure Theory, a completely novel approach to performing empirical calculations, is being developed. With this theory it is possible to simulate ab initio calculations and unlike conventional molecular mechanics, correctly describe bond dissociation and may ultimately be able to describe the entire potential energy surface. These are ambitious undertakings that have the potential to revolutionize computational chemistry. Chemical Informatics: Enormous amounts of chemical information are generated by computational methods, e.g., structures, heats of reaction, including complete reaction pathways. We are developing metrics that can easily and efficiently be used for chemical information retrieval, sub-structure searches, similarity measures, and for complex queries. A web-based National Comprehensive Computational Chemistry DataBase will be created, along with the tools for chemical information retrieval and storage. Such a resource will be of tremendous benefit to researchers in academia and industry, especially the pharmaceutical industry.