Computer modeling of Biochemical Reaction Rates

Period of Performance: 06/15/2008 - 06/14/2009

$100K

Phase 1 SBIR

Recipient Firm

Stewart Computational Chemistry
Colorado Springs, CO 80921
Principal Investigator

Abstract

DESCRIPTION (provided by applicant): An accurate and rapid computational method for predicting reaction rates for biochemical processes would be developed as a tool for biochemistry research. Existing semiempirical methods, while rapid, are not sufficiently accurate. The recently completed PM6 methods, while rapid, are not sufficiently accurate. The recently completed PM6 method has increased the accuracy of prediction of heats of formation of stable ground-state systems, but is still not accurate for predicting transition state energies. The purposed work would involve developing a method specifically designed for predicting activation barriers. This would involve the construction of a database of activation barriers, and using that database to optimize parameters to reproduce the activation barriers. Existing semiempirical methods use a consistent set of parameters in predicting geometries and energies. In the new approach, geometries would be predicted using PM6, and energies would be predicted using the proposed method. Although this two-step procedure would be less elegant than conventional methods if would allow activation barriers to be predicted with much better accuracy. Construction of the database of activation barriers would involve using results over very high-level ab-initio calculations. Other groups have already done calculations of this type, so this step would involve only literature research. Development of the new method involves parameter optimization only. The issues involved are now well understood, and available optimization programs should be sufficient. Finally, the new method would be integrated into the existing commercial program MOPAC2007.