Nmr Structure Refinement by Nonlinear Estimation

Period of Performance: 03/01/1998 - 02/29/2000


Phase 2 SBIR

Recipient Firm

Moldyn, Inc.
5720 Oberlin Drive
San Diego, CA 92121
Principal Investigator


The long-term objective of this SBIR project is to bring to market an innovative new software for protein three-dimensional (3D) structure determination and refinement from NMR data. This new code, named ASTER (Automated STructure Estimation and Refinement), realizes a fundamentally new paradigm for biomolecular 3D structure research that offers to provide not only significant speedups (up to three orders of magnitude) over conventional methodologies, but also a framework for the treatment of all known statistical model and measurement uncertainties (including uncertain proton resonance assignments). The fast and accurate determination of protein 3D structures has significant implications for drug design. The specific goal of the Phase II is to develop a commercial prototype of ASTER. This will be achieved in two stages. The first stage will result in complete refinement capabilities and involves: the full incorporation of all available chemical information (e.g., force field parameters); the development of more sophisticated memory management techniques to accommodate the large (but sparse) matrices required by the methodology: and, the development of interoperability with existing codes. The second stage completes the seamless determination/refinement functionality through the development and implementation of advanced techniques to expand the radius of convergence of the methodology.