Large Scale Optimization Methods with a Focus on Chemistry Problems
Abstract
Our research under this grant has developed general large-scale global optimization techniques in the context of solving protein structure prediction problems. Over the course of the research period the thrust of the research was primarily in two areas. The first area provided advances in our techniques for handling very large, complex proteins of arbitrary secondary and tertiary structure. The second area focused on advances in the efficiency of the algorithm on large-scale computing platforms. These advancements enabled us to significantly improve our effort in the CASP community-wide competition for protein structure prediction. During the CASP5 competition in summer of 2002, we attempted blind predictions of 20 proteins with sizes and structures far more complex than any previously attempted by our group. The results of assessment for the category of proteins which contain "new folds" (i.e. proteins which were not known by existing databases) show that our group's predictions were in the top 10% of groups which predicted these targets. Much of this research is joint work with the bio-chemistry group headed by Dr. Teresa Head-Gordon at UC Berkeley, and the computer science group of Dr. Silvia Crivelli at Lawrence Berkeley Laboratory.
Document Details
- Document Type
- Technical Report
- Publication Date
- Nov 01, 2002
- Accession Number
- ADA418451
Entities
People
- Richard H. Byrd
- Robert B. Schnabel
Organizations
- University of Colorado Boulder