Friday July 29/10:00

MS58/Grande Ballroom

Global Optimization Methods for Molecular Conformation and Protein Folding

To predict protein native structures and understand how they fold from the sequences of their constituent amino acids is one of the most important and challenging computational sciences problems of the decade. It is fundamental for almost all theoretical studies of proteins and protein related life process, and also has many applications in the biotechnology industry such as the structure-based drug design. The computational approach to this problem requires solving a global minimization problem, which is computationally intractable in general, and is extremely difficult to solve even on today's most powerful parallel machines. In the past several years, research efforts have been made in the field of computational science and optimization to develop algorithms which can exploit the problem structures and hence provide more efficient solutions to the problem. Interesting results have been obtained from both algorithmic development and application. The speakers will present several promising approaches to the global minimization problem for protein folding, including the Monte-Carlo search, stochastic sampling, constrained optimization, and continuation-based global optimization. Computational experience and results are discussed with the methods applied to some molecular conformation and protein folding problems. Organizer: Zhijun Wu Cornell University