TITLE:

THE HIERARCHICAL COLLECTIVE MOTIONS METHOD FOR SIMULATING LARGE BIOLOGICAL MOLECULES (No. 69)

DATE:

Friday, March 11th, 2005

TIME:

3:30 PM

LOCATION:

GMCS 214

SPEAKER:

Gary Huber, Department of Bioengineering, University of California San Diego

ABSTRACT:

Several obstacles stand in the way of performing atomic-level computer simulations of biological macromolecules. First of all, the large number of atoms challenges both processor and memory. Second, large molecules are characterized by multiple time scales spanning many orders of magnitude. Unfortunately, most of the interesting phenomena take place over the course of the long time scales, while the time steps taken by the simulation algorithm are limited by the small time scales.

Most of the interesting motions of biomolecules involve a few slow, large-scale collective motions of many atoms. If these motions are characterized, then information on the behavior of all individual atoms is not needed. We are devising a method, called the Hierarchical Collective Motions (HCM) method, to extract this information. It consists of running short simulations on small, overlapping segments of the molecule, analyzing the resulting information, getting rid of small, uninteresting motions, and defining a new set of variables. Next, the segments are joined into fewer and larger segments, and the information from the previous simulations is used to run new simulations at the coarser level. The analysis, elimination, and renormalization procedure is repeated until only one segment, encompassing the whole molecule, remains. Preliminary results obtained from simple model systems will be presented.

HOST:

Karen May-Newman

DOWNLOAD: