03/21/95

CONTACT: Stanford University News Service (650) 723-2558

Promising new approach for modeling molecules

STANFORD -- Sebastian Doniach, professor of applied physics, is collaborating with researchers at Los Alamos National Laboratory to develop a promising new approach for modeling the structure and behavior of biologically important molecules.

The approach includes the atom-by-atoms details that are critical for defining the overall structure of a molecule, as well as the influences of the surrounding water that affect a molecule's biological function. It allows a range of detail in a desktop simulation that is beyond the reach of even the most powerful, supercomputer-based molecular modeling systems currently available, according to the scientists. But its greatest attraction is that it is being designed to run on a standard scientific workstation, not a multimillion-dollar supercomputer.

The way in which the scientists have managed this feat will be described this week at the American Physical Society meeting in San Jose. The technique was developed by Doniach and Niels Gr_nbech-Jensen from Los Alamos. Their modeling program achieves major performance gains by taking calculated shortcuts that enable it to span the vastly different time and spatial scales that separate the level of the individual atom from the unfolding of an entire protein chain.

Normally, tracking the atomic dance of thousands of individual atoms slows simulations down to a crawl. The two scientists were able to vastly speed up the process by determining which kinds of atomic motions can be ignored. They have tested the results of their simplified model against simulations that include all the atomistic details but require Los Alamos' massively parallel computers to run. Such machines cost millions of dollars apiece and are not available to most researchers.

The scientists are working toward a product that will be available to others, a computer tool that could be extremely valuable for designing new drugs and other biochemical products. So far, they have shown that it works for simple peptides, short protein molecules built out of a couple of hundred atoms. If additional study confirms that the approach also works for more complex proteins, those containing thousands of atoms, then it is likely to find widespread application in the drug and biotechnology industry, the scientists say.

-dfs-

950321Arc5309.html


This is an archived release.

This release is not available in any other form. Images mentioned in this release are not available online.
Stanford News Service has an extensive library of images, some of which may be available to you online. Direct your request by EMail to newslibrary@stanford.edu.