View the LLNL home Back to the S&TR home Subscribe to Our magazine Send us your comments Browse through our index




Privacy &
Legal Notice

April 2001

The Laboratory
in the News

Commentary by Bert Weinstein

A New Kind
of Biological

The World's Most Accurate Lathe

Leading the
Attack on Cancer

Electronic Memory Goes High Rise



Bert Weinstein
Acting Associate Director,
Biology and Biotechnology Research Program

Computer Modeling Advances Bioscience

BIOLOGICAL scientists are beginning to reap the benefits of integrating high-performance computing with laboratory research. Work in Lawrence Livermore's Biology and Biotechnology Research Program is bringing computer modeling into the same essential role in bioscience that it plays in physics and engineering. Every complex physics experiment uses computer models to help design the experiment, guide its construction, predict the outcome, and suggest modifications. Every large engineering project, such as building an automobile, bridge, or integrated circuit, uses computational modeling to explore alternatives, optimize designs, and recognize flaws.
Over the past century, bioscience has evolved from a qualitative, observational discipline to a quantitative, predictive science. Molecular biology and genetics are generating vast amounts of complex data, and the generation rate is accelerating. Given the amount of information, more and more bioscientists recognize the need for large-scale computational tools. At the least, researchers need computers to collect, store, organize, and display their data and increasingly are using computers to accurately model the complex processes they are studying.
Throughout its history, the Laboratory has pioneered the use of powerful computers to solve complex scientific problems. The best known example is the extraordinarily complex modeling of nuclear detonations. At first glance, modeling nuclear weapons and modeling biological processes would seem to have little, if anything, in common. In truth, they are surprisingly similar. In both, many complex processes and variables work together to produce an end result with only a few measurable quantities, which are often averages spanning the entire experiment. Diagnostics are few—many of the most interesting quantities cannot be observed directly or take place on extremely brief time scales. Individual experiments are expensive and/or time consuming, so only a few can be conducted—not nearly enough to explore all the alternatives.
Modeling has many values. It fills the gaps in sparse experimental data, gives researchers insights into intermediate processes that cannot be directly observed, and allows many more virtual computer experiments than actual physical experiments. Computer models also force scientists to make sense of all available data at once, not just in piecemeal fashion. By integrating disparate bits of data, models inevitably provide insight into a problem.

As described in the article entitled A New Kind of Biological Research, advanced computation permits bioscientists to really "see" inside biochemical processes in breathtaking detail and learn how reactions take place. Modeling occurs at several levels of resolution. It can simulate the behavior of a few hundred atoms—for example, a few base pairs of DNA—with full quantum mechanical resolution. Or it can reveal the workings of larger systems with more abstraction and less precision. For example, the three-dimensional folding of a long chain of amino acids into a working protein is modeled partly from its physical and chemical properties and partly by comparing the sequence with known "folds"—structural patterns that are components of other functional proteins.
As in the physical and engineering sciences, confidence in biochemical computational models is developed by constantly testing them against experimental data obtained by biologists and biochemists working closely with computational experts. As the models are found to accurately predict measurable quantities, they begin to be trusted and relied on to guide experiments and to raise questions that suggest productive new lines of research.
Today, Lawrence Livermore scientists are at the forefront of integrating computation and experiment in bioscience. The challenge is to improve modeling accuracy and extend biosimulations to higher levels of complexity—for example, to groups of proteins working together to repair damaged DNA and, beyond, to intracellular components, structures, and communication paths. These research dreams extend to someday modeling the workings of an entire cell. The quest will, indeed, be an exciting one.






































Back | S&TR Home | LLNL Home | Help | Phone Book | Comments
Site designed and maintained by Kitty Tinsley

Lawrence Livermore National Laboratory
Operated by the University of California for the U.S. Department of Energy

UCRL-52000-01-4 | May 15, 2001