YouTube Facebook LinkedIn Google+ Twitter Xingrss  


By Jennifer Patterson Lorenzetti

November 15, 2003 | Drawing on the power of a cluster of microcomputers at Ernest Orlando Lawrence Berkeley National Laboratory in Berkeley, Calif., researchers have used computer modeling to point the way to new findings about the presence of an "adaptive control function" in the immune system. The findings not only shed new light on the functioning of the immune system, but also highlight the utility of computer modeling in biological research.

The in silico experiments, published in the September 25 issue of Science, were led by Arup Chakraborty, professor of chemical engineering and chemistry at the University of California at Berkeley and a leader in computer modeling for immunological research. His experiments used Berkeley Lab's Scientific Cluster Support program, which gives researchers access to an array of Linux-based microcomputers connected in parallel to perform as a single system. "From a computational standpoint, we didn't need a major breakthrough," Chakraborty says -- only an opportunity to harness existing computer power.

Chakraborty, together with immunologists Andrey Shaw (Washington University School of Medicine) and Michael Dustin (New York University School of Medicine), predicted the behavior of T cells under a variety of conditions, a process that would be impossibly cumbersome in traditional laboratory research. "We can test out hypotheses very quickly [and] follow every little detail," Chakraborty says. By looking at the predicted behavior of T cells as they are exposed to pathogens, the researchers were able to study the "immunological synapse" -- the gap between T cell and pathogen where the two cells interact.

The model predicted that this synapse is involved in both amplifying weak signals from the pathogen to the T cell, as well as damping overly strong signals, protecting the cell from death resulting from exposure to the strongest signals. The researchers described this function as an "adaptive controller," the presence of which was confirmed with studies using transgenic mice.

Chakraborty doubts that researchers would have thought to test for this adaptive controller mechanism without computer models pointing the way. "The computer model found the mechanism; the animals tested it," he says, adding that the in silico work "provided the insight." He believes that this sort of experiment points to the opportunities for closer work between traditional biological experimentation and computer modeling.

Although current computing power was sufficient to point to the adaptive controller mechanism, Chakraborty predicts that advances in computer technology will transform these computer models into more reliable facsimiles of biological subjects -- what he calls "digital transgenic animals."

"We need some advances in computer technology; we need to go beyond standard Linux boxes. We need the ability to simulate things in a much more efficient parallel way," he says, adding that greater processor speed will also enhance the use of computer modeling in biological research. Until these improvements come, however, Chakraborty is pleased with the use of these models. "[They are a] very useful complement to genetic and biological [research]."

 





For reprints and/or copyright permission, please contact  Jay Mulhern, (781) 972-1359, jmulhern@healthtech.com.