The National Institutes of Health has enlisted the University of Pittsburgh School of Medicine into an initiative to build a virtual immune system, a computer model of how the body responds to infectious agents.

Pitt has been awarded a five-year, $9.1 million contract and is one of four universities selected as Immune Modeling Centers.

By revealing in detail how the body, lungs and individual cells respond to pathogens such as influenza, tuberculosis and smallpox, mathematical representations of the immune system might suggest improved treatments or speed development of vaccines, said Timothy Gondre-Lewis, one of two project officers with the National Institute of Allergy and Infectious Disease.

The recent emergence of the virulent H5N1 strain of avian flu, for instance, has raised the specter of pandemic flu if a mutation of it should make the jump from bird transmission to human-to-human transmission. A tool such as a computer model of the immune system might help predict how humans would react to this strain.

"I think, in the end, this type of modeling will be able to tell us what we should think about in terms of immunotherapeutics or other interventions," should the H5N1 virus someday become transmissable from human to human, Dr. Gondre-Lewis said. "The investigators are well aware of this problem. It's going to be on their mind."

The modeling center at Pitt will focus on the lung, an organ that often is the site of exposure to infectious agents, and look at how it responds to pathogens such as the flu, tuberculosis and tularemia, a potential biological weapon also known as rabbit fever.

The Pitt center also will draw on existing collaborations with Carnegie Mellon University and the University of Michigan.

Researchers at two other centers, Duke University and the University of Rochester, will focus on the immune response at the cellular level, while researchers at the fourth modeling center, the Mount Sinai School of Medicine, will look at the immune response of the body as a whole.

Computer modeling is a tool of growing importance in many areas of science, including biology and medicine, because it enables researchers to closely analyze actions and reactions that occur at size and time scales that are either too small or too great to be easily observed. The results of physical experiments are translated into numbers to create the virtual models, which, in turn, can suggest further physical experiments.

"There's been a small but growing community of people trying to model the immune system over the last 15 to 20 years," said Dr. Penelope A. Morel, an immunologist who leads the Pitt center. Modeling individual immune reactions, such as those that can be performed in a test tube, has progressed, but modeling the more intricate interactions that occur in a living creature is a much bigger hurdle.

It's an effort that requires collaboration between immunologists, mathematicians, computer scientists and others. "An initiative like this is one way to get the top scientists in the field to work together," said Dr. Gondre-Lewis, who oversees the initiative with Dr. Alison Deckhut Augustine.

"We're very excited that they've seen fit to try to push this forward," Dr. Morel said.