Computer Scientist To Explore Intricacies Of Biological Evolution

The journey, Foster said, will help him learn nature's rules that he hopes will apply to his own research into genetic programming.

Although he is interested in using it to write computer code that can adapt to changing conditions, the major interest in genetic programming is as an inexpensive way to develop software. Companies could use genetic programming to evolve software to meet demands rather than hiring software engineers to develop the computer code.

"If I can learn the real rules of how nature works, maybe I can use them to write more robust programs that can cope with changes better," Foster said.

Foster's NIH senior fellowship, awarded through the National Institute of General Medical Sciences, is a rarity, said Laurie Tompkins, program director at Bethesda, Md. His is the only one awarded so far this fiscal year, which ends Sept. 30. The grant also is unusual because it spans two years, she added.

Although the NIH senior fellowship would allow him to pursue his sabbatical studies anywhere, Foster plans to pursue his studies in the lab of Holly Wichman, an associate professor of zoology whose lab is little more than a block away on the UI campus.

Wichman has won two NIH grants, including one through the National Institute of General Medical Sciences. Wichman said modern biology needs computer science just as much as computer science needs modern biology. She works with phiX 174, a bacteriophage that in 1977 became the first organism whose genome was decoded after months of labor. Now the university has a gene sequencer that could do the same job in hours and could analyze six genomes a day if necessary.

"There has been an explosion of data. One of the things molecular biologists need is better computer programs to analyze all this information. The idea is to get a computer scientist in the lab and get his ideas and hopefully get him to build better mousetraps," she said.

She uses the bacteriophage, a virus that infects bacteria, because its life cycle is so short. "We can go through 1,000 population doublings in 10 days," Wichman said, allowing her to trace the effects of higher or lower temperatures or changing hosts on the phage's success.

"We're trying to learn the rules of evolution on a short time scale," she said. There have been surprises. "The dogma is evolution is unpredictable but on a small scale it is not always unpredictable."

Microbes show a predictable pattern in developing resistance to drugs, for example, Wichman said. "You can see the same changes occurring independently in different patients. We're trying to understand the rules that govern when evolution is predictable."

Viruses can also evolve quickly and switch hosts, a well-known worry in the medical world. That's why HIV, hantavirus and flu viruses pose such a threat and a challenge for molecular biologists, she said. "These viruses are evolving on a time scale we can observe and hopefully control."

Foster said evolution is not the sole realm of biology. He's used evolutionary theory to design programs that can respond to changes and reprogram the hardware they're running on to adapt, much as living things can.

"That's the analogy and if I can understand the analogy better, then we will be able to milk it for all its worth," he said.

There's another parallel between molecular biology and computer science, Foster said. Most of an organism's DNA is silent, meaning it does not apparently contribute directly to development or behavior. Most computer programs designed with genetic programming reach a similar point once they grow complex enough, bulking up with code that has no apparent direct influence on their function.

Wichman's interest in transposable elements in DNA, popularly known as jumping genes, caught Foster's attention several years ago. They co-published a scientific paper together about transposable elements.

"I figure if you really understand something, you will be able to find ways to use it," Foster said. "What I'm looking at long term is to get a self-correcting computer chip." But understanding how things evolve is a worthy pursuit that reaches across several disciplines, including biology, math, business and computer science, Foster said, noting that his colleague John Dickinson was one of the original developers of genetic programming. "There is a real strength to this university in evolutionary studies," Foster added.