8/4/97

CONTACT: Janet Basu, News Service (650) 723-7582; e-mail basu@stanford.edu
Carol Moran, Interval Research Corp., (650) 842-6043 e-mail moran@interval.com

Stanford, Interval announce new "reverse spin-off" from industry: a research center in computational genetics

Technology transfer usually goes the other way.

Stanford scientists and engineers are accustomed to thinking of ways that the basic research carried out in their labs can be used to spawn new products and industries. In fact, when James Gibbons stepped down as Dean of Engineering in 1996, he took on a newly created role as special assistant to the university's president, defining and coordinating Stanford's myriad relations with industrial partners.

Now, Gibbons has announced what he calls a "reverse scientific transfer": a major project developed at an industrial laboratory will form the basis of a new academic research center at Stanford.

The project started at Interval Research Corp. of Palo Alto, a unique industrial research company founded in 1992 by software pioneers David E. Liddle and Paul G. Allen.

Thanks to a $2 million grant of seed money from the Paul G. Allen Foundation, the project will become the Center for Computational Genetics and Biological Modeling at Stanford, a cross-disciplinary, inter-school center to be based in the Department of Biological Sciences under the direction of Professor Marc Feldman and Aviv Bergman, a senior research scientist.

Scientists from Stanford and elsewhere will join forces there, using advanced computer modeling techniques to tackle complex questions about populations and human genetics ­ for example, about the spread of disease, the way organisms develop, the mechanisms of evolution and the history of the human species.

Interval's goal is to bring what the company calls "early-stage technologies" to life. Most of its projects are expected to become new companies or new industries.

This project, a five-year research program headed by Bergman, is spinning in the other direction: back to the university where Bergman put down roots while he was earning his Ph.D. in Feldman's lab.

Feldman said that the center's research will take advantage of an explosion of new information in biology, in an attempt to open up new areas of scientific understanding. Experimental and computational biologists are gathering massive data bases ­ for example, about the locations of tens of thousands of genes in humans and other organisms. At the same time, high-powered computer models using advanced, multiple-parallel-processor computers now have the capability to analyze these large data bases, studying many different variables at once, he said.

To make fruitful use of those capabilities, unique collaborations must be formed among computer scientists, geneticists and other biologists. That kind of research ­ bridging scientific fields and overcoming strikingly different ways of approaching scientific problems ­ is precisely the sort of work that Feldman's lab has promoted for years. Feldman is a founder, and he and Liddle both are board members, of the Santa Fe Institute ­ a think tank whose stated mission is to set new directions for science by fostering research that cannot be accomplished without knitting together a variety of scientific approaches.

Among the Stanford faculty working with Feldman and Bergman to plan the research agenda of the Center for Computational Genetics and Biological Modeling are Matthew Scott and David Kingsley in the Department of Developmental Biology; Luca Cavalli-Sforza and Neil Risch in Genetics, Samuel Karlin in Mathematics; and Yoav Shoham and Nils Nilsson in Computer Science.

At Interval Research Corp., Bergman worked with an international group of scientific collaborators. They began with studies of the theoretical conditions in which learning is beneficial to an organism, and continued with work on questions such as how geographical variations in environmental conditions affects the spread and growth of a population, and how evolution may be regulated in either gradual or sudden "punctuated" bursts.

Some of this work eventually may shed light on new ways to design artificial intelligence systems, as envisioned when the project began at Interval. Though it did not appear to have a commercial application in the near term, the company continued to pursue the research because of what Interval President and CEO Liddle called "excellent and high quality intellectual results." Now that the program is being moved to Stanford, Liddle said, it means that the agenda and spirit of the research program will continue in an atmosphere of academic free inquiry.

Bergman said that Stanford represents the ideal "gene pool of expertise" for the center. "While our research interactions with Stanford and other academic sites have been crucial so far, our plan for the future requires a much closer relationship with a number of disciplines and labs," he said.

The Allen Foundation seed grant will underwrite graduate students and collaborative research with several labs, taking advantage of the new center's "very strong computational facility," he said. A series of seminars will be used to cross-fertilize ideas among scientists and engineers in many fields.

Among the new and continuing projects the center is expected to tackle:

• Studies of how disease organisms spread and develop resistance. "There are well-defined methods for predicting the trajectory of an epidemic," Feldman said, but in the past scientists have not had the computational tools or the research data needed to analyze several trajectories at once ­ for example, to include such factors as genetic variation in the disease organism, and how that might be reflected in variations in the host that carries the disease, or genetic variations in patterns of immune response and resistance among those eventually infected by the pathogen.
• Studies of the genetic history of animal populations. Bergman has been working with evidence collected on the Channel Islands off Santa Barbara by scientists from the University of California-Los Angeles and Oxford University in the United Kingdom. "We are analyzing how genetic variations among gray foxes are related to the known history of their biogeography," Bergman said.
• Studies of development and speciation. Feldman said that models for the formation of phenotypes from genotypes could be used to analyze the changes in genes, body type and behavior that differentiate two populations, so that they eventually become different species. "This question ­ how the genetic basis for speciation connects to the phenotype ­ has proved to be very elusive. It's a problem of interest to many people," he said.
• Studies of variation in human populations. Feldman's lab has already begun this work, in close collaboration with Cavalli-Sforza, the pioneering geneticist who has long studied the patterns of human migration. They are working to analyze facts from linguistics, archaeology and other fields as well as a vast amount of data on small differences in DNA among people within a population, and small variations between different populations. Their goal is to show what can be learned about human history by statistically combining this disparate information.

Bergman said, "We are fortunate to live in an era where vast strides are being taken in the biological understanding of molecular evolution as well as the methodologies for computation. This is a very exciting period: we have a lot of data screaming for analysis."

Said Feldman, "As we bring together all these disciplines to think about what is happening at the forefront of biology, this promises to be an extremely exciting and productive field over the next five to 10 years."

For information about Interval Research Corp. on the World Wide Web, go to http://www.interval.com.

For information about the Feldman Lab, go to http://www-evo.stanford.edu/.

-30-

By Janet Basu