Research

Research

Organizing and Sharing Biological Data

Dr. James Brinkley's field didn't have a name, so he invented one. In 1991, Brinkley coined the term "structural informatics."

His field integrates computer technology with biomedical information. The UW has received a grant of $1.8 million from the National Library of Medicine, a branch of the National Institutes of Health, to plan a Center for Excellence in Biomedical Computing.

"The NIH has become aware of a huge need for information processing in the basic sciences," said Brinkley, a UW research professor of biological structure and of medical education and biomedical informatics. He also is an adjunct professor of computer science and engineering.


	An image from the Digital Anatomist Dynamic Scene Generator shows the thoracic vertebral column, the aorta, and its branches.

Brinkley is principal Investigator on this planning grant. Dr. Peter Tarczy-Hornoch, an associate professor of pediatrics and head of the Division of Biomedical Informatics in the Department of Medical Education and Biomedical Informatics, is co-principal investigator.

One of the first steps in creating the center will be determining organizational aspects, such as leadership.

"Another goal is to talk to various biologists and find out what their information needs are, to assess what information problems they might have, and then try to understand how we might help them with some of our techniques." Brinkley said.

According to Brinkley, the center will to combine efforts from three groups: the Division of Biomedical Informatics, the database group in the Department of Computer Science, and the Structural Informatics Group in the Department of Biological Structure. Efforts by the Structural Informatics Group, which was originally founded as the Digital Anatomist Project by Dr. Cornelius Rosse, professor and former chair of biological structure. The anatomy project includes the Human Brain Project, which is preparing a 3-D, computer-based template that could contain overlays of various kinds of information about brain function.

The Structural Informatics Group is also devising methods for understanding and managing information about other structures of the body and is using those methods as a framework for organizing data.

"The new center will extend the methods that we and our collaborators have been developing in particular areas to a wider range of biological applications," Brinkley said. "The structure of the body should be a helpful way to organize a lot of useful biological information."

The planning grant funds three project. These include applying techniques that Brinkely, along with others, have already created for brain mapping to the lab of Dr. John J. Clark, professor of biological structure, who studies cataract formation.

"We building an experiment management system to help him manage and share his data," Brinkley said.

Another issue the group is working on in Clark's lab is predicting protein function. Members of the group are using computer science techniques originated by several scientists, including Linda Shapiro, professor of computer science and engineering and of electrical engineering. These techniques were initially applied to computer vision to recognize objects and scenes. The goal is to see if Shapiro's techniques can reveal a protein's shape function.

"This is an example of taking techniques that were developed for a completely different area, computer science, and applying them to another area, biology," Brinkley said.

Dr. John Gennari in Biomedical Informatics leads a second project, called Ontology Alignment. This project intermeshes two different knowledge sources.

"There are ways to represent knowledge in the computer," Brinkley said, "but we would like to be able to link our way of representing knowledge with somebody else's way of representing knowledge."

"The third project, Peer Data Management, which is led by Tarczy-Hornoch and Alon Halevy in Computer Science, recognizes that there will never be a centralized database for all the information in biomedicine, so we need to find a way to link multiple databases to enable them to talk to each other," Brinkley said. He explained that this idea is analogous to Napster and other peer-to-peer sharing programs.

"We're hoping to generalize that idea beyond the notion of shared music files," Brinkley said. "We're looking for ways to share the complicated data and knowledge that's recorded in biomedicine. The data and knowledge could be stored in local experiment management systems like the one we're creating for John Clark, but then we'd let it be searched in a peer-to-peer manner similar to the way you would search for a music file."

"This research will prove challenging," Brinkley said, "and is therefore of great interest to our collaborators in computer science."

"I don't know if we'll ever see a completely seamless information resource," he added, "because knowledge is changing so fast"