Making the Data Computable

The National Institutes of Health has announced funding of $18.8 million to develop a new National Center for Biomedical Ontology (cBiO), part of NIH's Roadmap Initiative. The principal investigators of the new center are Mark Musen of the Stanford University Medical Center and Suzanna Lewis, who is moving from the University of California Berkeley to join Berkeley Lab's Life Sciences Division. Associate director Sima Misra is also moving from UCB to Berkeley Lab.

Ontology, applied to computation, means much what it did to Aristotle, says Lewis — rules of logic and definition allowing knowledge from many sources to be integrated in a way computers that can use to address specific questions.

"Stanford is responsible for the theoretical computer-science aspects of cBiO," Lewis says. "Our core will be bioinformatics — developing practical applications researchers can use." Both these cores are driven by the biological projects of the third core: researchers working to connect data from model organisms to their human correspondents. The final focus is on training and educating researchers in developing ontologies. "We'll be developing methods and tools, not the content," Lewis says. "Content is determined by the needs of the research community."  

While working with the Berkeley Drosophila Genome Project and its rich Flybase dataset, Lewis explored ways to relate genomic information from a variety of model organisms, including mouse, fly, worm, zebrafish, human, and a host of others — all developed independently. "It was apparent that all of us would get more bang for the buck if we could coordinate our efforts."

The Gene Ontology Project, started even before there were complete genomes for some organisms, has become what Lewis calls "the de facto standard for describing the functional role of gene products." Other ontology projects followed, including Sequences (of DNA, RNA, and proteins); Cell Types, Pathways, and Anatomy, the last of which Lewis calls "a tough one, but it's important to be able to compare, say, mouse gut with human gut." Out of these efforts has grown the Open Biomedical Ontologies, an important cBiO resource.

In the Life Sciences Division, Lewis will be applying ontological concepts and methods to microarrray datasets, tumor cell lines, and many other disparate databases -- work that will keep her in contact with the larger community of biologists worldwide.  

"This is inherently a community effort," Lewis says. "A language that's spoken by just one person is not very good for communication."

-- Paul Preuss