Dana-Farber researchers view genetic info via a portal

A cancer that strikes white blood cells and eventually bone marrow, multiple myeloma can be difficult to treat. Now, the Dana-Farber Cancer Institute in Boston is harnessing the dual power of business intelligence and Web 2.0-based scientific search tools to gather complex, scattered data to better treat patients and work toward a cure for this formidable disease.

Dana-Farber is a treatment, research and teaching facility affiliated with Harvard Medical School. Its physicians and researchers regularly slog through complex calculations to find connections between data gleaned from tumor biopsies and other clinical samples and the vast genetic research housed within the organization or spread among three massive public-domain databases.

Dana-Farber officials are working to leverage grant money and other resources to blend data warehousing capabilities with Web-based data-collection tools, since vital connections between patient samples and analytical data will almost certainly prove the crux of both effective patient treatment and any potential breakthroughs tied to the disease, according to researchers.

The immediate goal is to clear tedious tactical challenges that get in the way of data sharing. "We face a real problem of data fragmentation or data balkanization. Often, the data we need is divided into lots of different subsets, since there are so many groups active in this research," says John Quackenbush, a professor of computational biology and bioinformatics at Dana-Farber's Department of Biostatistics.

Not only is data on multiple myeloma and other diseases often far-flung and fiercely guarded, it is also incredibly complex, says Joseph White, a senior research scientist at Dana-Farber. "A single gene may be represented by several different name sequences," he explains. "To gather all of the information on any one particular gene, a researcher must look at many sources and different expressions for the gene."

To make the hunt for precious genetic information easier, Dana-Farber officials have stitched together a system that wraps in Oracle's Healthcare Transaction Base, a service-oriented architecture that supports the medical industry's HL7 standard for the electronic exchange of clinical data. The project is partly funded through a US$1 million grant from Oracle Corp.

To search for genetic data contained in the National Institutes of Health's repository and other databases, Dana-Farber researchers use a Web portal designed by InforSense Ltd. to generate the data and analyze workflow.

"What we have been doing is pulling summary data on samples, along with links to public domain data with the InforSense Web-based tools," Quackenbush says.

Tying publicly available data with information gathered in-house is no easy or inexpensive task, says Wayne Eckerson, director of research at The Data Warehousing Institute in Renton, Wash. "In cases such as these where the data is really valuable, organizations often want to integrate publicly available information into their data warehouse, but that is an expensive approach," he explains. The alternative, Eckerson adds, is to link to the public databases in real time via an application programming interface.

For more enterprise computing news, visit Computerworld. Story copyright Computerworld, Inc.