NIH makes big deal of big data

Creates position of data scientist
By Anthony Brino
07:50 AM

The National Institutes of Health (NIH) is putting a fresh emphasis on health informatics, with Director Francis Collins, MD, creating a new advisory position and recruiting an associate director for Data Science.

Collins, a physician and geneticist, said there’s recently been an “exponential growth of biomedical research data” from genomics, imaging and electronic health records, with the new position being focused on building the NIH’s related research projects.

"There is an urgent need and increased opportunities for advanced collaboration and coordination of access to, and analysis of, the rapidly expanding collections of biomedical data," Collins said. "NIH aims to play a catalytic lead role in addressing these complex issues — not only internally, but also with stakeholders in the research community, other government agencies, and private organizations involved in scientific data generation, management and analysis."

Not unlike the Office of the National Coordinator, the NIH is trying to foster collaboration in an ecosystem of numerous commercial and public projects, as a way to help grow certain subfields. In the early 2000s, the NIH helped create and fund national centers of biomedical computing. One of those, the University of Michigan’s center for integrative bioinformatics, recently spun-off into a part of an international public-private pharmaceutical and genomic research project, called tranSMART.

As providers and the broader healthcare industry focus on standardization and interoperability with electronic health record systems, the larger players in the pharmaceutical industry are recognizing a need for a shared research warehouse enterprise, said Brian Athey, chair of the University of Michigan’s bioinformatics department.

Last year, tranSMART was released into the public domain under open source GPLv3 license, offering a data repository with demographics, clinical observations, clinical trial outcomes, adverse events and biomarker data like gene expression and metabolism. "Pharma feels that they need all this, but that they shouldn’t develop 10 different platforms for 10 different pharma companies,” Athey said.

[See also: NIH establishes National Center for Advancing Translational Sciences.]

The NIH Common Fund research project recently spotlighted the work of the nine original biomedical computing centers, which included informatics research in imaging, oncology and other specialty areas, as several of the research centers evolved into other multi-organization projects, like Michigans.

Athey  says the organizational model for biomedical research has changed, to become much more multi-disciplinary across multiple institutions, public and private.

In the era of Big Data, amid the country’s medical, economic and policy challenges, the NIH’s new data scientist will be looking for valuable informatics research projects with the potential to spur innovation, as the agency tries to do the same in multiple areas of medicine on a lean budget. As modern technology heads toward the "1,000 genome" one main biomedical challenge will be finding ways to actually use it in the clinical setting, by providing unique risk profiles or a basis for customized therapy.

[See also: White House launches 'big data' initiative ]

Collins has appointed Eric Green, MD, to serve as the acting associate director for data science. A bacteriologist by training, Green has kept a feet in both medicine and molecular biology and he was one of several lead researchers on the Human Genome Project.