Last year, a research project led by David Goldstein of Duke University found that variations in the IL28B gene had profound effects on how people with hepatitis C respond to treatment. More specifically, those with certain mutations were twice as likely to respond to prolonged drug therapy (Nature 461, 399–401, 2009). The link's impact on patient care and drug development served as a wake-up call for investigators to ramp up efforts to identify genetic variants associated with disease occurrence and treatment response.

Spurred by this homegrown insight, investigators at the Duke University School of Medicine in Durham, North Carolina decided to form Discovery Genomics. The new center—a collaboration between the Duke Clinical Research Institute (DCRI) and the university's Center for Human Genome Variation (CHGV)—launched in October. Although projects within the university are its main focus, Discovery Genomics also operates as a fee-for-service business for pharmaceutical companies interested in finding genetic variations associated with drug efficacy and toxicity.

The commercial packages include genome-wide association studies and whole-exome sequencing to search for common and rare genetic variants, respectively, among DNA samples collected in clinical trials. Discovery Genomics will interpret the data and prepare publications authored by faculty, with industry ties disclosed.

The center's pharmacogenomic studies for drug efficacy cost about $500 per sample. Those for toxicity, meanwhile, cost about $6,000 per sample for whole-exome sequencing and $15,000 per sample for genome-wide scans. According to Tom Kaminski, DCRI's manager of business strategy, the costs are competitive with wholly commercial entities offering similar services, such as Expression Analysis, also in Durham.

Discovery Genomics highlights its search for rare genetic variants as a pioneering approach. “Little by little, more people are [agreeing] that it's not common but rare variation that will be associated with many complex diseases,” says Kevin Shianna, director of CHGV's genotyping facility.

But according to David Altshuler, a geneticist at Massachusetts General Hospital (MGH) in Boston, the approach echoes the technology-driven direction of pharmacogenomics overall. “It's not so much a change in strategy as an evolution of technology,” says Altshuler. “Five years ago, there was no way to do the sequencing to study rare genetic variants.”

Several other academic institutions have created engines to facilitate the incorporation of pharmacogenomics into clinical research. Similar efforts include MGH's Center for Human Genetic Research and the Penn Genome Frontiers Institute at the University of Pennsylvania in Philadelphia.

The Duke center is among the first to convert its faculty expertise and technological know-how into a business model. According to Kaminski, Discovery Genomics' strength lies with its storehouse of 200 full-coverage sequences to which new sequences can be compared, enabling researchers to glean meaningful data from a small number of new samples.

Still, some skepticism remains about the current approach. Christopher Newton-Cheh, a cardiologist and geneticist at MGH, cautions about the high potential for oversimplified or incorrect interpretations of the data. “Sometimes a result of a genetic variant in one tissue could counteract the result in another tissue,” says Newton-Cheh. “[Pharmacogenomics] may open potential avenues... but it also has the potential to close [them].”