Rescentris and Ohio State University have been awarded a three-year, $6 million grant from the state of Ohio’s tobacco settlement coffers for a collaborative bioinformatics project, Rescentris said last week.
The award will support development of the Biomedical Informatics Synthesis Platform, “a large IT framework that allows clinical annotations to be layered over the map of the genome,” according to Rescentris CEO Adel Mikhail.
Ohio’s share of the 1998 settlement with the US tobacco industry is expected to total $9 billion between 2000 and 2025. The state identified biomedical research as one of its funding priorities, and anticipates spending $24 million through its Biomedical Research and Technology Transfer trust fund in fiscal year 2003. BISP is expected to enable the study of tobacco-related disease by helping researchers identify genetic risk factors associated with the onset of disease, as well as factors that influence response to treatment and patient outcomes.
The goals of the infrastructure effort — as spelled out by the BISP project abstract on the OSU website — are ambitious. Clinical records, genomic data, information from radiological imaging systems, and digital microscopy data will be integrated via a system of ontologies and controlled vocabularies. The platform’s client/server architecture will use web and grid service middleware to provide distributed access to the federated database. The concept of a “virtual data center” will permit users to create specialized views of the federated database to support particular studies. A computational front end will give researchers access to tools such as Blast, PCR primer selection tools, and multiple alignment algorithms.
Joel Saltz, a professor in the biomedical informatics department at OSU, is leading the university’s contribution to the project, but was out of the country last week and unavailable to comment. Several researchers from the Ohio Supercomputer Center are also working on aspects of BISP’s grid infrastructure.
Mikhail said the participants are just beginning to put the project plans and requirement documents together. In addition to the current contributors, Mikhail said, “We are looking for large corporate partners to help commercialize the technology that comes out of this grant.” At first, data will come from clinical studies at OSU, but Mikhail said that other hospitals and clinical research organizations are expected to participate as the project gets rolling.
Rescentris’ BSML (Bioinformatics Sequence Markup Language) will serve as the data model for the genomics end of the platform. Mikhail said the standard might have to be “modified or extended” in order to map clinical information to human genome sequence data.
Components of the BISP infrastructure already support a number of research studies targeting tobacco-related diseases at OSU. Wolfgang Sadee, chair of pharmacology at OSU’s College of Medicine and Public Health and director of the university’s pharmacogenomics program, said he is using BISP’s data warehouse backend for a clinical research project on coronary artery disease.
The project, in which researchers are exploring genetic predisposition to coronary artery disease and genetic determinants of response to statins and other therapeutics, has deposited DNA samples and clinical records for 1,000 patients into the data warehouse so far — a move that is “already beginning to pay off,” according to Sadee. “We’re tapping into [the data warehouse], we’re making calculations, we’re getting risk factors out, we’re getting a much better idea as to how these patient populations evolve and what it all means,” he said.
While all the patients in the study population have been genotyped, Sadee said there’s plenty of room for improvement on the genomics aspects of the infrastructure. Only a few genes have been genotyped for each patient, he said, but “we now know that there are literally hundreds of candidate genes, and each patient may have a different mix of different genes and each one has a little bit of an effect, which makes it very difficult to resolve what’s important.” Ideally, Sadee said, the project will be able to genotype hundreds, if not thousands of genetic variants for each patient.
Some of the funding for the project has been earmarked to increase OSU’s genotyping throughput, Sadee said. “Now that we have the informatics capacity, we need to have the genomics capacity to bring the data in there.”
For the few SNPs that the researchers have so far, mapping between the clinical and genomic information “is working very well,” Saltz said. However, he was quick to note, “There is no pat solution yet. If you look at hundreds of genetic variants in a complex clinical setting, to actually come up with a solution to say, ‘This patient has to be treated this way’ — that is not possible yet, but that’s our goal.”
BISP, Sadee said, will help provide “the first steps in that direction.”