NEW YORK (GenomeWeb) – Joining a growing number of population-based genomic medicine projects, Yale School of Medicine and the Yale New Haven Health System are planning to launch a large-scale initiative next year that will recruit at least 100,000 patients and analyze their genomic data, along with electronic health record information, for both research and return of clinically actionable results.
Michael Murray, formerly director of clinical genomics at Geisinger Health System's Genomic Medicine Institute, is coordinating the launch of the project. Earlier this year, he left Geisinger to become director of clinical operations at the Center for Genomic Health at Yale School of Medicine and Yale New Haven Hospital. At Geisinger, he was intimately involved in the MyCode Community Health Initiative, one of the first large-scale genomic medicine projects that returned actionable results to participants.
The new initiative at Yale, which has yet to be named, will build on the experience Murray and his colleagues gained at Geisinger. "The project will be using the Geisinger model as sort of a template," he said, "but hopefully, we'll take the lessons learned there and build up a large biobank quickly, and with a couple of differences."
Having the successful model of the Geisinger MyCode initiative has been helpful because "this kind of a project is still an idea that a lot of people are not deeply familiar with," Murray said. "So to be able to point back to the work we did there, to be able to speak confidently about what the opportunities and deliverables are, has made it great to interact with people."
The Yale project could itself become a blueprint for similar efforts in the future. "We hope that we'll be able to build a model for people at academic health centers," he said.
The project is supported by Yale School of Medicine and the Yale New Haven Health System, which are putting up some of the initial funds, and the state of Connecticut, which is providing a grant to get the project off the ground. Combined, these institutions have committed more than $20 million in total, Murray said. Once the initiative has started, the plan is to look for other sources of funding, including research grants and collaborations with industry and academia, he added.
Starting in January of 2019, the project will enroll 400 to 500 patients per week, or 20,000 to 25,000 per year, to a biobank that will store peripheral blood samples. Participants will be recruited throughout Connecticut from Yale New Haven Health's hospital network, which includes Yale New Haven Hospital, Bridgeport Hospital, Greenwich Hospital, and others. They will be broadly consented for research, as well as for the return of genomic results that can be used to drive their healthcare.
The initial goal is to recruit 100,000 participants, but depending on interest and how the project goes, that number might be increased later on. "We think that at 100,000, there will be opportunities to collaborate with other big biobanks, as well as do a lot of projects on our own," Murray said.
While Connecticut has a relatively small population compared to other states — about 3.6 million — its ethnic and racial makeup pretty much matches that of the US. "We think that getting a diverse cohort across Connecticut will allow us to do research and other precision medicine initiatives that will be good pilots for the entire country," he said. This, he added, makes the project a little different from Geisinger's MyCode initiative, whose cohort is almost exclusively of European ancestry.
Another difference is that the sequencing will be done in house in a CLIA environment, at the Yale Center for Genome Analysis. The center is already involved in other large-scale projects, for example the National Human Genome Research Institute-funded Yale Center for Mendelian Genomics, and offers CLIA-compliant exome sequencing. "We have a team in place that is doing clinical genomics as well as big research projects all the time," Murray said. "They will be engaged as full partners to do the sequencing and to deliver those results into the electronic health record."
Like the Geisinger project, which outsourced the sequencing to its partner, Regeneron Pharmaceuticals at the Regeneron Genetics Center, the Yale initiative will sequence participants' exomes and generate array-based genotyping data, at least initially. "We've talked about [sequencing] whole genomes, and everybody aspires to that, but it's a cost issue," Murray said.
Allen Bale, who has been overseeing clinical genome sequencing services at the Yale Center for Genome Analysis, and his team will be in charge of making variant calls and generating medical reports for the project, which will go into patients' EHRs. All the pieces for doing so are basically in place, Murray said, "and what we're spending our time doing this year is getting the pipeline created to do this setup at a high volume."
The data will be used twofold: for research projects — both involving de-identified genomic and EHR data and identified data — and to inform patients of clinically actionable results.
One big difference to Geisinger, which is essentially a hospital system, and the Yale effort, which involves a hospital network as well as a large medical school and university, is that the research community is much larger.
"I’m really excited that there are people who will want to participate in research related to this biobank, who extend beyond the typical medical and basic science researchers," Murray said.
For example, social scientists at Yale have already expressed an interest in the project, and early conversations with other faculty are ongoing about research in other areas. "There are lots of questions about health economics, as well as ethical and legal implications" of returning and managing genomic results, Murray said, which could be of interest to scholars at Yale's school of public health, law school, or business school.
Regarding return of results, the project will focus on rare variants that increase the risk for monogenic diseases, polygenic scores involving common diseases, and pharmacogenomic variants. "We may do it in a phased manner, but those three things are the top priorities," Murray said.
In terms of rare variants, a list of genes and conditions has not been finalized yet, he said, but will be based on the American College of Medical Genetics and Genomics' ACMG-59 list to some extent.
The array data will be used to calculate polygenic risk scores, which could potentially guide care, in addition to their use in research studies. Last month, for example, researchers at Massachusetts General Hospital, Harvard Medical School, and the Broad Institute published a study on new genome-wide polygenic scores for five common diseases. "The whole field of polygenic scoring is very quickly evolving, so exactly what we will deliver back, and how we will deliver it back, will depend a little on how that goes both at Yale and elsewhere, but it's definitely a priority of the project," Murray explained.
Overall, the hope is that the project will provide participants with information that will help them prevent disease. "We think that this kind of a biobank and associated screening is really going to get people to be aware of some significant health risks, such as heart disease and cancer risk, and get care for it," Murray said.
In a paper published in the American Journal of Human Genetics last month, for example, the Geisinger team wrote that 542 participants out of a total of 200,000 participants in its project, or about 0.25 percent, received a report of at least one pathogenic or likely pathogenic variant that is associated with a monogenic disease through their EHR.
In addition, two years ago, they published a study in Science reporting that 229 individuals in the Geisinger study were found to have a known or predicted pathogenic variant related to familial hypercholesterolemia (FH), suggesting that the prevalence of the disease is 1:256 in the cohort. Only 15 percent of these individuals had a diagnosis of FH, and only 24 percent would have qualified for such a diagnosis based on their clinical EHR data.
Finally, unpublished data on the BRCA 1 and 2 genes that has been presented by Geisinger at conferences suggests that the project's screening approach identified five times more individuals with a BRCA mutation than current approaches.
"It will take a while to develop the long-term outcome data that says that this benefits people in very specific ways," Murray said, "but we can say in general that knowing that you're at risk for heart disease or cancer, and then getting appropriate screening and prevention, is expected to impact some of the most serious health risks in the US."
In the meantime, he and his colleagues are making sure the infrastructure for the new project at Yale gets built, which includes getting protocols finalized; hiring new staff, including five to 10 full-time recruiters; and getting sequencing and variant calling pipelines up and running. Also, a website describing the project will go up within the next couple of months.
"It's going very fast," Murray said. "I arrived in April and we hope to be online right after the first of the year. That's a pretty good timeline. We don't want to rush it too much but we also want to take advantage of the way the field is evolving and of the opportunities that we know are there."
Murray said he expects little overlap with recruitment for the National Institutes of Health's All of Us research program, which has similar goals in terms of genomic analysis and return of results. That project is enrolling patients at Community Health Center (CHC), a primary care provider in Middletown, Connecticut. "We've spoken to them and we're going to look for ways to collaborate," he said. "We don't anticipate that there will be any competitive situation there."