Field of Genes
The Penn State Hershey Institute for Personalized Medicine benefits from a unique “build it as you need it” growth pattern
I didn’t start my career intending to focus on personalized medicine. For many years I was a yeast geneticist, first at Cold Spring Harbor Laboratory, then at the State University of New York at Stony Brook, and finally at Princeton. I studied numerous topics – how cells respond to their environment, how their genes are regulated – but all within the purview of basic research in model organisms. That changed about four years ago, when the vice dean for research at Penn State Hershey asked if I wanted to come out and start an institute for personalized medicine. It was too good to pass up – a wonderful challenge and an opportunity to take everything I had learned about genomics in model organisms and apply it to people.
Starting from scratch
I had to build from the ground up when establishing the Institute for Personalized Medicine (IPM). In Pennsylvania, we have access to money from the “tobacco settlement.” The Tobacco Master Settlement Agreement provides the Milton S. Hershey Medical Center with a significant amount of funding each year that can be spent at the discretion of the vice dean for research. Fortunately, because he’s the one who suggested the program, he was very supportive of our financial needs. We also had a grant from the National Institutes of Health (NIH) to renovate the physical space we needed to house the institute – so having the initial resources available early on helped us get the IPM off the ground.
Early on, I recruited Glenn Gerhard. He’s a clinical geneticist with a strong research program in the application of genomics to medical issues, and he also had experience in setting up and running a biobank. So he knew all the ins and outs of obtaining patient consent, establishing databases to handle samples, working with the institutional review board for approvals, and generally understanding the ebb and flow of the biobanking process. He was a great addition to the program, and I credit him with much of our early success.
My job at that point was to recruit genetically interesting patient populations from the various clinical departments – those who looked like they might be predisposed to a particular disease or who could be stratified by response to a particular treatment. I used that knowledge first to identify the questions we wanted to ask, and then to select the patients whose data might help us answer those questions.
This is very different to the way most other biobanks and institutes for personalized medicine operate. Many use what I call a “field of dreams” approach – they say, “if we build this institute and get 100,000 samples, then somebody smart will come along and figure out what to do with it.” That requires a lot of upfront infrastructure and funding without any assurance of downstream income to offset it. Ours, by contrast, was a focused, “build it as you need it” approach, so that we could closely link funding to individual projects.
We currently have about 25 different collaborations, with another 15 or so in preparation. We started with amyotrophic lateral sclerosis (ALS, or Lou Gehrig’s disease) and have expanded outward – first into other neurological diseases, and now into cancer, cardiovascular, pulmonary, urinary and more. Ever since people realized we were a resource that would enable them to pursue their more ambitious research questions, I haven’t had any trouble finding collaborators.
ALS is a syndrome, not a disease – it’s diagnosed based on patient presentation, rather than underlying etiology. That may be why it’s so difficult to treat; trying to find a single drug to address a range of diseases that we classify as “ALS” would be like trying to find one drug to cure all of cancer. Our hope was that, by applying genomic tools, we might be able to identify the etiologies of the disease – and then come up with new treatment approaches.
As a geneticist, I recognized early on that the known ALS-associated genes were all dominant alleles. From genetics in model organisms, we know that if you select for mutations that generate a given phenotype, only 10 percent are dominant alleles – so we were just hitting the tip of the ALS gene iceberg. To overcome that, we focused heavily on examining the genomes of patients previously considered “sporadic” and their families, in order to see whether their disease was actually the result of recessive mutations. And we’ve been successful in that. In fact, we’ve increased the number of potential genes that are causative in ALS by a factor of 10.
Handling the hype
Despite our early successes, running a center like the IPM isn’t easy. Funding is always an issue, for instance – there are always more collaborations than we have money to pursue. There’s also the issue of translating basic science into clinical applications. Our work to date has been strictly research, but precision medicine’s real impact is on patient treatment – so we have to get the genomic information back to the clinic. That’s not happening as quickly as we’d like; it’s taken a while to get the IPM’s clinically certified laboratory established, and in many cases where we’ve established potential tests, clinicians aren’t always ready to integrate genomic information into patient care.
There’s been a lot of hype around precision medicine, and there’s an expectation that it’s going to revolutionize medical treatment. But that’s a slow process, and precision medicine is still a long way from being commonplace and having the massive impact that we anticipate. It’s becoming state-of-the-art care for cancer – you shouldn’t get cancer treatment without at least having your tumor sequenced over the potential drivers, because therapy really does depend on the nature of the tumor. But in other areas, progress is slower. Ultimately, the more we understand about polygenic diseases, the more we’ll be able to use that information to be able to tailor treatment to the patient.
That’s where patient self-advocacy can play a role, too. I think that there’s going to be increased recognition that genomic information can be informative about patients’ disease predisposition and treatment options. Connecting genomic information to medical records so that it’s clinically useful is one step in the right direction, but the next is to create and implement a mechanism for getting that information back to the patient as quickly as possible. It’s their data – their genome – that we’ve sequenced, and patients have a growing interest in obtaining as much information as possible so that they can advocate for their own medical care. I think it’s important for them to have that kind of information. Here at Hershey, we have open portals so that patients can look at their own medical records. I think that kind of openness needs to extend to genomic information as well, particularly when it’s linked to medical care. We’re moving into a new era of healthcare; instead of an all-knowing physician and a patient who listens, we now involve a team of people – and the patient is just as important as any other member of the team.
Penn State Hershey Institute for Personalized Medicine
Located: Penn State College of Medicine, Hershey, Pennsylvania, USA.
The Institute for Personalized Medicine features a biorepository that stores blood and tissue samples along with a database of patient information, and a genome sciences core that handles nucleic acid analyses from quality assessment to next-generation sequencing. The IPM’s ongoing projects include investigations into ALS, autism, diverticulitis, epilepsy, aneurysms, Parkinson’s disease, osteoporosis and cancer – and researchers anticipate projects examining psychiatric disorders, bone healing and age-related macular degeneration.
The institute includes one pathologist, three CLIA lab staff members, four consenters, seven genomics core members, 10 bioinformaticians, and approximately 12 students and 25 clinicians conducting IPM-based research projects.
Dealing with data
One concern precision medicine raises is the need for protocols regarding data collection and processing. We’re in an exponential phase of data accumulation, but we don’t know how to take full advantage of it. Part of the problem is that we can generate data for a particular purpose – for instance, treatment or billing – but when we try to mine that data for research, it’s not useful. Collecting data that can be valuable from multiple perspectives is a big challenge, and disparate data often contain key information. Twitter is better at predicting flu outbreaks than epidemiological data. So there are ways to take advantage of the information out there, but there’s also a lot of noise. Separating signals from noise is one of the exciting directions I think informatics will take in the next couple of years.
There’s a lot of data out there, if only we can learn to use it properly. For example, the IPM has an alliance with the New York Genome Center for studying ALS. We share detailed phenotypic and genomic data – and we do it in an established common language, so that researchers from any site can understand the information provided by any other. There is a strong push to establish common data languages, so that it becomes easier to aggregate information on patient populations.
I think it’s important for medical researchers to be well versed in both computer analysis and biology. At the IPM, we spend a lot of time training the next generation of physicians and basic researchers to have a foot in both doors, because it’s so important for laboratory medicine professionals to understand what information is already out there and how best to take advantage of it. The people who can cross the boundaries between life sciences and computing are the ones who will lead the personalized medicine programs of the future.