Researchers from the University of Maryland's Baltimore and College Park campuses have opened a new informatics and bioimaging center that will focus on facilitating genomic research, medical information management, and translational science.
The new Center for Health-related Informatics and Bioimaging will combine computing resources at UM College Park with clinical data and biomedical expertise at UM Baltimore.
At the center, cross-institutional teams will work on projects in genomics, health records management, and image analysis using multicore systems and cloud computing infrastructure as well as new methods of organizing, visualizing and analyzing massive amounts of data.
Leadership of CHIB will be split between the two campuses: Owen White, associate director for bioinformatics at the Institute of Genome Sciences and professor in the School of Medicine, will direct CHIB’s activities in Baltimore, while Amitabh Varshney, a professor of computer science and director of the university's Institute for Advanced Computer Studies, will lead efforts in College Park.
Recently, BioInform spoke with UMB’s White about his institution’s role in the joint center. What follows is an edited version of the conversation.
What will your activities at UMB be specifically as part of this partnership?
It’s going to be a broad set of ventures because we are organizing a bunch of different disciplines together into a more focused activity. My current image of what that means is we have assets such as data from medical imaging, genomics, medical informatics, [and] patient health records and [we want to expose] that to a larger research community represented by people here at UMB as well as down in College Park.
A scenario that you could imagine is this: as people get access to anonymized patient records they could develop machine-learning algorithms to look at that data and identify people who are at some type of health risk, and then that information is reported back into the electronic medical record and becomes part of the decision-making process that’s available to the doctor and to the patient.
Where is all the data coming from?
A constellation of data sources that’s as large as this city or the mid-Atlantic region. The hospital system here [has] a number of different hospitals, all of which have medical records that will eventually be centralized, de-identified, and made available to the research community. There [are] a large number of cohorts that are being recruited by community outreach activities … and there are a lot of researchers that would be interested in getting access to that information to find people who want to participate in different types of studies.
There are also lots of data generators at College Park, including people who are generating genomic sequence or collecting geographic or social information. So the data streams are going to be quite varied.
You seem to have a pretty robust informatics group at UMB’s Institute of Genome Sciences. How is this partnership a good fit for you and College Park?
It’s beneficial in both directions. First of all, we do a lot of basic research and I have collaborators at College Park that I’ve been working with for some years, generating publications and working on data that also goes toward generating research grants. My group of about 35 people is composed of engineers, analysts and faculty, so in addition to doing research here we are also providing fee-for-service activity activities that people at College Park could take advantage of. We also have a lot of experience of taking research-grade software that has been developed by researchers like people at College Park and … folding it into production analysis systems that are part of our fee-for-service paradigm. So there is a lot of exchange that’s going in both directions already and we are hoping that will start to grow into a larger operation.
Tell me a little bit about the software and services you have in place at UMB.
In my group our bread and butter is mostly genomic analysis; we do annotation of bacterial and eukaryotic genomes; we’ve done a lot of comparative analysis. Another activity that’s been going on here is generation of microbiome data and we do a lot of data processing for that, where we identify genes or look at community compositions.
There is a clinical informatics group that helps people with everything ranging from clinical database management to statistical analysis to study design. There is also a very large imaging infrastructure that’s managing the imaging data in the hospital and handling the transactions where people are analyzing that information and generating reports that are then passed on to the doctor. There are electronic medical records that are being managed within the hospital [and] we’ve got a lot of outpatient clinics that are gathering information about patients, variables and test results, all of which are being encoded into electronic systems. There’s also general enterprise IT … handling email, calendar systems, document storage within the different schools here.
Do you think you’ll need to increase the infrastructure you have on the ground in order to handle all the data that’s coming in?
I’d like to think of it more as a consolidation where the more that you could be centralizing the computational infrastructure the better … right now it's fair to describe the information landscape as stove-piped — data is distributed in lots of network closets and underneath desks throughout the campus. I have about 17 racks of computing equipment and we are joining that in with the enterprise computing equipment that’s also on campus. There is a lot of benefit from centralization that we hope to gain.
How about staff? Are you hiring?
A lot of questions regarding the direction of the center still remain to be developed. It’s going to be the result of partnership with us and the state as well as different people that may very well become members of the center. I have every expectation that we’ll be able to expand because typically when you start to engage in centralization of these types of activities, more and more people start to buy in and the justification is there for expanding the group. So I can’t answer that question just yet; we just have to look to the future.
Will the new center focus on particular disease areas?
The center I think in a lot of ways should be disease agnostic. I don’t think we would want to constrain ourselves to be exclusive to any one disease area. We have a lot of different strengths that are basically as diverse as the research that’s here on campus.
Have you identified any specific projects for the center?
I think the first place where we are going to be very successful is the clinical informatics division led by Kathleen Tracy. She is focusing on centralization of a lot of diverse information but especially patient records, cohorts, and different outpatient and inpatient data resources, linking that between the hospital and the research community. Her work will be one of the center’s flagship areas of focus.
Are you opening up a separate space for the new center?
Right now, it looks like there will be something on the order of one-third to one-half of floor space in the health sciences library where we can set up shop. Membership into the center is going to be happening at a few different levels. In some cases, faculty and staff may have space over at the library. There will be other categories of faculty who will be very active members of the center but would remain in their departmental location. We will start viewing a whole lot of the informatics staff and activities that are happening at my institution as also being part of the operation. So it’s going to be brick and mortar as well as virtual.