Here are four strategies that BME leadership could adopt
Data science has entered a new era, says BME Professor Philip Bourne, head of the University’s Data Science Institute. It is one characterized by unprecedented computing power, vast datasets and new tools to interrogate and manage that data. Organizations now have the ability to tap the wealth of open, complex and diverse digital data, combine them in ways that produce statistically significant conclusions, and translae those findings into actions that improve the human condition. “For BME departments, the latest evolution in data sciences represents an unmatched opportunity,” he says. “It has the power to transform everything we do.”
Bourne is in a position to know. As head of the University’s Data Science Institute, he has a window into the application of data science in virtually every possible field of research. His understanding of its applications to medicine was shaped during his tenure as associate director for data science at the National Institutes of Health and as a faculty member at the University of California, San Diego, where he pioneered the development of databases, including the Protein Data Bank (PDB) and algorithms to determine the three-dimensional alignment of whole protein structures and/or their binding sites.
Bourne believes that academic institutions have been slow to appreciate the benefits of data science. In his view, the business world has been quick to see the potential of this new environment and has taken the lead, not only in developing digital tools—he rates Google’s publicly available Correlate data mining tool highly—but also in applying data science to maximizing their productivity. “Organizations like Google, Amazon and Facebook quite early on recognized the commercial value of data science,” he notes, “but now data science approaches have been widely adopted by more traditional companies. It’s a factor in every single industry I can think of. It should matter more in academia.”
Bourne sees a number of potential applications of data science that could not only galvanize BME research but also change the way the way BME departments operate. Here are strategies that BME leadership could adopt:
Four Strategies for BME Leaders
Mine administrative information.
Collaboration is a necessary response to the growing complexity of the challenges that BME researchers undertake, but collaboration often arises serendipitously. Every university keeps a repository of grant proposals. Using data mining techniques, BME departments could sift through these repositories and locate faculty members in other disciplines who are working on related subjects.
Encourage faculty to make better use of publicly available data.
Total data from NIH-funded research in 2016 is estimated to have reached 650 petabytes. At the same time, there is approximately 2,700,000 million petabytes of data in the world. Tools now exist to combine biomedical data with other publicly available datasets to drive conclusions that would have been impossible to reach previously. “When you put disparate datasets together, you start to see something new,” Bourne says. This approach has applications for education and fundraising as well as research. For instance, departments could combine student transcripts with LinkedIn data to gain a better sense of courses that shaped careers.
The role of computation in guiding experimental work is now well-established, and the most successful departments have strengths in both areas. Chairs should look to hire faculty members who combine a BME specialty with expertise in applied mathematics, computer science, statistics and other data science fields.
Create a platform.
Airbnb revolutionized the hospitality industry by bringing together suppliers and consumers of lodging. Department chairs might bring together suppliers and consumers of relevant information by creating a platform that describes digital material resident in the department, including datasets, tools, protocols, workflows and course syllabi. “Research and education are incredibly inefficient,” Bourne notes. “At any given time, there is the danger of recreating something that has already been developed.” Such a platform might foster collaboration both within and without a department, while highlighting its activities.
The overarching promise of data science is that it can break down barriers that separate people and information in ways that were difficult or impossible previously. As a tool to maximize the productivity and impact of a department, Bourne believes it is unequaled.
Did you know?
The department is home to two nationally funded student training programs in systems biology and biomedical data science.
NIH Training Grant in Biomedical Data Science, Jason Papin (PI)
A pre-doctoral training program focused on teaching scientists to work at the interface of computer science, statistics, big data and biomedicine.
NSF REU in Multiscale Systems Bioengineering, Timothy Allen (PI)
Each summer, UVA trains undergraduates from a variety of STEM backgrounds in the skills, confidence, and mentorship necessary for successful careers in the exciting field of systems bioengineering.
UVA Biomedical Engineering Home
UVA Biomedical Engineering Research