CBB Program Evolves to Expand its Reach.
For decades, computational processing power has doubled every two years, a trend known as Moore’s law. Moore’s law has enabled bigger and bigger problems to be tackled with computation, but recent advances in genomics are a source of concern: the big data generated by techniques like next-generation sequencing are growing much faster than Moore’s law, far outpacing the power to process them.
The explosion of data represents a boon for science and medicine, but it also poses a challenge for science education. Alex Hartemink, the recently appointed director of Duke’s Program in Computational Biology and Bioinformatics (CBB), thinks the type of training students receive has to reflect this rapidly changing environment.
“It doesn’t really make sense for universities to teach people how to operate only in the world that exists, but rather how to operate effectively in any world that might exist,” Hartemink said. “The question then becomes, how do you achieve that? How do you train students to constantly be surveying the landscape, seeing what opportunities and challenges exist, and train them to be adaptable?”
In his new position, Hartemink inherits a program that has already shown an ability to evolve. When the program was first established a dozen years ago, it was called the Bioinformatics and Genome Technology Program (BGT). As the program grew, it became clear that more and more faculty in computational biology were teaching classes and supervising PhD students.
“We made a mid-course adjustment and renamed the program, not as a marketing exercise but as a more accurate reflection of the areas in which Duke was growing,” explained Hartemink. “We constantly have to adapt. As these new technologies come along, we have to think about whether we are going to be able to teach the same courses we taught six years ago. That’s not likely.”
One thing that has been a constant all along is the collaborative and interdisciplinary nature of the CBB program, which Hartemink believes has contributed to successes both inside and outside the program. For example, he says that having computational biology students around to facilitate connections between faculty in different disciplines was one factor that helped crystallize other efforts like the Center for Systems Biology.
“It doesn’t really make sense for universities to teach people how to operate only in the world that exists, but rather how to operate effectively in any world that might exist.”
In Hartemink’s own laboratory, the presence of CBB students has helped forge collaborations with five or six different people representing different departments around Duke. He says those collaborations might have had different outcomes if they had involved a computational student who didn’t know much about biology and a biology student who didn’t feel very capable in quantitative analysis. That kind of an approach could also work well, but Hartemink thinks that having CBB students who are eager to do both has been more effective in leading to greater innovation.
“This program brings in really interesting students that don’t want to be put in a box,” Hartemink said. “They want to think about multiple things at the same time. They want to innovate in new ways. That culture then becomes self-sustaining – because we have those kinds of people here, those kinds of people are attracted to the program going forward.”
Most recently, the program has attracted three new faculty in Barbara Engelhardt, Raluca Gordân and Tim Reddy. Engelhardt, a Berkley-trained biostatistician who studies human genetic variation and its impact on genome regulation, officially joins the faculty this fall. She said collaboration is the greatest asset to her work at Duke.
“Once or twice a week I may be sitting down with people who are entirely wet lab scientists or who are doing completely different things than I do, but we all have a common goal,” Engelhardt explained. “There is so much cross-talk because there are so many joint activities for us to be involved in. It makes it easy to collaborate, even though we may have completely different expertise.”
Engelhardt is a purely quantitative scientist, which sets her apart from the other recent recruits. Both Gordân and Reddy are setting up labs that have both a dry (quantitative) and a wet (experimental) lab component. Gordân, who trained on both sides of the aisle as a graduate student under Hartemink and a postdoc under Harvard’s Martha Bulyk, is now using a specific experimental technique in her own lab to explore how different DNA binding proteins operate. She is the first at Duke to use the technique, which was pioneered by Bulyk.
“Raluca has the expertise to do these experiments, and she wanted to run the lab herself,” Hartemink said. “With the rise of computational biology programs around the world, people are being trained for comfort in both wet and dry science. Well-trained computationalists have become accustomed to doing experimental science. They want to stay on the cutting edge and generate their own data. At the same time, they are sufficiently comfortable with the analysis end, so they can do that as well.”
Reddy, who has set up his genome biology lab to be half dry and half wet, says that having the experiments and the analysis in lockstep makes it possible to generate an incredible amount of new knowledge in a short period of time.
“Having all of those resources in one place means the loop is just a little bit closer,” said Reddy, who studies mechanisms of gene regulation. “I make a point of training the people that are good at doing the experiments in the finer points of analysis, and teaching the people that are good at doing the analysis how to do the experiments as well. Because there is a better understanding, the follow-up can happen immediately. I also think that it helps people get a bigger picture view of the research and develop keener instincts about some of the questions that are going to be most interesting to pursue.”
“I make a point of training the people that are good at doing the experiments in the finer points of analysis, and teaching the people that are good at doing the analysis how to do the experiments as well.”
The inclusion of wet as well as dry approaches in training is one of many steps in the evolution of the CBB program. In his new role as director, Hartemink wants to look at every aspect of the program, to ensure it is the most impactful computational biology program possible.
“How are we doing in regard to our curriculum, how are we doing in rotations, how are we doing in first year advising?” Hartemink asks. “How are we doing in maintaining a good culture and community for our students? How are we doing in placement, how are we doing in recruiting? These are some questions we ask to ensure that everything we do, we do with excellence.”
Once he is confident that the existing CBB program is continuing to operate smoothly, Hartemink wants to consider expanding the program’s reach, to integrate better with offerings at the undergraduate and postdoctoral levels. Though he has no formal plans in place, he is interested in exploring relationships with undergraduate programs, certificates, and departmental majors. He also wants to think about how to make Duke a destination for postdocs coming from other computational graduate programs, creating a community for them to connect on campus and training them to be highly effective interdisciplinary scholars.
“It’s important that we not think on too short a time horizon, focusing on just the current stage of a graduate student’s career,” explained Hartemink. “We want to be training people to go off and have a long career in productive science, and hopefully innovative science, even as new technologies and ideas burst onto the scene. If we’re preparing people for that going forward, I think we’ll be really successful.”