Training the Next Generation of Computational Biologists


By Michaela Webb, NASA Space Grant Intern for CyVerse

A University of Missouri undergraduate research course used CyVerse technology to teach computational biology.

Gavin Conant, associate professor of Animal Sciences and Bioinformatics at the University of Missouri (MU), used CyVerse technology to teach a summer research course for undergraduates aimed at preparing the next generation of biologists to tackle the future's computational challenges.

The nine-week-long course, titled Educating for the Grand Challenges at the Intersection of Biocomplexity and High-Performance Computing, brought 12 undergraduates from around the country to MU to study computational biology. Conant used Atmosphere, CyVerse's cloud-computing platform, to teach the computational aspects of the course.

To help students tackle the complexities of computational biology, the class paired biology students with computer science students so that teams had an understanding of both the science and the computational aspects of the research.

The experience began with an intensive two-and-a-half day workshop about how to use CyVerse and other digital tools. Students then went to work in their respective labs, reconvening on Wednesdays for computation workshops taught by Conant.

Projects ranged from researching changes in gene expression of plants infected with a parasitic nematode to investigating the genetic diversity of broccoli. “These projects all had big datasets that were hard to analyze,” Conant said. “The students were trying to learn new ways of analyzing these large datasets to find more comprehensive results.”

The students wrote software scripts to simulate genetic drift using Perl programming language. They then modified those scripts to add the effects of natural selection. They also experimented with running programs in parallel to increase the speed of the analysis.

Although Conant taught the same course during the previous two summers, this year was the first time he used CyVerse. Conant said instructors had struggled to find adequate technology for teaching the class in its first two iterations. They tried using MU's computing cluster, which is a set of connected computers with more computing power than a single machine, but found that it didn’t fit the needs of the course very well.  

“One of the problems with the cluster was that there was a whole level of complexity beyond just using a UNIX computer. When you’re on the head node, which is the primary computer in the cluster, if you want to do a compute job you have to ask for compute resources and log into another node. That was confusing to the students,” Conant explained.

Students using the cluster for their research also worried that if they made a mistake, they might damage the machines or cause another researcher’s program to crash.

Conant said that CyVerse’s platform, Atmosphere, was the perfect solution. “With Atmosphere, I could tell the students: Look, you cannot break the system. I think the students felt more comfortable with that, because they knew that they were playing in sort of a sandbox.” 

Atmosphere also helped Conant tailor the experience to fit the needs of the students. “The really nice thing about Atmosphere is that I could set up exactly the environment that I wanted them to see,” Conant said. “I could install the tools that I wanted, I could pre-stage the data wherever I wanted to, I could test everything in advance, and I could know that when they started working, the computer interface was going to look exactly like mine.” 

Students who hadn’t previously thought of using CyVerse in their research left the course with new skills and ideas, and an interest in bringing CyVerse into projects at their respective institutions, Conant explained.

And because Atmosphere is a cloud-computing tool, collaboration between students didn’t necessarily stop when the course ended, he added. “The fact that it’s a shared resource, that they can go back to their home institutions and they could still be working together even if they’re in very different places – that’s powerful.”

Image above: Brandon Nielsen and Jacob Rosenthal present their project at the University of Missouri's Undergraduate Research Summer Forum. (Photo courtesy of Gavin Conant/MU)