Mike Langston

UTK electrical engineering and computer science

Throw a lot of data anyone’s way and the first thing they’ll ask is whether you can help them visualize what it’s all about. In the case of huge genome-scale data sets that is a thoroughly challenging job.

But Mike Langston and graduate research assistants Joshua New and Jordan Lefebvre are up to the task with novel algorithms, faster computational techniques, and web-based graphic tools such as the Ontological Discovery Environment that they’ve helped develop for biological scientists.

Langston’s LDRD companion project leader Elissa Chesler generates large quantities of unique, high-quality laboratory data in the course of her experiments for characterizing the effects of host genetic diversity on gut microflora. In her team’s LDRD work, researchers are faced with the task of integrating two massive and diverse data sets—one of intestinal gene expression in the tissue of mice and the other the genetic identity of gut microbiota. From such studies they hope to determine how susceptible an individual might be to intensely debilitating disorders such as Crohn’s disease and irritable bowel disorder.

This is where Mike Langston and his research team come in.

"I view my role as a computational scientist as something of a middle man in the tool chain," Langston says. "The biological scientist produces data in huge supply. My team uses statistical and graph theoretical tools to eliminate noise and irrelevant data, reducing the problem to its core. We crack this core with our most efficient methods implemented on our fastest computing platforms."

The JDRD team returns relatively small and highly distilled solutions (sets of genes and proteins) back to the biological scientist, who then is able to verify novel results with more traditional "wet lab" techniques.

"This is a wonderful time to be a scientist,” says Langston. "We’re confronted by an amazing confluence of emerging technologies that challenge us to find new solutions for new problems, every day."

JDRD project: High-performance computational tools for systems genetic analysis;
LDRD Project: Host genetic diversity as a variable selection environment for the gut microbiome, Elissa J. Chesler.

Mike Langston on one of his many overseas research trips;
Sweden’s Drottningholm Palace is in the background.

Langston and his colleagues devise novel combinatorial algorithms for high-performance computing platforms and use them to unravel subtle connections between genetic variations and gut microbial ecology. These techniques make it possible to pinpoint genetic polymorphisms that play key roles in gut microbial composition. The team then produces correlation graphs and other structures, and explores them using their clique-centric toolkit. They also apply new, tunable algorithms to visualize resultant clusters using tools from Jian Huang’s SeeGraph laboratory (illustrated above).