Mushegian Lab

Arcady Mushegian, Ph.D.

Director of Bioinformatics Research

Professor, Department of Microbiology, Molecular Genetics & Immunology
  The University of Kansas School of Medicine

Horizontal Tabs

Profile


Arcady Mushegian, Ph.D., did his graduate studiesin what he calls a “game-changing” era. “In the 80’s we went from a time when you worked a year to generate 10 reliable data points about one molecule to the development of high-throughput technologies that could generate millions of data points for every gene in the genome,” he says.

Hence, the need for computer-based technologies to crunch biological data and the emergence of the field of bioinformatics. As head of bioinformatics research, Mushegian collaborates with colleagues and develops computational strategies to analyze DNA sequences, compare genomes, or monitor global gene expression changes.

Those tasks require mathematical expertise, but Mushegian sees himself as a biologist first and says his undergraduate and graduate education in his native Russia taught him not to avoid mathematics. “In this country there is a sense that math is not for everyone, even not for every scientist,” he says. “But there we were expected to learn math and not be intimidated by it.”

His early interests were in microbiology and viruses and in 1989 he earned his Ph.D. at Moscow State University studying RNA virus movement in plants. To Mushegian, viruses were a logical place to start in comparative genomics, since the small genomes of many plant viruses had been sequenced by the late 80’s.

Mushegian came to the US in 1991 for postdoctoral training in plant biology at the University of Kentucky and then at the University of Washington. He spent his early postdoc learning “wet lab” techniques and studying plant pathogens. Concurrently, he started using the sequence analysis program BLAST to examine evolutionary relationships between plant and animal proteins. “At that time there were no webservers,” he notes. “We submitted sequences for analysis by email!”

Soon after Mushegian undertook a third postdoc with evolutionary biologist Eugene Koonin at the NIH. “By 1995, we knew what was coming,” he says. “The first bacterial genome was sequenced that year and every month afterwards new genomes were reported.” To make sense of that onslaught of information Mushegian and Koonin published a series of studies comparing bacterial genomes, as part of the emergence of the field evolutionary genomics.

Mushegian also became interested in a hypothetical uni-cellular creature now known as Minimal Genome, or the smallest bacterium that can be engineered to live on rich medium. In a 1996 PNAS study Mushegian and Koonin compared sequences of two bacterial species to identify and compute the minimal number of genes that the minimal genome may need—at the time thought to be 256. That number has been revised upward both by Mushegian and others to around 300. Researchers at the J. Craig Venter Institute are working on synthesizing a minimal genome from scratch.

Mushegian, a book lover, likens the field of comparative genomics to the study of language. “Look at me and E. coli and you’ll find that 2/3 of our genes are related,” he says. “It’s analogous to the idea that all Indo-European languages are related.”

On completing his postdoc, Mushegian worked with bioinformatics groups at San Diego biotech companies before being recruited to Stowers to set up a bioinformatics group in 2001. “I was impressed with the vision of the founders, and it didn’t hurt that resources were good and sustainable, “ says Mushegian, who was among the Stowers Institute's earliest faculty recruits. “We had a feeling we were creating something that would last.”

At Stowers Mushegian continues to pursue his interest in evolutionary genomics in addition to collaborative studies. One particularly satisfying collaboration was with former Stowers investigator and developmental biologist Olivier Pourquie, with whom he analyzed the formation of precursors to the mouse vertebral column, called somites. Somites rhythmically bud off—or segment—during embryogenesis, a process driven by oscillating gene expression. The goal was to determine how those oscillating signals give rise to somite patterns.

“That data was noisy and we could not fit an algorithm to the periodic pattern of gene expression,” says Mushegian. “But then we found a formula used by astronomers to determine periodic motions of planets and adapted it to computational analysis of gene expression patterns.” The work was published in Science in 2006.

A 2011 Development study was an evolutionary follow-up analyzing segmentation genes expressed in mouse, fish, and chicks. Mushegian’s focus on evolutionary biology is also reflected in his 2007 textbook, Foundations of Comparative Genomics, a historical overview of computational analysis of genomes.

Mushegian emphasizes that the goal of bioinformatics is not to just generate data but to reveal unanticipated patterns. “There is a quote somewhere that science is figuring how to discern the difference in things that appear similar and the similarities in things that appear different,” he says. “Discovering that is very rewarding.”