By Charles E. Reineke
Back in 1993, Dmitry Korkin, then a teenage numbers whiz from the Soviet Socialist Republic of Kazakhstan, couldn’t imagine anything more thrilling than beginning his studies with the world-renowned mathematics faculty at Moscow State University. And it was indeed an exciting time, especially after the shooting started.
Kids from Kazakhstan weren’t typically admitted to Moscow State, Russia’s most prestigious institution of higher learning. Korkin defied the odds. Not only did he ace the entrance exam to an elite, Moscow State-affiliated prep school, he went one better, pointing out errors and false assumptions in the test questions. His precociousness paid off. He was admitted to Kolmogorov High and a year later, began his studies at the venerable university.
Korkin’s parents and little sister, ethnic Russians nervous about political instability in Kazakhstan, packed up and moved with him to the capital. As fate would have it, the family arrived just ahead of an attempted coup d’état.
“The route of my bus to the high school was basically next to the bridge where the tanks stood,” Korkin says. “For me it was fascinating. If it were today, of course, I’d be as far away from that place as possible. People were shooting, you know? But I remember riding in the bus, seeing the tracer bullets over the Moscow River and thinking, ‘Wow, this is really cool!’”
Korkin, now a 31-year-old assistant professor of informatics and computer science at MU, hasn’t lost his taste for adventure, though these days he pursues the less-perilous path of pushing the frontiers of computing power. Korkin works in the rapidly evolving field of bioinformatics, a discipline that employs huge databases, machine-learning algorithms, complex statistical assays and other computationally intensive activities to illuminate the inner workings of biological systems.
His most recent area of investigation, for which Korkin was awarded a five-year, $613,000 CAREER grant from the National Science Foundation, involves the curious phenomenon of “molecular mimicry,” a process by which pathogenic proteins bypass a host cell’s defenses by adopting the structure and behavior of more welcome polypeptides. Just as some animals change colors to hide from predators, Korkin says, so do pathogenic proteins use mimicking to disguise their true identities. “With successful mimicking,” Korkin says, “the pathogen can not only avoid being attacked by the host’s immune system, it can also disrupt important functional processes in the host that are governed by protein-protein interactions.”
Because molecular mimicry is thought to be common to all microbic invaders, he adds, computer-assisted predictions of pathogenic interactions could, potentially, lead to the conquest of infectious diseases in plants, animals and people.
Korkin says his passion for using computers to solve problems began in childhood, but became more pronounced during his years at Moscow State.
“I was interested in things that went beyond the traditional mathematical approaches,” he says. “When I learned that there was a professor, Alexander Grusho, who used the principles of information theory to detect and predict security vulnerabilities of computer systems, I was very interested.”
After earning a master’s degree, Korkin left Moscow State to pursue a doctoral degree at Canada’s University of New Brunswick. There, at the urging of a friend, Korkin was introduced to Lev Goldfarb, an expert in the branch of artificial intelligence known as machine learning and pattern recognition. Goldfarb’s research, the friend said, was “absolutely the coolest thing ever.” Korkin agreed, and soon found himself happily toiling away on projects designed to teach computers to “think” for themselves. It helped that New Brunswick’s small campus was scenic but staid. “I arrived there in late 1999,” he says. “Because all I did, basically, was study I was able to finish my dissertation in the summer of 2002.”
That dissertation laid the groundwork for Korkin’s subsequent use of machine learning to help computers recognize patterns in the structure and function of proteins. He later refined these skills during postdoctoral stints at two big-city institutions: New York’s Rockefeller University and the University of California, San Francisco. At UCSF especially, Korkin was able to use his expertise to assist some of the world’s best life scientists. Among them was Elizabeth H. Blackburn, with whom Korkin later published a paper demonstrating that computers could accurately predict certain protein-protein interactions. Blackburn, an expert on telomeres, protein structures at the end of chromosomes that play a crucial role in genetic stability and cell reproduction, was awarded the Nobel Prize earlier this year.
Korkin’s time in the Bay Area yielded other benefits, too. While completing his postdoc at UCSF he met his future wife, Helen, at a party in Berkeley. The couple’s first child, Daniel, was born in September.
Leaving the big city for Columbia has been something of a lifestyle adjustment, especially since Helen, a native of Moscow who moved to San Francisco at age 14, had never lived in a small town. But Korkin says MU’s research environment, one in which he could actively work with engineers, mathematicians, bioinformaticians and life scientists, was just too good to pass up. “The collaborative spirit I saw during my first visit was very much like that back at UCSF — very distinct from the other universities that I applied to,” he says.
Thankfully, it turns out that Helen, too, has found Columbia to be a congenial place. “After we made our first trip out here together she said, ‘OK, this is not too bad,’” Korkin recalls with a smile. “And now we are very happy.”