Recommended by: Michael Brudno, University of Toronto
Improving the sensitivity of structural variation detection is a key driver of Paul Medvedev's research. "There are certain variants that we currently just cannot detect," says Medvedev, a postdoc at the University of California, San Diego. "Being able to detect these will give us an opportunity to better understand the formation mechanisms of structural variants and how these are disrupted in tumors."
A computer scientist by training, Medvedev found himself initially "excited by the life sciences because of the impact it can have on our lives," he says. But the interdisciplinary nature of his work has at time proved challenging. "I am always conscious of how to present it to audiences of different fields," Medvedev says. "The biggest challenge is when these audiences are intermixed."
Publication of note
In 2010, then a PhD student under the supervision of Allan Borodin and Michael Brudno at the University of Toronto, Medvedev and his colleagues published a paper in Genome Research describing CNVer, a tool that allows users "to integrate depth-of-coverage with mate-pair information for detecting structural and copy-number variants," he says. "There were already tools using each of these signals separately, but we showed that one can really improve the accuracy if you use both of them."
That high-performance computing is on the upswing has been a boon for genomics research, Medvedev says. "Two years ago, I would have wished for a machine with infinite memory, but now we have machines with a terabyte of memory, so that dream has pretty much come true," he says. That storage solutions for large volumes of data are also no longer a pressing issue, he adds, "had made my research a lot more streamlined by allowing me to focus on other issues."
And the Nobel goes to…
While the prize committee doesn't give awards for computer science, Medvedev says "it would be great if someone using our methods won it."