A team from the Oak Ridge National Laboratory led by computational systems biologist Dan Jacobson developed a genomics algorithm that has achieved record-breaking computational speeds. On the Summit supercomputer, the team’s Combinational Metrics (CoMet) application achieved a top speed of 2.36 exaops – or 2.36 x 1018 (2.36 billion billion) calculations per second. It is the fastest science application ever reported.
For that effort, Jacobson’s team has been named a finalist for the Gordon Bell Prize, the highest award for supercomputing.
As the chief scientist for computational systems biology at the laboratory, Jacobson works to compare genetic variations within a population to uncover hidden networks of genes that contribute to complex traits. His career in science might never have happened but for a throat infection during his days at FSU that caused him to switch majors. “I initially enrolled at FSU as a music major. Halfway through my freshman year, this infection made me lose my voice for several months, and I wound up making a small transition to biochemistry,” he says with a chuckle.
Although biology makes up a large component of Jacobson’s duties, he must also have a firm grasp of mathematics, statistics and complex supercomputing. “We train people to think across multiple disciplines,” Jacobson says.
Jacobson is collaborating with researchers at the U.S. Department of Veterans Affairs to find people who have used opioids, whether or not they developed an addiction. Trying to discover the complex genetic architecture responsible for human addiction requires poring over many possible genomic molecular variations in cells.
Jacobson’s research efforts regarding sustainable biofuels call on him to routinely team up with research partners at 15 other institutions in the U.S. “A lot of the collaborative efforts are also with labs overseas,” he says. “No one group can do it alone. The more data you have, the more robust your answers are.”