Brett Teng Gao, a senior Engineering student at Washington University in St. Louis, is making the most of his summer in the San Francisco Bay area. In addition to an internship with Roche Diagnostics, Gao spent a weekend at the Google-sponsored Artificial Intelligence (AI) Genomics Hackathon — the first of its kind— and his team won. It also was Gao's first hackathon.
Ben Hsu, Nandita Damaraju, Jo Varshney and Teng Gao (Photo courtesy of Jo Varshney)
Gao, who is majoring in computer science and biology with a minor in mathematics, was one of four members of the team, which also included a cancer researcher and two computer scientists. Their prize was a Titan X Pascal GPU — a very high-powered graphical computer processor designed to run virtual reality and complex algorithms quickly.
The AI Genomics Hackathon, held June 23-25 at Google Launchpad in San Francisco, brought together about 150 participants who organized into 40 teams. All came together to use AI, computations and biology to advance the understanding of neurofibromatosis type 2 (NF2), a rare disease that causes noncancerous tumors in the nervous system.
Gao said only about 20 percent of the hackathon's participants were undergraduate students, but the interdisciplinary collaboration he has learned at WashU was an asset to the team.
"This hackathon is really contributing to an ongoing patient cause for a rare disease," Gao said.
"Computer scientists and biologists are from two different worlds, and it's so rare to get them together. It is amazing how this event brought both sides together to solve a problem."
Hackathon participants had a unique set of materials to use in their quest. The hackathon's organizer, Onno Faber, an entrepreneur in the Bay Area, was diagnosed with NF2 and made the genome sequencing data from his tumor available to the participants.
Gao said that data was valuable because there is not enough data available on NF2, making it difficult to study, but the team combined its interdisciplinary skills to find the winning solution.
"Our approach to address this problem is to use a machine learning technique called transferred learning," Gao said. "We found a similar dataset for a sister disease that's far more common. We trained the machine learning algorithm using a bigger data set first, then we applied the algorithm on the rarer disease that we are tackling. Hopefully this will be useful for others studying the disease."
The School of Engineering & Applied Science at Washington University in St. Louis focuses intellectual efforts through a new convergence paradigm and builds on strengths, particularly as applied to medicine and health, energy and environment, entrepreneurship and security. With 94 tenured/tenure-track and 28 additional full-time faculty, 1,300 undergraduate students, 1,200 graduate students and 20,000 alumni, we are working to leverage our partnerships with academic and industry partners — across disciplines and across the world — to contribute to solving the greatest global challenges of the 21st century.