When he was in the third grade, Green came across a book of number theory problems in a local public library. He was instantly hooked. Mathematics seemed like a game to him, and when he learned a few years later that people did it for a living, he vowed to become a mathematician himself. But he also had an early exposure to computational biology. Green’s next-door neighbor was a statistical geneticist at the University of North Carolina in Chapel Hill. During high school, Green worked at a summer job arranged by his neighbor that involved writing computer programs to analyze genetic data.

Green earned an undergraduate degree in mathematics from Harvard and a doctorate from the University of California, Berkeley, where his dissertation combined work on abstract algebra and functional analysis. But after several years as an assistant professor at Columbia, he became dissatisfied with pure mathematics. “The math I was doing didn’t seem to have any connection to the real world,” he says. “I started feeling that I’d like to be working in an area where I could apply mathematical ideas to something more concrete.”

About that time he read the book Molecular Biology of the Gene, by James Watson. The intellectual rigor and “almost mathematical” elegance of the book rekindled his interest in genetics. He moved back to the University of North Carolina, where he began working with research teams studying the genetics of heart disease and the regulation of the human fibrinogen gene. He spent a lot of time at the bench, getting a feel for the problems and potential of biological research. “I was learning to think like a biologist,” he says.

In 1986, he moved to the company Collaborative Research, in Waltham, Massachusetts, to help construct a linkage map of the human genome. There he met Lander, who at that time also was making the transition from mathematics to biology. Green and Lander wrote a key computer program that allowed large amounts of data to be incorporated into the map. It was Green’s first exposure to genomics—the study of large-scale maps and sequences of genetic information.

In 1989, finding that he missed the collaborative atmosphere of academia, Green moved again, to Washington University School of Medicine, in St. Louis. There he joined a genetics department consumed by the emerging challenges of genomics. Washington University was becoming a leader in the new field of high-throughput genetic sequencing. But these initial sequencing efforts were encountering major difficulties. The data emerging from automated sequencing machines were of variable quality; some “calls” for individual DNA bases were rock solid but others were ambiguous. Furthermore, geneticists did not have a good way of combining the raw data, or “reads,” from the sequencing machines into longer sequences.

Green designed two computer programs to solve these problems. PHRED (for “Phil’s revised editor”) assigns an error probability to each base call so that the reliability of different parts of the read can be assessed. PHRAP (for “Phil’s revised assembly program”) takes the output of PHRED and combines the reads into longer stretches known as contigs. Their names don’t have the Madison Avenue slickness of mega-selling software titles, but they don’t need it.

	PAGE 4 OF 9	Go Back | Continue