in the Cascades. He wore a flannel shirt, black pants, and trail runner shoes, which he casually put up on a coffee table in his office, next to a stack of books and scientific papers as we sat down to talk. The place gives off the irreverent vibe of many labs, with yellowed clippings taped to the wall from the satirical newspaper The Onion, and photos of lab members climbing some picturesque peaks.
Baker grew up not far from his current-day office, in Seattle’s Montlake neighborhood. His parents were professors at the UW, in physics and atmospheric sciences. Baker didn’t take much interest in science as a kid, although he got good enough grades at Garfield High School to get into Harvard University. He concentrated there on social studies, particularly philosophy, until he got bored his senior year.
“It seemed like a lot of talk,” Baker says. “I decided it was just talk and there wasn’t a lot of content to it.” Biology, in contrast “seemed like something you could really learn.”
Still, he had no idea of his career path when he graduated. He took off to see the world, and spent about six months in 1984, when he was 22, traveling through China, India, and Nepal. It was the first year China was opened up to individual outside visitors, and he called the experience “exhiliarating, but not really constructive.” He met other Westerners who would teach English for a while, make some money, and pursue their next travel adventure. “By the end of that time I was eager to do something that was actually going to contribute to the world,” Baker says.
On his return, Baker entered UC Berkeley for graduate school in biochemistry. He dabbled in various fields, from brain science to developmental biology. During graduate school, he met his wife, Hannele Ruohola-Baker, then a biochemistry grad student at Yale. (She has gone on to become an accomplished scientist in her own right, as a UW stem cell researcher at the new South Lake Union labs.)
Baker got his doctorate in 1989, and did a five-year postdoctoral stint at UC San Francisco. He says he used that time to “learn as many things as possible” and didn’t particularly focus on protein structure.
By 1994, when he was 31, he took a faculty job in the biochemistry department at the UW. All the stars aligned for him. His wife got a job at the UW, he got to come back to Seattle where he could enjoy skiing and hiking mountains, and it enabled his kids to be close to their grandparents. But there was more to it than that—the UW allowed him the freedom to sink his teeth into his emerging interest in protein folding.
“I like living in Seattle, and I’ve always felt the people here let me do the kinds of things I wanted to do,” Baker says. “Basically, everybody left me alone, which is what I wanted. I was able to pursue the things I was interested in.”
Protein structure and folding soon captured his interest at the UW. Baker and his students did experiments unraveling proteins and watching how they go back together again.
Baker was intrigued by how proteins obey the laws of physics, and like to settle into the lowest energy state possible. One analogy he uses to help people envision this is if balls were dropped onto the planet from high in the atmosphere, they’d want to settle in the lowest place above sea level, like a ditch in the Dead Sea. Scientists have long used imaging techniques to analyze protein structure, but it’s time-consuming and expensive, and there is a vast and unknown number of proteins in nature—maybe 10 million to 100 million, Baker says. The way to get a handle on the structure of a large number would be through computer models that can run algorithms that predict the 3-D structure based solely on an amino acid sequence.
“The DNA sequence alone doesn’t tell you anything about what the protein does, or how it does it. It’s just like the instructions for making a 3-D structure,” Baker says. “You need to be able to go from the genome sequence to proteins and their structure and how they work as machines to really understand how biology works.”
This requires a lot of calculating, on the level of supercomputing, and even then, many efforts to predict structure have been plagued by inaccuracies. One big step forward came in 2003, when the Baker lab showed,
