ERROR 1
ERROR 1
ERROR 2
ERROR 2
ERROR 2
ERROR 2
ERROR 2
Password and Confirm password must match.
If you have an ACS member number, please enter it here so we can link this account to your membership. (optional)
ERROR 2
ACS values your privacy. By submitting your information, you are gaining access to C&EN and subscribing to our weekly newsletter. We use the information you provide to make your reading experience better, and we will never sell your data to third party members.
Earlier this year, David Baker was sitting at his computer at the Institute for Protein Design, in Seattle, worrying about a situation familiar to anyone who’s ever been a graduate student: it was his turn to present at the weekly group meeting, but he didn’t have any results to show.
“I really hadn’t made much progress. And I was totally stressed,” Baker recalls. Knuckling down over the next few days, he coaxed some results from his computational project and met his deadline, enabling some productive conversations about what direction to head next.
The odd thing about this situation is that Baker is not a grad student, nor even a postdoc. He’s a group leader and head of the University of Washington’s Institute for Protein Design, with a staff of just over 130. After finishing his computations, Baker practiced the talk that he would give at the annual TED conference in Vancouver, British Columbia. At that conference, which took place in April, the nonprofit announced that Baker and his institute would receive a 5-year, $45 million grant from its Audacious Project, which collects money from philanthropic sources to boost innovative ideas. Baker’s grand idea? Design and synthesize proteins from scratch instead of tweaking what nature already produces. He believes that in doing so, his team will be able to create proteins with novel folds and structures that might find use in applications such as a universal flu vaccine, a smart drug-delivery system, and even solar cells.
Baker isn’t the only researcher leading a team that creates designer proteins, and the idea isn’t brand new. For instance, Baker’s lab has already spun out eight companies working to advance designer proteins to market as therapies for conditions such as multiple sclerosis and celiac disease. But scientists agree that he has been one of the catalysts in the field. He has, they say, jump-started a lot of the foundational work in protein design.
“It’s clear that Baker has had remarkable success,” says Mohammed AlQuraishi, a researcher at Harvard University who works on machine learning in biology. “I’m not shocked that he and his lab are the main driver” in the field.
And yet Baker is still running his own computations. He is in the lab every day, checking in to see what people are up to. While he admits that his group might not always appreciate such an active interest, it is part of what makes the Baker lab the Baker lab—along with daily social events and weekly chocolate tastings. He wants to foster a connectedness among not only his group members but also the protein design community. “I’m kind of fascinated by the sociology of scientific creativity,” Baker says.
In fact, Baker, who was born and grew up in Seattle, says science wasn’t his first love. During his time at Harvard as an undergraduate in the early 1980s, Baker started out majoring in philosophy and social science. It wasn’t until his final year in college that he decided to switch focus from trying to understand how the human brain makes decisions. This was after taking a course in developmental biology and encountering the book Molecular Biology of the Cell. “The first edition came out then, and it was really, really cool,” Baker recalls. Now in its sixth edition, the textbook has become a biology classic.
Still enchanted with the brain, Baker says that when he went to graduate school, “I thought I wanted to study developmental biology or neurobiology.” Instead, he joined the lab of cell biologist Randy Schekman at the University of California, Berkeley, to study how the cell is organized. As Baker got more interested in biological organization and the molecular structures that make it possible, he moved on to carry out a postdoc with structural biologist David Agard at the University of California, San Francisco, in 1989.
Originally, Baker explains, he planned to spend a year as a postdoc working on structural biology and then apply what he learned to cell biology as a principal investigator at a university. But his interests changed. “When I started doing structural biology, I got really fascinated,” Baker says. “I sort of came to think of it as the simplest case of biological self-organization.” So when he joined the University of Washington as a new group leader in 1993, he focused his lab on structural biology and the question of how proteins fold.
The original question he aimed to answer was how the amino acid sequences within proteins can result in such a variety of 3-D protein shapes. To try to understand protein folding in terms of amino acid sequence, Baker’s new lab started off doing classical biology experiments with techniques such as mass spectrometry and nuclear magnetic resonance spectroscopy before expanding into new areas. Only then did lab members begin building computer models to study the correlation between short amino acid sequences and the structures they formed.
The key realization came when Baker and his group started developing a software program called Rosetta in the late 1990s. On the basis of their previous work with protein fragments, Baker and his team realized a full protein folds when each of its sections is in the right shape at the same time as every other section. Starting from a Monte Carlo sampling algorithm, Rosetta could simulate this coordinated dance for any amino acid sequence inputted.
Rosetta first made waves in the world of protein-structure prediction in 1998 at the third biennial Critical Assessment of Protein Structure Prediction, or CASP, competition by successfully predicting how a sequence would fold. Over the next few years, as the code was refined and better computing power became available, the lab continued to refine its structure predictions.
But then, Baker says, “we realized we could go backwards, from structures to sequences.” The software could also design proteins according to the shape you wanted to create.
The Rosetta software began to grow as group members worked to add code that would design proteins from scratch and model interfaces between proteins. Jeffrey J. Gray, now a professor at John Hopkins University, joined the Baker lab in the early days of its success predicting protein structure from amino acid sequence. Because he came with a chemical engineering background, Gray says, he “wanted to build things.” Together with other team members, Gray added code to the Rosetta program that enabled it to model so-called protein-protein docking, which occurs when pieces of a larger protein assemble.
When Baker’s original team members began to leave the University of Washington and form their own research groups at other institutions, the question arose: What’s the best way to share credit for the code—the intellectual property—that went into optimizing Rosetta? The answer was RosettaCommons. Universities, companies, and other institutions can download the Rosetta software, but if they want to use the code commercially, they have to pay a fee. That money is pooled and fed back to support the community of users who edit, add to, and maintain the Rosetta code every day.
“Every day, we see what gets checked in from around the world,” Gray explains, “and you reach out to people if you want to work with them. Or if you see something relevant, you might share it with somebody who you think might be interested.”
For the past 7 years, using Rosetta, the Baker lab has been focusing on designing new proteins from scratch, a process known in the science community as de novo design. But it would be a mistake to think that all the work done in Baker’s lab is computational. Although Baker is hesitant to give a firm head count, the team is large (C&EN counted more than 80 members on the group home page at press time). And it contains both computational as well as experimental biologists, who work from the models outputted by Rosetta to synthesize proteins and then optimize them for their desired application. A viral capsid that the Baker lab first built for drug delivery in 2017, for example, was optimized through the Nobel Prize–winning experimental technique called directed evolution.
“Every problem that we work on, you have to make some creative leap to actually solve it,” Baker says. “So the question is, How do you build an environment where people will do that?”
Perhaps because of his earlier interest in philosophy and neuroscience, Baker describes the organization of his lab like a brain. “In a big group, you can think of each researcher as a neuron,” he explains. When you get many people working closely together and exchanging ideas all the time, he says, you can get emergent ideas—those that come about only because all the members of the group are interacting, sort of like how neurons work together to create functions that a single neuron couldn’t provide on its own. You wouldn’t get this if team members were isolated, Baker says, so he continually tries to get people talking across projects.
That philosophy applies to the Baker lab’s scientific work but also to its social interactions. The social activities that take place every afternoon and the weekly group meetings are meant to contribute to getting a “maximum information flow” across all the team members to make sure the “brain” is functioning maximally, Baker says.
Just in the past 7 months, that brain has created proteins that can zipper together like DNA (Nature 2018, DOI: 10.1038/s41586-018-0802-y), designed pH-driven switches for transforming protein conformation (Science 2019, DOI: 10.1126/science.aav7897), produced small proteins that tune receptor responses in the body (Science 2019, DOI: 10.1126/science.aav7532), programmed protein arrays on mineral surfaces (Nature 2019, DOI: 10.1038/s41586-019-1361-6), and predicted unknown protein-protein interactions in Escherichia coli bacteria (Science 2019, DOI: 10.1126/science.aaw6718).
Now working at Amazon, Baker lab alumna Franziska Seeger first found out about the lab and its Rosetta software when a lab mate introduced her to the game Foldit. Foldit, launched by Baker’s team and collaborators in 2008, provides a video-game-like interface for the Rosetta software that allows players to compete to build an optimally folded version of a particular protein or to design completely new structures that the Baker lab can then confirm in the lab (Nature 2019, DOI: 10.1038/s41586-019-1274-4). Wanting to design proteins that would bind to cell-signaling biomolecules called cytokines, Seeger joined the Baker lab as a postdoc in 2014. Along with other group members, she created from scratch a protein that would bind to the cytokine human interleukin-17, thought to play a role in autoimmune diseases.
“Chocolate Wednesdays definitely helped me to stay motivated,” Seeger admits. “But I also really believe in the vision of building computational tools to advance drug and vaccine development” and the idea that an entire team can solve problems together.
Ideas can come from anyone in the lab, Baker says. “I don’t believe in any kind of hierarchy. Anyone can have an idea that everyone else can listen to.”
If the Baker lab is a brain, the community that builds and supports the Rosetta software could be seen as an ecosystem of linked and cooperating brains, each contributing to the tool that is now licensed tens of thousands of times.
In early August, just like every year, more than 250 protein-folding experts—all members of RosettaCommons—will head to the mountains near Seattle. Turning off of Highway 2 onto Icicle Creek Road, the scientists will carry their hiking boots as well as their laptops for a 4-day retreat to discuss new ideas for advancing Rosetta.
Over 40 international labs continue to refine the software as part of the RosettaCommons community, adding insights from genetics and methods from machine learning. New ideas and applications are constantly emerging.
Collaboration on Rosetta happens all year long at events such as hackathons, but the main event, called RosettaCON, is still held every August. Somewhere between a conference and a summer camp, its culture and the broader culture around Rosetta is unique, Gray says.
RosettaCommons is much more than one man—David Baker—Gray emphasizes. Members drive forward scientific developments and also diversity and inclusion across members’ labs. But he adds that many of the principal investigators who actively keep the RosettaCommons community growing have been affected by their interactions with Baker. “His influence on the culture is amazing,” Gray says, “and has spread far beyond his lab.”
Baker believes his early passion for philosophy and social science has helped him build not only his lab but also the Rosetta community and connect with the Seattle biotech scene, which currently contains spin-off firms from Baker’s lab. The most recent is Neoleukin Therapeutics, which hopes to start clinical trials on a Rosetta-designed cancer therapy in 2020.
Successful spinouts as well as a grand vision may have been one of the reasons why Baker and the Institute for Protein Design found success as a TED Audacious Project this year. The cash adds to funding that the institute receives from the likes of the Bill & Melinda Gates Foundation and the US National Institutes of Health. Baker sees the TED money as a short-term way of ramping up the work done at the institute, which is why he plans to use it to hire more staff and improve the Rosetta software. “We’re just at the beginning of this technological revolution in protein design,” Baker says. “We want to attract really smart and creative people from all around the world to work on it.”
Baker’s aim, he says, is to build up the institute to be the Bell Labs of protein design, and he hopes to add one person per week for the rest of the year. In the 1950s, Bell Labs drove innovation in physics and information theory and helped lay the foundations of the modern digital era. Baker hopes his institute can become a similar hub.
But predicting the pace and direction of scientific progress is a harder problem than predicting protein structures from sequences or designing proteins, Baker admits. “I’m not one of those visionaries who can see 10 years out about where things should go,” he says. “I like to try and see 3 months out, and that gives me some intuition about what problems are good and what directions will be profitable.”
Join the conversation
Contact the reporter
Submit a Letter to the Editor for publication
Engage with us on Twitter