Saving real lives with artificial intelligence

This is the home of the UW’s Institute for Protein Design (IPD), founded and led by Professor David Baker, a winner of the 2024 Nobel Prize in Chemistry. This cross-disciplinary institute is harnessing artificial intelligence to shape the future of medicine and beyond, work that will improve people’s lives. At the heart of that work are the faculty members, doctoral students, postdoctoral researchers, staff members and undergraduates like Faruq, who come from all parts of the country and the world.

Baker describes the IPD as a “communal brain”— a network of connected people from different fields of study, each contributing their unique perspectives, ideas and expertise to solve complex challenges in medicine, technology and sustainability. “We look at some of the biggest problems that confront humanity,” says Baker, a Seattle native who has spent a lifetime studying proteins and founded the IPD at UW Medicine in 2012. “And then we think, ‘Could we design proteins to address those problems?’”

A trailblazer in protein design

Protein design is a new field of biology, one that Baker helped found. Scientists have been trying to construct synthetic proteins since the late 1980s — and while Baker’s team made many breakthroughs, one piece of the challenge eluded them: quickly predicting the protein’s unique shape. The way a protein coils itself up into a specific shape determines what it can or can’t do. It’s the key to creating proteins that carry out specific tasks, like binding to cancer cells so they can be found and treated.

About seven years ago, IPD researchers started experimenting with artificial intelligence. Using a deep-learning tool — an AI approach that analyzes patterns in data to make predictions — they trained it to predict how proteins would fold.

Suddenly, complex work that might have taken years in the lab could be done in minutes on a computer. This discovery was named the 2021 Breakthrough of the Year by the American Association for the Advancement of Science and helped lay the groundwork for Baker’s 2024 Nobel Prize.

Today, the team’s suite of homegrown AI tools for modeling and generating proteins are even more sophisticated. “It’s fun to see how fast things are moving,” says Baker about the progress he’s seen over the past few years. “We release our AI methods for anyone to use, so we’re seeing an acceleration of protein design across the world.” More people working in protein design means more new medicines in the future.

Designing hope for patients

That possibility is what brought Christina Savvides, who’s working toward a UW medical degree and a doctorate in molecular engineering, to the IPD.

“I want to be what is called a physician scientist,” says Savvides, who chose the UW for its top-notch medical training program and collaborative research environment. “A core principle is the idea of ‘bench to bedside,’” she explains, referring to a medicine’s journey from being developed in the lab to becoming available to patients. “I want to offer something to patients who have no viable treatment options.” The protein research she’s currently working on could be that hope.

Because the human body runs off of proteins, they’re the key to unlocking new ways to prevent and treat diseases, including cancer. “One of the challenges in recognizing cancer cells has been they don’t look very different from normal cells,” Baker says. “We’ve recently developed a technology that can measure key markers of cancer proteins. This gives us a whole new way of targeting cancer cells.”

Other IPD breakthroughs that might save or improve lives could include an effective antivenom for snakebites, a universal flu shot you’d need only once in your life, antibiotics designed to combat drug-resistant bacteria, a better vaccine to prevent malaria (the number-one killer of young children worldwide) and treatments for autoimmune disorders like rheumatoid arthritis.

Savvides describes the current treatment options for rheumatoid arthritis as a “sledgehammer” approach. While these medications work well, they result in collateral damage to tissue and organs, and many people opt to stop treatment. Her goal is to find the key to target just the immune system “and avoid that destruction.”

The process of getting a new medical treatment from a university lab to patients is complex, requiring clinical trials and approvals from regulators. The IPD has a robust pipeline to kick things off by creating spinout companies that can, for example, test and commercialize new medicines so they can get to doctors and patients. So far, Baker has launched 21 biotech companies, many of them in Seattle — with more on the horizon. That includes Xaira Therapeutics, the single largest new biotech firm in 2024, with $1 billion in committed funding.

The UW is the ideal place for new medical discoveries to take off, Baker says, because of its collaborative culture and location. He adds: “Seattle is a great place for this work because of the resources and expertise in the city.”

Accelerated by AI

These exciting new developments rely on fast, accurate AI tools — and the expertise of research scientists like Andrew Borst, ’19, with a giant microscope that takes up half of a basement room.

Borst leads a team that uses a high-powered electron microscope to see if the lab proteins match their computer-generated blueprints. He often thinks back to a pivotal moment in December 2022: the first time every AI-designed protein that came under his microscope was behaving in real life exactly as the computer models had predicted. It meant the institute’s AI tools weren’t just fast — they were also accurate. And that opened exciting possibilities, especially for new medicines.

Part of what sets the IPD apart is this combination of computational and lab work, where promising AI-designed proteins are produced in the lab and then tested. On one side of the IPD, researchers huddle around computers to talk through programming code and examine vibrant models of possible proteins — cylindrical, circular and ribbonlike shapes that dance on screens. And nearby through gleaming glass doors, white-coated team members navigate around each other in a lab as they set up experiments to validate and modify the AI designs.

Testing is often an iterative process, requiring the designer to go back to the drawing board or tinker with the algorithm to get the protein’s shape and function right. It’s equal parts exciting and frustrating. And it’s what brought Borst to the UW.

“I chose the UW for graduate school because I wanted to be where groundbreaking research was performed,” says Borst, who grew up in Everett, earned his UW doctoral degree in biochemistry and is now IPD’s head of electron microscopy. Over his 15 years at the UW, he’s watched how AI has magnified the impact of that research.