Can AI Crack The Biology Code?

Summary

Short Wave Podcast Summary: "Can AI Crack The Biology Code?"

Episode Information:

• Title: Can AI Crack The Biology Code?
• Host: Emily Kwong
• Producer: Burleigh McCoy
• Release Date: June 3, 2025
• Description: Exploring the transformative impact of AI on protein science, highlighting breakthroughs, applications, and future possibilities.

---

Introduction to the Protein Folding Challenge

In this episode of Short Wave, hosts Emily Kwong and producer Burleigh McCoy delve into the intricate world of protein science and how artificial intelligence (AI) is revolutionizing our understanding of the fundamental building blocks of life.

Burleigh McCoy (00:29): "Today I want to dig into how AI has shaken up the field of protein science. As in the fundamental building blocks of life proteins."

Burleigh explains that proteins, composed of amino acids, are essential for various biological functions, and understanding their precise shapes is crucial for deciphering their roles in processes like oxygen transport and photosynthesis.

Burleigh McCoy (01:05): "The ability of a protein to do its specific job... relies wholly on its unique, complicated shape."

---

The Complexity of Protein Folding

Determining the three-dimensional structure of proteins has been a long-standing challenge. Traditional experimental methods are time-consuming and labor-intensive, often taking months or even years to map a single protein's structure.

Burleigh McCoy (01:24): "They can take years and years... Essentially, a scientist needs to take the equivalent of a molecular photo of the protein."

Emily underscores the daunting task, likening it to visualizing a complex assembly from millions of possible configurations.

Emily Kwong (05:07): "There are so many theoretical ways a single protein could fold that it's a big problem to solve."

---

AI Breakthrough with AlphaFold

The narrative shifts to the groundbreaking contributions of Google's DeepMind and their AI model, AlphaFold2, which dramatically improved protein structure prediction.

Burleigh McCoy (02:30): "And for most of that 30-year history, participants have really only made incremental progress. But in 2020, Google DeepMind used AlphaFold2, that's its AI protein prediction model. And AlphaFold2 blew the other competition out of the water completely."

AlphaFold2's deep learning architecture enabled it to predict protein shapes with unprecedented accuracy and speed by analyzing the distances between amino acids and leveraging vast datasets of known protein structures.

Burleigh McCoy (06:13): "This is a type of AI called a deep learning program... it learns about proteins from a huge collection of protein structures that scientists have been building on for decades from their experimental data."

---

Real-World Applications and Impact

The introduction of AlphaFold2 has had a seismic impact on the scientific community, accelerating research across various fields.

Julian Bergeron (07:48): "I input a sequence, and then a few hours later, I had the model, and I was like, oh, my God, this just did it. And we'd been struggling with that problem for months, if not years."

Julian Bergeron, a structural biologist at King's College London, highlights how AlphaFold2 transformed his lab's research trajectory by enabling rapid protein modeling, which was previously a bottleneck.

Emily Kwong (09:46): "The mission statement that we have for the science program at Google DeepMind is to leverage AI to accelerate and advance science."

The accessibility and user-friendly nature of AlphaFold have facilitated its widespread adoption, allowing scientists to tackle complex problems in disease research, drug discovery, and environmental science.

---

Advancements with AlphaFold3

Building on the success of AlphaFold2, DeepMind introduced AlphaFold3, which extends its capabilities beyond single proteins to include interactions with other biomolecules.

Burleigh McCoy (10:07): "DeepMind released a new version, AlphaFold 3, which can predict the 3D structure of proteins and other kinds of biomolecules that they attach to."

AlphaFold3 provides a more detailed and accurate depiction of protein interactions within the biological milieu, enhancing our understanding of cellular processes.

Pushmeet Kohli (10:38): "It really gives you a more detailed and more accurate picture of what is happening inside the body... They are interacting in a very rich biological space."

---

Limitations and Challenges of AI in Protein Prediction

Despite its advancements, AlphaFold has certain limitations. The model performs best with proteins that have a single defined structure and may struggle with proteins that exhibit multiple shapes or have flexible regions.

Burleigh McCoy (11:31): "The model works best when a protein has a single defined structure. But some proteins have more than one shape or they have sections that are kind of flimsy."

Additionally, the accuracy of predictions is contingent upon the availability of extensive training data. Proteins or biomolecules with limited structural data pose a greater challenge for the model.

Burleigh McCoy (12:13): "The prediction ability depends on the amount of what's called training data available."

---

Designing Novel Proteins with AI

Beyond predicting existing protein structures, AI is enabling scientists to design entirely new proteins with functionalities not found in nature, addressing contemporary challenges such as disease and climate change.

David Baker (12:42): "We can now create really new proteins that solve these problems that weren't really relevant during evolution to make the world a better place."

David Baker, a biochemist and director at the Institute for Protein Design, discusses how his lab leverages AI models like RosettaFold to engineer proteins with specific purposes, including:

• New Protein Antibodies: Enhancing the immune response against infections like influenza.
• Switch Proteins: Serving as environmental sensors.
• Carbon-Storing Proteins: Aiding in carbon sequestration to combat climate change.

Burleigh McCoy (13:31): "They've made proteins that could help store carbon, which is a huge hurdle for fighting climate change."

---

Conclusion: The Future of AI in Protein Science

AI models such as AlphaFold2 and AlphaFold3 have fundamentally transformed protein science, making structure prediction more accessible and enabling the design of novel proteins with tailored functionalities. While challenges remain, particularly in handling proteins with complex or flexible structures, the ongoing advancements promise to unlock new frontiers in biology, medicine, and environmental science.

Burleigh McCoy (14:08): "Predictive and generative AI models have fundamentally changed the protein science landscape."

Emily Kwong wraps up the episode by acknowledging the profound impact of AI on understanding the "little things in life," emphasizing the potential for future breakthroughs.

Emily Kwong (14:35): "Thank you so much for bringing us this big, big story about the little things in life."

---

This episode of Short Wave offers an insightful exploration into how AI is not only solving age-old scientific puzzles but also paving the way for innovations that could address some of humanity’s most pressing issues.

Transcript

[Sponsor Voice]

This message is from NPR sponsor Spectrum Business. From one employee to thousands, Spectrum delivers connectivity solutions at a great value for any size business. And it's all backed by the Spectrum commitment. Learn [e-mail] business.

[Emily Kwong]

You're listening to Shortwave from NPR. Hey, hey, Short wavers. Emily Kwong here with producer Burleigh McCoy. What's up early?

[Burleigh McCoy]

Hey, Emily. Hello.

[Emily Kwong]

What do you have for us today?

[Burleigh McCoy]

Okay, so, Emily, today I want to dig into how AI has shaken up the field of protein science. As in the fundamental building blocks of life proteins.

[Emily Kwong]

I've heard of them, yeah. I mean, this is like what you studied back in your scientist days.

[Burleigh McCoy]

Yes, yes. I love proteins.

[Emily Kwong]

Oh, we love that you love them. How has AI moved the needle in this field, though?

[Burleigh McCoy]

Well, scientists have used it to dig into a problem that protein scientists have struggled with for more than 60 years. And that is, what do these building blocks, of which there are millions, look like?

[Emily Kwong]

Like their shape.

[Burleigh McCoy]

Like their shape. Yeah, exactly.

[Emily Kwong]

And why is that so important?

[Burleigh McCoy]

Well, the ability of a protein to do its specific job, so, like, carry oxygen through your body or turn light into sugar, that relies wholly on its unique, complicated shape. So to understand how it works, you need to know its shape.

[Emily Kwong]

But why can't scientists just run an experiment to determine the shape?

[Burleigh McCoy]

They can for some proteins. But those experiments can take years and years. And, Emily, that's because a scientist essentially needs to take the equivalent of a molecular photo of the protein to map its complicated shape. But getting the protein to cooperate to get that photo. So, like, to hold still, for example, without falling apart, that can be super tricky. And it could take a grad student's entire PhD program to figure out a single protein. And other proteins were just abandoned because they would never cooperate.

[Emily Kwong]

Proteins sound difficult, honestly. So the challenge is, how do you figure out a protein's shape without running these super tedious experiments? Is this where AI comes in?

[Sponsor Voice]

Yeah.

[Burleigh McCoy]

And to give you a sense of kind of how AI has changed the protein game, there's this protein competition that scientists run every other year.

[Emily Kwong]

Get out. A protein competition.

[Sponsor Voice]

Okay.

[Burleigh McCoy]

Yeah. And they've run it for the past 30 years, where groups will basically compete on who can accurately guess the most protein shapes. It's like nerd central. For sure.

[Emily Kwong]

We love.

[Burleigh McCoy]

And for most of that 30 year history, participants have really only made incremental progress. But in 2020, Google DeepMind used AlphaFold2, that's its AI protein prediction model. And Emily, AlphaFold2 blew the other competition out of the water completely.

[Emily Kwong]

Wow.

[Burleigh McCoy]

Okay, game changer. And now the Google DeepMind team has taken this AI tool to the next level by expanding it beyond proteins.

[Emily Kwong]

So today on the show, how scientists have taken a huge step to understanding the building blocks of life using AI.

[Burleigh McCoy]

Plus how other researchers are using the tech to design brand new proteins, ones never before seen in nature.

[Emily Kwong]

And how AI could help us solve the biggest problems we face today, from disease to climate. You are listening to Shortwave, the science podcast from Nvidia pr.

[Sponsor Voice]

This message comes from the Hartford every day. They make business insurance easy or easier anyway. And when it comes to business insurance for midsize and large businesses, they know just how complex it can be. And businesses rely on them to make the complex challenges less so. That's why at each step, from quotes to claims to everything in between, the Hartford is devoted to making it easier. Connect with your underwriter at the hartford or visit thehartford.com this message comes from Thuma. Create your oasis with Thuma, a modern design company that specializes in furniture and home goods by stripping away everything but the essential. Thuma makes elevated beds with premium materials and intentional details with clean lines, subtle curves, and minimalist style. The Thuma bed collection is available in four signature finishes to match any design aesthetic. To get $100 towards your first bed purchase, go to Thuma Co, NPR. This message comes from Insperity, providing HR services and technology, from payroll benefits and HR compliance to talent development. Learn more at Inspiration.

[Emily Kwong]

Okay, Burley. So scientists, it seems, have been trying to figure out the complicated shapes of proteins for decades to better understand how they work. Why has this been such a complicated thing to figure out?

[Burleigh McCoy]

Well, the short answer, Emily, is that there are so many theoretical ways a single protein could fold that it's a big problem to solve. So if you unfolded a protein, it would kind of look like a bunch of beads on a long string. And those beads are little molecules called amino acids.

[Emily Kwong]

Oh, I remember this from biology. There are like 20 types of amino acids.

[Burleigh McCoy]

Yep.

[Emily Kwong]

Each one is a little different.

[Burleigh McCoy]

Right. So each one has a slightly different shape, and that kind of dictates how that part of the string can be folded up. Because proteins often have 100 or more amino acids. You can see how imagining all the ways it could fold would get complicated.

[Emily Kwong]

Yeah, it just sounds like thousands of different shapes are, what, hundreds of thousands of different shapes.

[Burleigh McCoy]

Okay, try billions of trillions, Emily. Like there are theoretically more ways for one single protein to fold than there are stars in our night sky.

[Emily Kwong]

This sounds like a glorious nightmare.

[Burleigh McCoy]

Right?

[Emily Kwong]

I'm so curious. Okay, so you Said that AI has helped us make some leaps and bounds towards a solution. How does this technology work?

[Burleigh McCoy]

So this AlphaFold model is a type of AI called a deep learning program, which is this huge network of data processing points called nodes. And the purpose of this network is to learn and then make predictions based on what it's learned. In AlphaFold's case, and other models like it, it learns about proteins from a huge collection of protein structures that scientists have been building on for decades from their experimental data.

[Emily Kwong]

Okay. So the idea is that after these models use all of that carefully gathered experimental data to learn, they can then predict the shapes of proteins they do not know yet.

[Burleigh McCoy]

Exactly.

[Emily Kwong]

Okay. And going back to the protein competition in 2020, how did AlphaFold blow away the competition?

[Burleigh McCoy]

So they essentially changed the whole architecture of their model. They had been using AI before, but remember the beads on a string analogy? If amino acids are the beads, even if one bead is far from another on the string, when it all folds up, they could be right next to each other. So with AlphaFold 2, the model looked at distances between all the different amino acids and previous knowledge from solved protein structures.

[Emily Kwong]

Awesome.

[Burleigh McCoy]

And the accuracy and speed of the predictions went way up.

[Emily Kwong]

Okay, I'm assuming that made a huge difference for scientists everywhere studying proteins.

[Burleigh McCoy]

Totally. Julian Bergeron, a structural biologist at King's College London, is one of them. He studies the tail, like appendage that propels bacteria. So it's called a flagellum, and it's pretty complicated.

[Julian Bergeron]

It's this huge assembly, so it's longer than the bacterial cell itself. It consists of 20 to 25 different proteins, but many of them have hundreds of thousands of copies of that protein.

[Burleigh McCoy]

And these huge propeller machines are what give some bacteria the ability to make you sick or build plaque on your teeth. So Julian's lab is trying to figure out how these giant machines work, what their pieces look like, and how it all fits together. And so when the AlphaFold 2 model came out, he just had to try it.

[Julian Bergeron]

And I input a sequence, and then a few hours later, I had the model, and I was like, oh, my God, this just did it. And we'd been struggling with that problem for months, if not years, years. And all of a sudden, I messaged my lab and I said, we model everything. And we've had dozens of projects that immediately progressed thanks to this.

[Emily Kwong]

Okay, so it sounds like overnight alphafold changed the trajectory of his lab. Yeah, but how did he know that using AlphaFold2 would actually work?

[Burleigh McCoy]

Yeah, so the accuracy is super important. Right. Especially when you're basing all of your other experiments on the results. And it's important to note that like other AI, AlphaFold2 isn't right 100% of the time. So you can't just take the results at face value. But unlike some other AI, included in the results is a score, basically telling you how accurate each part of the structure is.

[Emily Kwong]

Okay. And are others in the field using alphafold too?

[Burleigh McCoy]

Yeah. So this is something that actually sets alphafold apart from other protein prediction AI models. It's extremely user friendly. So essentially anyone who works on a protein or even just has a sequence of a protein can plug it in and get results. I talked to Pushmeet Kohli, vice president of research at Google DeepMind, and he told me why it was important for them to make this tool open access.

[Pushmeet Kohli]

The mission statement that we have for the science program at Google DeepMind is to leverage AI to accelerate and advance science.

[Emily Kwong]

Okay, so I'm scrolling through the AlphaFold website and I'm seeing scientists using this model for all kinds of things. They're working on malaria and cancer research, drug discovery, plastic eating enzymes.

[Burleigh McCoy]

And last week DeepMind released a new version, AlphaFold 3, which can predict the 3D structure of proteins and other kinds of biomolecules that they attach to.

[Emily Kwong]

Why are those other biomolecules important?

[Burleigh McCoy]

Yes. So I know we talked about how much proteins are super important. I love them. But I have to admit, they rarely work alone. And if we actually want to know how biology works as a whole, we need to understand how proteins work with their partner molecules.

[Pushmeet Kohli]

So it really gives you a more detailed and more accurate picture of what is happening inside the body, where proteins are just not just sort of existing in isolation. They are interacting in a very rich biological space, or soup of RNA and DNA and small molecules. And it really sheds light into those rich interactions.

[Burleigh McCoy]

Now, previous versions of these protein prediction softwares would model where each amino acid was located, but in this new version, Alpha Fold 3, it maps things on an even smaller level. So it models where individual atoms are. Wow. So they can predict the structure of multiprotein complexes like the bacterial flagellum, or something like proteins in the blood which attach to iron atoms.

[Emily Kwong]

That is powerful. Okay, what are the limits to AlphaFold predictions?

[Burleigh McCoy]

Yeah, there are definitely limitations. Pushmeet says that the model works best when a protein has a single defined structure. But some proteins have more than one shape or they have sections that are kind of flimsy. Think cooked versus uncooked spaghetti.

[Emily Kwong]

Okay, so the model has Sounds like some trouble with prediction in some cases. And the results show that.

[Sponsor Voice]

Yeah.

[Burleigh McCoy]

So the idea is that these results would say, hey, I'm not so confident in this area of the protein, just so like, users know. And another limitation is that the prediction ability depends on the amount of what's called training data available. So I mentioned that there's a lot of training data for proteins, but some.

[Pushmeet Kohli]

Categories have much less training data available. For example, there's much less structural data available for RNAs.

[Emily Kwong]

Okay, so the prediction is only as good as the data.

[Burleigh McCoy]

Exactly, exactly. But, Emily. But, Burleigh, there's another way scientists can use AI in the protein world.

[Emily Kwong]

Okay, what's that?

[Burleigh McCoy]

To generate brand new proteins. Ones like, not found in nature anywhere.

[Sponsor Voice]

Huh.

[David Baker]

Humans face new problems today. And, you know, we live longer, we're polluting and heating up the planet. And it's reasonable to think that if with another million more millions of years of evolution, that some of these problems would be. Would be solved. But we don't want to wait that long. So the idea is that we can now create really new proteins that solve these problems that weren't really relevant during evolution to make the world a better place.

[Burleigh McCoy]

So this is David Baker. He's a biochemist and the director of the Institute for Protein Design at the University of Washington. And he's been working on proteins for years. He actually developed one of the earlier protein prediction models. His lab has a similar AI program to AlphaFold 3. It's called RosettaFold, All Atom. But his big focus is designing these brand new proteins.

[Emily Kwong]

This sounds so futuristic, right? Like, what kind of new proteins?

[Burleigh McCoy]

So far, they've done things like design new protein antibodies, which are important for fighting infections, in this case, to fight influenza. They've made something called a switch protein that could be used as an environmental sensor. And they've also made proteins that could help store carbon, which is a huge hurdle for fighting climate change.

[David Baker]

I think really across medicine, sustainability technology, I think there's huge opportunities to transform the current ways we do things with protein design.

[Burleigh McCoy]

So these predictive and generative AI models have fundamentally changed the protein science landscape. And again, there's definitely room for improving the prediction power. But with what the field has shifted to, like, in terms of prediction accuracy and design potential, I mean, it's really gotten this retired protein fanatic, like, missing my science days.

[Emily Kwong]

Burley, thank you so much for bringing us this big, big story about the little things in life.

[Burleigh McCoy]

Thanks, Em.

[Emily Kwong]

This episode was produced by Rachel Carlson. It was edited by our showrunner, Rebecca Ramirez Burleigh checked the facts. Ko Takasugi Chernovin was the audio engineer. Special thanks to Jeff Brumfield. Beth Donovan is our senior director and Colin Campbell is our senior vice president of podcasting strategy. I'm Emily Kwong. Thank you for listening to Short Wave from N.

[Sponsor Voice]

This message comes from Mint Mobile. Mint Mobile took what's wrong with wireless and made it right. They offer premium wireless plans for less and all plans include high speed data, unlimited talk and text and nationwide coverage. See for [e-mail] Switch this message comes from Capella University. At Capella you'll learn relevant in demand skills you can apply in the workplace right away. A different future is closer than you.

[Burleigh McCoy]

Think with Capella University.

[Sponsor Voice]

Learn [e-mail]. this message comes from NPR sponsor Viori featuring the performance jogger. Visit viori.com NPR for 20% off your first purchase on any US orders over $75 and free return exclusions apply. Visit the website for full terms and conditions.