Biotech and AI startups cradle found success with its generative approach to protein design, landing major customers and a whopping $24 million in new investment.
The company left stealth more than a year ago, just as the hype of major language models was heating up. Many AI companies in biotech train models to understand molecular structure; Cradle’s insight is that the long sequence of amino acids that make up the proteins in our bodies is “like a foreign programming language.”
It may not be possible for a human to learn that language, but an AI model can – and a human can work with that. While they still can’t say “make a protein that can do this,” they can ask which of 100 interesting proteins are most likely to survive at room temperature or in an acidic environment.
The approach seems to have caught the eye of major drug development companies such as Johnson & Johnson and Novozymes. Creating a useful and functional protein from scratch is often a fairly involved process, perhaps years and hundreds or thousands of experiments in the wet lab.
Cradle says its technology can cut that time and the number of experiments needed to be reduced. Although this does not actually prove the claims of half the development time, it provides an illustrative example from the development within it.
They used their software to create alternative versions of T7 RNA polymerase, an RNA production enzyme, that are more resistant to high temperatures. Typically, they say, a team might expect less than 5 percent of the intended tweak molecules to have the desired aspect, but 70 percent of the variants produced in the Cradle show increased stability. That’s the equivalent of running four or five such experimental runs in one.
In addition to T7, Cradle works inside “a dehalogenase that can be used to decontaminate the soil, a growth factor that promotes growth through cell division that is often used in cultured meat products, a transaminases that regulate metabolic pathways and also help to understand certain diseases. as an antibody therapeutic,” said Cradle CEO and co-founder Stef van Grieken in an email to TechCrunch. “We benchmarked our models against an in-house engineered protein using existing tools and saw significant improvements in Generative AI-based designs.”
(This is it, if you’re wondering 🙂
Such large improvements are possible, and small, even fractional improvements are acceptable to companies that invest millions in these processes. But of course there is more to the drug development process than the generation of likely candidate molecules.
“We have already been able to demonstrate the potential of our platform to accelerate the R&D phase and help our partners bring bio-based products to market faster and more cost-effectively,” said van Grieken. “In fact, as we ourselves and many colleagues have now completed several phases of experimentation with our platform, we have seen models that generalize very well to different types of proteins and tasks, which are more exciting.”
The technology is not limited to drug development and can be used in food and industrial applications as well. Like other devices of this type, part of the draw for customers is that the Cradle does not require a machine learning engineer to operate, but can be placed directly in the hands of scientists and engineers. lab.
I asked van Grieken his thoughts on building an EU-based biotech company, with many of the team having previously worked at large Silicon Valley technology companies.
“We found that setting up in the EU has advantages and disadvantages. Fundraising for a deep tech business in Europe is more complicated in Europe than in the US, where there are many modern ‘tech-bio’ investors that companies like Cradle are interested in. There’s also a much larger community of like-minded founders in the Bay Area,” he said.
“However, from a talent point of view I think Europe is underrated,” van Grieken continued. “For example, here in Zurich, you have all the big tech companies (Apple, Google, Facebook) represented by thousands of engineers. You have an amazing talent pool coming from ETH and EPFL, which are some of the best universities for computer science and molecular biology in the world. And the competition for talent is certainly less intense than in the Bay Area. Finally, many of the world’s largest pharma and biotech companies are located in Europe, so we are close to our customers. I am sure that the European ecosystem is developing rapidly.”
Cradle’s $24 million A round follows a $5.5 million seed round last year. Previous investor Index Ventures led the round, with Kindred Capital (also a seed investor) participating, along with individual investors Chris Gibson, Tom Glocer, and others. The company says it will use the capital to grow its team and sales, as you do.