There is a new company on the block whose mission is to make programming biology easy. The European startup called Cradle is emerging from stealth after building out their platform. It has just announced a €5.5 million ($5.4m) seed funding round led by Index Ventures, Kindred Capital, and angel investors including John Zimmer, co-founder and president of Lyft
Do not be deceived by the company’s seemingly narrow focus on proteins. They are not just something we eat – although engineering animal-free meat, egg, and dairy products is indeed a big focus of synthetic biology. Proteins are also versatile biological machines that underpin every almost function in living cells and have equally as many uses outside of biology. Think of the enzymes used in detergents, cosmetics, and textiles; or antibodies that make powerful therapeutics; or, in fact, any other area of biotechnology where proteins catalyze reactions to make products like bulk and specialty chemicals, flavors and fragrances, biofuels, materials, and more. There are countless uses for these biomolecules, and Cradle wants to enable even more applications with the ability to design custom proteins that perform versatile tasks.
Cradle’s co-founder and CEO Stef van Grieken is a self-described “purveyor of fine protein”. He spent the last decade working at Google AI leading the development of several machine learning applications, as well as at X, Google’s “moonshot factory”, assessing the feasibility of early-stage projects. During his tenure at Google
Designer proteins are a multibillion dollar industry: the market is projected to reach $3.9 billion by 2024, driven in large part by protein-based therapeutics. But it could be even bigger: there is a huge potential for branching out into other areas of synthetic biology, if only designing custom proteins was not so difficult. The way protein engineering is currently done is through trial and error in the lab, and the typical success rate of hitting the design specifications is less than 1%. To increase the chances of success, biologists can use software tools like Rosetta or AlphaFold to predict protein structure based on its sequence. Proteins start out as just strings of amino acids that fold into 3D shapes like origami. But predicting the folding pattern is an incredibly complex problem, and a program like Rosetta requires years of training and thousands of computers to run.
Cradle is approaching the problem differently: they are using a generative model to “reverse engineer” proteins. You may have heard of or even used generative models like DALL-E that can create new images based on a descriptive input. Cradle’s founders thought to apply the same principle to design new protein architectures. Instead of using sequence-structure models, they use machine learning algorithms trained on real data. The user can specify what kind of protein they want to design, and the platform will provide a list of possible sequences that can create that structure. And the best part is – you do not have to be a machine learning expert to use it:
“Cradle’s self-teaching, self-improving generative machine learning models draw on recent advances in ‘natural language processing’ to predict which parts of a protein’s genetic code a biologist will need to alter, significantly improving a scientist’s chances achieving positive experimental results without the need for a machine-learning background”, said the CEO in a press release. “Through this method, Cradle believes it can reduce the time and cost of getting a synthetic biology product to market by an order of magnitude.”
Today, most biotech and synthetic biology companies are left to their own devices when it comes to engineering proteins. Major players in the protein engineering field include Thermo Fischer, Danaher, Agilent Technologies
Cradle itself is neither a synthetic biology nor a machine learning company – they are both. “We didn’t want to be just a machine learning company; you really have to understand the biology as well,” said Stef. With expertise in machine learning technology and superior lab skills that their team members brought from companies like Google, IBM
To match the Cradle team’s diverse backgrounds, the company has attracted investors from different areas of tech, including the founder of the DNA synthesis company Twist Bioscience Emily Leproust and Lyft’s president John Zimmer. The interest from the ride share company may be surprising at first; but a lot of the advancements in machine learning have come from other areas of tech. The company’s co-founder Jelle Prins himself came from Uber
And that is what happens when different areas of deep tech collide: a galaxy of new possibilities is born. Stef envisions his company empowering synthetic biology innovations in the chemicals and ingredients space, materials science and engineering, and other areas: “Hopefully we’re going to be a catalyzer for many more companies to be built because the cost of getting [products] to market should go down. If you can build a bio-based product with a team of 15 people in a couple of years and just a few million dollars, that would be a success.”
Cradle’s software is already being used by several companies, and they want to distribute it widely as possible. This is why the platform is free to use for academics. Cradle is also offering friendly IP terms, where the users do not have to pay royalties on any products developed using the platform, as well as complete privacy and security to protect trade secrets. “We want to make it available to everybody to democratize protein engineering,” is the vision of Cradle’s CEO. Stef will be speaking at the SynBioBeta conference next year, the place where synthetic biology leaders and visionaries gather to bring about a more sustainable future. Let’s see what kind of new ideas Cradle’s technology will inspire.
Thank you to Katia Tarasava for additional research and reporting on this article. I’m the founder of SynBioBeta, and some of the companies that I write about, including Twist Bioscience, are sponsors of the SynBioBeta conference and weekly digest.