Biotech and AI startup Cradle has seen significant success with its generative approach to protein design, securing major customers and a substantial $24 million in new investment.
The company emerged from stealth mode just over a year ago, amidst increasing interest in large language models. While many AI companies in biotech focus on training models to understand molecular structure, Cradle took a different approach, likening the long sequences of amino acids in proteins to an “alien programming language.”
While it may be challenging for a person to comprehend this language, an AI model could, and individuals could collaborate with it. Although they couldn’t simply request “create a protein that does this,” they could inquire about which of 100 intriguing proteins appears most likely to endure specific conditions.
This unique approach has attracted the attention of major drug development companies such as Johnson & Johnson and Novozymes. Creating a functional protein from scratch typically involves a lengthy and labor-intensive process, often requiring years and numerous wet-lab experiments.
Cradle claims that its technology can significantly reduce the time and number of experiments needed. While it hasn’t substantiated its claims about halving development time, it did provide an example from its in-house development to illustrate its capabilities.
They utilized their software to generate alternative versions of T7 RNA polymerase, an RNA production enzyme, with enhanced heat resistance. Normally, a team might expect less than 5 percent of intentionally modified molecules to exhibit the desired trait, but 70 percent of the variants produced by Cradle displayed increased stability, achieving the equivalent of four or five experimental runs in one.
In addition to T7, Cradle is internally developing “a dehalogenase for soil decontamination, a growth factor for cell division in cultured meat products, a transaminase for metabolic regulation, and an antibody therapeutic,” stated Cradle CEO and co-founder Stef van Grieken in an email to TechCrunch.
Such notable advancements are feasible, and even small improvements are valued by companies investing heavily in these processes. However, the drug development process encompasses more than just generating potential candidate molecules.
“We have already demonstrated the potential of our platform to expedite the R&D phase and assist our partners in bringing bio-based products to market more quickly and cost-effectively,” said van Grieken.
The technology is not limited to drug development and can also be applied in food and industrial sectors. Like other similar tools, one appealing aspect for customers is that Cradle can be used directly by scientists and labs without requiring a machine learning engineer.
Van Grieken shared his thoughts on establishing a biotech company in the EU, considering that many team members had previously worked at major tech firms in Silicon Valley.
“We have found that building in the EU has its pros and cons,” he said. While fundraising for a deep-tech venture in Europe is more complex than in the US, Europe offers a rich talent pool and proximity to major pharma and biotech companies.
Cradle’s $24 million A round follows a $5.5 million seed round last year. Index Ventures led the round, with Kindred Capital and individual investors Chris Gibson and Tom Glocer, among others, participating. The company plans to utilize the capital to expand its team and sales efforts.