Latent Labs aims to replace trial-and-error drug discovery with push-button design

Today Latent Labs launches Latent-X, a frontier AI model for push button protein design, which outperforms competing models under identical laboratory conditions. 

Latent Labs is working to transform the expensive, labour-intensive, and high-failure-rate processes of drug discovery into automated drug design. The company emerged from stealth in February with $50 million in funding to accelerate its progress in programmable biology.

The team consists of former AlphaFold 2 co-developers, ex-DeepMind team leads, and brings rich experience from Microsoft, Apple, Stability AI, Exscientia, Mammoth Bio, Altos Labs and Zymergen.

Traditional drug discovery requires screening upwards of tens of thousands of random molecules—a process where hit rates are typically well below 1 per cent and each experiment takes months and costs thousands of dollars.

Latent-X is a general-purpose frontier model that creates binders from scratch for unseen or previously untargeted proteins, solving the geometric puzzle of binding at the all-atom level. The model generates designs over 10x faster than previous methods and co-samples sequence and structure simultaneously, allowing for close to real-time computational experimentation. 

Latent-X generalises beyond nature's repertoire by generating all-atom binder structures that obey atomic-level biochemical rules, opening doors to other therapeutic modalities that depend on target-specific binding—nanobodies and antibodies being prime examples. 

Specifically, Latent-X generates functional, high-affinity de novo binders with breakthrough laboratory performance. In extensive wet lab experiments across seven therapeutic targets, Latent-X achieved hit rates of 91-100 per cent for macrocycles and 10-63 per cent for mini-binders.

With Latent-X, drug designers can generate high-confidence binders with the push of a button, achieving what would typically require testing tens of thousands of candidates by testing as few as 30 candidates per target.

The model is available for early access on Latent's no-code AI protein design platform, where users can upload protein targets and generate cyclic peptides and mini-binders directly in the browser.

Through the platform, users can generate, explore, and score binder designs, selecting top-ranked structures for further lab testing. The platform includes a free tier for both commercial and non-commercial users.

"We envision a future where effective therapeutics can be designed entirely in a computer, like semiconductors or space missions," said Simon Kohl, CEO and founder of Latent Labs. 

"Our platform empowers scientists with lab-validated protein binder design at their fingertips, whether they're experts or new to AI-powered drug design, and without needing AI infrastructure. This is the first step on our mission toward making biology programmable in order to make drug design instantaneous."

The company is now open to partnerships to bring these expanded capabilities to new drug applications.

Lead image: Latent Labs. Photo: uncredited.