Latent Labs Launches Web-Based AI for No-Code Protein Design

Latent Labs, a cutting-edge artificial intelligence (AI) laboratory, has emerged from stealth mode with a groundbreaking web-based AI model for programming biology. Six months after securing $50 million in funding, the company has launched Latent-X, a powerful tool that streamlines the design of novel proteins, marking a significant leap forward in drug discovery. Simon Kohl, Latent Labs' CEO and founder, previously co-led DeepMind’s AlphaFold’s protein design team. He emphasized that Latent-X has achieved state-of-the-art (SOTA) performance across various metrics when tested in a physical lab. Unlike AlphaFold, which excels in predicting the structure of existing proteins, Latent-X focuses on creating entirely new molecules, including nanobodies and antibodies, with precise atomic structures. This capability significantly accelerates the development of new therapeutics, offering a more efficient alternative to traditional methods that involve screening millions of random molecules, a process characterized by low hit rates, long timelines, and substantial costs. The Latent Labs platform, accessible via platform.latentlabs.com, is designed to be user-friendly and does not require advanced AI infrastructure. Users can upload protein targets and generate cyclic peptides and mini-binders directly in their browser. The platform supports target upload, hotspot selection, binder design, and computational ranking, making it a comprehensive solution for protein design. Key features include structure visualization, predicted structure overlays, and computational metric rankings, which allow scientists to replicate the AI workflows used to produce validated binders. In extensive wet lab experiments across seven therapeutic targets, Latent-X demonstrated remarkable performance. It achieved 91-100% hit rates for macrocycles and 10-64% hit rates for mini-binders. These binders showed strong target specificity, with picomolar binding affinities for mini-binders and single-digit micromolar affinities for macrocycles. Compared to existing generative models, Latent-X outperformed in both in silico and laboratory validations. The success of Latent-X is particularly notable for its ability to create high-affinity de novo binders, which are crucial for developing new therapeutic agents. The platform's approach is revolutionary, as it transforms a traditionally labor-intensive process into an automated, rapid design workflow. Traditional drug discovery methods often require months and high costs for each experiment, with hit rates typically below 1%. Latent-X, however, can generate high-confidence binders with the push of a button, reducing the number of candidates needed for lab testing from millions to just 30 per target. This efficiency not only saves time and resources but also increases the likelihood of discovering effective drugs. One of the key advantages of Latent-X is its flexibility and general-purpose nature. It can design binders for unseen or previously untargeted proteins, solving complex geometric binding puzzles at the atomic level. The model generates designs up to 10 times faster than previous methods and can co-sample sequence and structure simultaneously, enabling real-time experimentation. By generating all-atom binder structures that adhere to biological rules, Latent-X opens new avenues for therapeutic modalities such as nanobodies and antibodies, which are highly sought after due to their versatility and potential for oral delivery and tissue permeability. The significance of Latent-X extends beyond its technical achievements. It democratizes access to AI-driven drug design, empowering a wider range of organizations, from academic institutions to biotech startups and pharmaceutical companies. Kohl highlighted that not all companies possess the resources to build their own AI models, infrastructure, or teams, making Latent-X a valuable tool. While currently offered for free, the company plans to introduce advanced features and capabilities for which it will charge in the future. Latent Labs is backed by prominent investors, including Radical Ventures, Sofinnova Partners, Google’s Chief Scientist Jeff Dean, Anthropic’s CEO Dario Amodei, and Eleven Labs CEO Mati Staniszewski. The team comprises experienced individuals from leading tech and biotech companies, such as former AlphaFold 2 co-developers, ex-DeepMind leaders, and professionals from Microsoft, Apple, Stability AI, Exscientia, Mammoth Bio, Altos Labs, and Zymergen. This diverse expertise positions Latent Labs to drive innovation in the field. Industry insiders hail Latent-X as a game changer in drug discovery. The ability to design effective therapeutics entirely in a computer, akin to designing space missions or semiconductors, is seen as a transformative step. By removing barriers to AI infrastructure and expertise, Latent Labs is poised to accelerate the development of new drugs and treatments, potentially addressing critical health issues more rapidly and cost-effectively. The company's strategic focus on partnerships underscores its commitment to expanding the impact of Latent-X across the biomedical landscape. In summary, Latent Labs' launch of Latent-X marks a pivotal moment in AI-driven drug discovery. Its combination of technological superiority, user-friendly platform, and broad accessibility could redefine how novel proteins are designed and how quickly new therapeutics reach the market. As Latent-X continues to evolve and gain widespread adoption, it holds the promise of making drug design more instantaneous and programmable, aligning with the company’s vision of a future where biology is as manageable as other advanced fields.