Latent Labs Launches Latent-X: No-Code AI for Protein Design

Today, Latent Labs has launched a groundbreaking AI model called Latent-X, designed to streamline the complex process of protein design for drug discovery. The model is available for early access on Latent's no-code platform, enabling users to generate cyclic peptides and mini-binders directly in their browsers. Through this user-friendly interface, researchers can upload protein targets, explore binder designs, and select top-ranked structures for further lab testing. Importantly, the platform includes a free tier accessible to both commercial and non-commercial users, sign-ups for which are now open at platform.latentlabs.com. Latent-X stands out for its ability to produce high-confidence binders with unprecedented efficiency and accuracy. Traditional drug discovery involves screening millions of random molecules, a process with extremely low hit rates (typically below 1%) and high costs (each experiment taking months and costing thousands of dollars). In contrast, Latent-X can generate binders that would normally require testing millions of candidates by testing just 30 per target, significantly reducing the time and cost involved. The model's performance has been rigorously validated through extensive wet lab experiments conducted across seven therapeutic targets. These tests demonstrated hit rates ranging from 91-100% for macrocycles and 10-64% for mini-binders. The generated binders also showed picomolar binding affinities for mini-binders and single-digit micromolar affinities for macrocycles, along with strong target specificity. This level of performance surpasses that of existing generative tools, both in computational evaluations and laboratory validation. Macrocycles are particularly valuable in drug design due to their potential for oral delivery and tissue permeability, while maintaining high specificity. Mini-binders, on the other hand, offer versatility and high specificity, making them a flexible option for various therapeutic applications. The success of Latent-X in generating these types of binders opens new avenues for developing effective therapeutics more quickly and at lower costs. The Latent Labs Platform is an intuitive tool that integrates the latest advancements in AI protein design. Users can easily upload their target proteins, select hotspots, and generate binder designs, which are then computationally ranked for quality. The platform provides structure visualization, predicted structure overlays, and metric rankings, replicating the AI workflows used to create the lab-validated binders. This empowers researchers, whether they are AI experts or novices, to leverage cutting-edge technology without needing dedicated AI infrastructure or expertise. Latent-X is a general-purpose model capable of designing binders from scratch for any protein, including those that have never been targeted before. It solves the intricate geometric puzzle of binding at the all-atom level, generating designs over 10 times faster than previous methods. By co-sampling sequence and structure simultaneously, Latent-X allows for swift computational experimentation, producing novel designs that adhere to atomic-level biochemical rules. This capability extends to creating entirely new molecule designs, such as nanobodies and antibodies, which are crucial for developing therapies for various diseases. Simon Kohl, CEO and founder of Latent Labs, emphasizes the transformative potential of Latent-X. "We envision a future where effective therapeutics can be designed entirely in a computer, much like how space missions or semiconductors are designed today," Kohl stated. "Our platform empowers scientists with lab-validated protein binder design at their fingertips, eliminating the need for specialized AI infrastructure and teams." Kohl, a scientist who previously co-led DeepMind's AlphaFold 2 protein design team, highlights the distinction between Latent-X and AlphaFold. While AlphaFold is primarily a predictive model for visualizing existing protein structures, Latent-X goes a step further by generating new proteins with precise atomic structures. This makes Latent-X uniquely suited for accelerating the development of new therapeutics. Unlike other AI-driven drug discovery companies such as Xaira, Recursion, or Isomorphic Labs, which focus on proprietary medicine development, Latent Labs adopts a collaborative approach. They aim to license their model to a wide range of organizations, including academic institutions, biotech startups, and pharmaceutical companies. "Not every company has the resources to build their own AI models and infrastructure," Kohl explained. "By providing Latent-X, we democratize access to state-of-the-art AI for protein design." Initially, Latent-X is available for free to encourage widespread adoption and experimentation. However, the company plans to introduce paid features and capabilities as it expands. This business model aligns Latent Labs with other open-source AI foundational model providers like Chai Discovery and EvolutionaryScale, fostering innovation and collaboration across the industry. Latent Labs' success and vision are supported by a robust funding round of $50 million, co-led by Radical Ventures and Sofinnova Partners. Additional investors include prominent figures such as Google’s Chief Scientist Jeff Dean, Anthropic’s CEO Dario Amodei, and Eleven Labs CEO Mati Staniszewski. The team behind Latent Labs comprises experienced professionals, including former AlphaFold 2 co-developers and ex-DeepMind team leads, bringing a wealth of knowledge from tech giants like Microsoft, Apple, and Stability AI, as well as biotech companies like Exscientia, Mammoth Bio, Altos Labs, and Zymergen. Industry insiders have hailed Latent Labs and Latent-X as a significant advancement in the field of AI-driven drug discovery. The platform's ability to generate high-affinity binders with minimal laboratory testing is seen as a game-changer, potentially revolutionizing the drug development process. Companies lacking the resources to build and maintain their own AI models and infrastructure stand to benefit greatly from Latent-X, making this technology accessible to a broader array of researchers and developers. Latent Labs is poised to play a pivotal role in advancing the field towards programmable biology and instantaneous drug design.