Isomorphic Labs is a biotechnology company that applies Artificial intelligence and Machine learning to drug discovery and development. Founded in 2021 as a spin-out from Google DeepMind, the company aims to use computational models to understand biological systems at a molecular level, thereby reducing the time and cost associated with traditional pharmaceutical research. Headquartered in London, United Kingdom, Isomorphic Labs operates independently while maintaining close ties to its parent organization, Alphabet Inc.
The company's name derives from the mathematical concept of isomorphism, which describes a structure-preserving mapping between two sets. In the context of drug discovery, this reflects the ambition to create a digital representation of biological processes that faithfully mirrors their real-world behavior. By building predictive models of molecular interactions, Isomorphic Labs seeks to identify promising drug candidates and optimize their properties before physical synthesis and testing.
Background and Founding
Isomorphic Labs was announced in November 2021 by Demis Hassabis, co-founder and CEO of Google DeepMind. The venture was established to leverage breakthroughs in AI, particularly the AlphaFold protein structure prediction system, for therapeutic applications. AlphaFold, which had demonstrated remarkable accuracy in predicting protein 3D structures from amino acid sequences, provided a foundational tool for the new company. Hassabis stated that the goal was to reimagine the entire drug discovery process through an AI-first approach.
The company was incorporated as a separate legal entity under Alphabet, with initial funding and resources provided by the parent company. Its formation followed years of research at Google DeepMind on applying neural networks to biological problems, including work on protein folding and molecular dynamics. The spin-out allowed the team to focus exclusively on commercial drug discovery while maintaining access to DeepMind's ongoing research.
Technology and Approach
Isomorphic Labs employs a range of Deep learning techniques, including neural networks and transformer models, to model molecular structures and interactions. The core platform integrates data from public and proprietary sources, including genomic sequences, protein structures, and chemical libraries. The company's models are trained to predict binding affinities, toxicity, and other key properties of drug candidates.
A central component is the use of residual networks and U-Net architectures for processing 3D molecular data. These models can capture spatial relationships between atoms and residues, enabling accurate predictions of how small molecules interact with protein targets. The company also utilizes generative AI to propose novel molecular structures that meet desired criteria, such as high potency and low off-target effects.
Isomorphic Labs emphasizes a closed-loop approach, where computational predictions are validated through experimental assays, and the resulting data is fed back to refine the models. This iterative cycle aims to improve accuracy over time and reduce reliance on trial-and-error in the lab. The company collaborates with contract research organizations and academic institutions to conduct wet-lab experiments.
Partnerships and Collaborations
In January 2024, Isomorphic Labs announced a strategic partnership with Eli Lilly and Company to discover small molecule therapies for multiple targets. The deal included an upfront payment and potential milestone payments totaling over $1 billion, reflecting the high expectations for AI-driven drug discovery. Under the agreement, Isomorphic Labs would use its platform to identify lead compounds, which Eli Lilly would then develop and commercialize.
In the same month, the company entered into a collaboration with Novartis to develop small molecule drugs for three additional targets. This agreement also involved substantial upfront and milestone payments. These partnerships marked Isomorphic Labs' transition from a research-focused entity to a commercial drug discovery organization, providing validation for its technology and business model.
The company has also maintained ongoing collaborations with Google DeepMind, particularly around the development of AlphaFold and its successor, AlphaFold 2. These tools are used internally and shared with the broader scientific community through open-source releases and publications.
Leadership and Team
Demis Hassabis serves as the founder and chief executive officer of Isomorphic Labs, while also retaining his role at Google DeepMind. The company's leadership team includes experts in computational biology, medicinal chemistry, and AI research. Notable hires include John Jumper, who led the AlphaFold project and joined Isomorphic Labs as a research scientist, and Pushmeet Kohli, who became head of AI for science.
The team comprises researchers from diverse backgrounds, including former academics from institutions such as University of Cambridge and University of Oxford, as well as industry veterans from pharmaceutical companies. The company has grown to over 100 employees as of 2024, with plans for further expansion.
Impact and Future Directions
Isomorphic Labs represents a significant bet on the potential of AI to transform drug discovery, an industry traditionally characterized by high failure rates and long development timelines. The company's approach, if successful, could lead to faster identification of novel therapeutics for diseases with unmet medical needs, including cancer, neurological disorders, and rare genetic conditions.
However, the field of AI-driven drug discovery is highly competitive, with numerous startups and established companies pursuing similar goals. Isomorphic Labs differentiates itself through its access to DeepMind's research infrastructure and the AlphaFold technology, which provides a unique advantage in structure-based drug design. The company's partnerships with major pharmaceutical firms provide financial stability and validation.
As of 2024, Isomorphic Labs has not yet advanced any drug candidates into clinical trials, but its pipeline is expected to mature in the coming years. The company continues to publish research and develop new methods, contributing to the broader scientific understanding of molecular interactions. Its long-term vision is to create a comprehensive computational model of human biology, enabling the design of personalized therapies tailored to individual patients.
See Also
- Google DeepMind
- Artificial intelligence
- Machine learning
- Drug discovery
References
- Isomorphic Labs official website (accessed 2024).
- Alphabet Inc. press releases and investor communications.
- Hassabis, D. (2021). Announcement of Isomorphic Labs.
- Jumper, J. et al. (2021). Highly accurate protein structure prediction with AlphaFold. Nature.
- Eli Lilly and Company press release, January 2024.
- Novartis press release, January 2024.