Can IsomorphicLabs Revolutionize Drug Discovery with AlphaFold and AI Partnerships?

On January 8th, Alphabet's AI pharmaceutical arm, IsomorphicLabs, made headlines by announcing two groundbreaking drug-discovery deals, worth a staggering $3 billion, with industry giants Eli Lilly and Novartis. These partnerships mark a significant milestone for the company, which was spun off from DeepMind, and signify its entry into a new phase of pharmaceutical innovation.

The collaboration with Eli Lilly involves the discovery of therapeutics targeting multiple disease-related proteins and pathways. IsomorphicLabs will receive a $45 million down payment from Eli Lilly, with the potential for $1.7 billion in milestone-based payments. This partnership goes beyond mere pilot projects or academic collaborations, with the primary goal being the creation of actual drugs.

Similarly, the collaboration with Novartis focuses on the discovery of small-molecule therapies against three undisclosed targets. Novartis will make an upfront payment of $37.5 million, with a total milestone potential of up to $1.2 billion. Novartis, known for its partnerships with other tech giants like Palantir and Microsoft, sees great potential in working with IsomorphicLabs to tackle the challenges of targeting proteins that are structurally unresolved or difficult to address with traditional chemistry methods.

IsomorphicLabs leverages its advanced AI technology, particularly its AlphaFold model, to drive drug discovery efforts. This model utilizes high-performance computing to virtually screen compounds and identify molecules that can effectively target disease-associated proteins. While conventional screening methods typically cover around 1 million compounds, virtual screening with AlphaFold can cover a staggering 20 billion compounds. Moreover, IsomorphicLabs explores de novo design, enabling the creation of molecules from scratch that precisely match protein-binding pockets.

The collaborations with Eli Lilly and Novartis also serve another crucial purpose for IsomorphicLabs. The company aims to adapt its AI models and determine the extent of fine-tuning required to target specific drug targets effectively. By working closely with these pharmaceutical partners, IsomorphicLabs seeks to understand whether a more general-purpose AI model or a specialized model tailored to specific data or programs is the optimal approach. Ultimately, IsomorphicLabs aims to strike a balance that maximizes scalability and enhances the efficiency of its drug discovery process.

The potential of IsomorphicLabs extends beyond predicting static protein structures, as it aims to link structure and function to advance drug design. While protein structure prediction is valuable, predicting how small molecules bind to specific proteins holds even greater significance. IsomorphicLabs is also developing additional models, currently undisclosed, to predict properties such as pharmacokinetics and potential toxicity of compounds. These models, combined with the power of AlphaFold, hold immense promise for accelerating biomedical breakthroughs and enabling "digital biology."

Looking ahead to 2024, IsomorphicLabs has set ambitious goals. It aims to continuously improve the accuracy of its prediction system, expand its internal pipeline, and advance its own drug program. While details about IsomorphicLabs' specific drug targets remain undisclosed, the company has narrowed down a list of around 20 targets and is utilizing contract research organizations (CROs) to make and test potential compounds. IsomorphicLabs acknowledges that it is still in the early stages of testing its platform's predictions and is focused on ensuring the reliability of its AI models before pushing forward with internal projects.

IsomorphicLabs, born out of DeepMind, has quickly established itself as a leader in the field of AI-driven drug discovery. The success of its AlphaFold model, which made accurate predictions of protein structures, has garnered significant attention and recognition. The establishment of IsomorphicLabs and its partnerships with industry giants like Eli Lilly and Novartis have further solidified its position as a frontrunner in the intersection of AI and pharmaceuticals.

With a talented team of experts, including Chief Scientific Officer Miles Congreve and Chief Artificial Intelligence Officer Max Jaderberg, IsomorphicLabs continues to attract top talent and expand its operations. The company currently employs around 90 individuals across its London headquarters and Lausanne, Switzerland offices. Notably, it has also secured the expertise of renowned scientists like Dr. Jennifer Doudna, a Nobel Prize winner in Chemistry, and Dr. David MacMillan, a recent Nobel laureate.

As IsomorphicLabs forges ahead, it aims to overcome the challenges facing the AI pharmaceutical industry, both technically and commercially. CEO Demis Hassabis envisions that with additional breakthroughs akin to AlphaFold, the time and cost required for drug development can be reduced significantly, while the success rate can be enhanced. IsomorphicLabs strives to revolutionize drug discovery, leveraging the power of AI and its partnerships to find treatments for some of the most devastating diseases humanity faces.

In a world where AI and pharmaceuticals intersect, IsomorphicLabs stands at the forefront, poised to reshape the landscape of drug discovery and usher in a new era of possibilities.