Bristol Myers Squibb and Takeda Collaborate to Accelerate AI-Based Drug Discovery

A key development set to change the pharmaceutical research landscape occurred this morning, with Bristol Myers Squibb (BMS) and Takeda Pharmaceuticals announcing plans to share their respective proprietary data to facilitate AI-based drug discovery. This partnership, announced on October 1, 2025, aims to leverage the power of AI to accelerate the discovery of new drug molecules, potentially transforming the way drugs are developed.

A New Era of Collaboration

The alliance brings together:

• BMS
• Takeda Pharmaceuticals
• Astex Pharmaceuticals
• Industry leaders AbbVie and Johnson & Johnson

Main Goal:
To train an advanced AI model, OpenFold3, by sharing thousands of experimentally determined protein-small molecule complexes. These datasets will help train the model to predict protein-small molecule interactions, a crucial step in understanding drug discovery.

This program is part of the broader AI Structural Biology Network, an industry-led consortium collaborating closely with the AlQuraishi Lab at Columbia University.

“Preventing global crises like pandemics requires a new type of network that benefits from AI to understand protein structures and detailed biological models, which will help develop drugs against diseases such as cancer,” said Greig.

Federated Data Sharing: A Secure Way to Share Big Data

One of the standout features of this collaboration is the federated data-sharing model, managed by Germany-based Apheris.

Key Highlights:

• Companies contribute generated data without sharing sensitive information.
• Instead of exchanging raw data, companies share model updates, keeping proprietary data secure and private.
• This model addresses the critical need for data privacy in collaborative analytics.
• Encourages a more open research environment, fostering cooperation between institutions and research groups.

Enabling AI in Drug Discovery

Pooling diverse datasets from multiple pharmaceutical companies is expected to dramatically enhance the predictive capabilities of OpenFold3. With a broader range of interactions, the AI model can:

• Make better predictions
• Identify promising drug candidates faster

Expert Insights:

• Payal Sheth, VP of Discovery Biotherapeutics at BMS:
  “The federated model enables companies to innovate with predictive models for small molecule discovery that no single entity could accomplish alone.”
• Hans Bitter, Head of Computational Sciences at Takeda:
  “We are an industry that needs to do more working together than we could ever accomplish alone.”
  This underscores the initiative as a prime example of how the pharmaceutical industry can collaborate on R&D to better serve patients.

The Role of OpenFold3

OpenFold3 is the latest iteration of its predecessor, OpenFold, and represents a significant advancement in AI-based drug discovery. Co-created with Columbia University’s AlQuraishi Lab, the model is designed to:

• Predict protein folding structures
• Analyze interactions with small molecules

These predictions are crucial for:

• Identifying candidate therapeutic targets
• Designing molecules capable of modulating biological functions

Success Factor:
The effectiveness of OpenFold3 heavily depends on the quality and diversity of the data it is trained on. By incorporating datasets from multiple drug companies, the model can learn more diverse relationships, enhancing accuracy and reliability.

Considerations for Future Drug Development

The collaboration between BMS, Takeda, and their partners marks a key turning point in integrating AI into pharmaceutical research.

Benefits Include:

• Faster development of new therapies
• More efficient path for bringing treatments to market
• Demonstrates that collaborative data-sharing can advance science while protecting proprietary information
• Sets a precedent for future industry partnerships as AI becomes increasingly prominent in drug discovery

Conclusion

The partnership between Bristol Myers Squibb and Takeda Pharmaceuticals, along with other leading pharma companies, represents a new face of drug discovery. By leveraging AI and adopting best practices for secure data sharing, these companies are not only enhancing their own research but also contributing to the broader scientific community’s understanding of complex biological systems.

As this collaboration progresses, it holds the potential to expedite the development of next-generation therapeutics, ultimately benefiting patients worldwide, particularly those affected by cancer.