AlQuraishi lab’s OpenFold3 to Be Fine-Tuned with Pharma Industry Data in a Secure AI Collaboration Powered by Apheris

OpenFold3, a structure prediction system developed by AlQuraishi Lab at Columbia University, will be fine-tuned using proprietary data from AbbVie and Johnson & Johnson in a confidentiality-preserving and secure federated environment powered by Apheris.
Marie Roehm
Marketing
Published
Last updated

Germany, BERLIN, March 27, 2025 – Apheris GmbH announced today an industry-sponsored initiative by the AI Structural Biology (AISB) Consortium. OpenFold3, a structure prediction system developed by AlQuraishi Lab at Columbia University, will be fine-tuned using proprietary data from AbbVie and Johnson & Johnson in a confidentiality-preserving and secure federated environment powered by Apheris.

One major challenge in advancing AI algorithms for modern drug discovery today is the limited availability of data on protein and ligand structures necessary for training AI models that are impactful in industrial settings. Public databases are commonly acknowledged as insufficient for the precision required in drug discovery portfolios.

The goal of this collaboration is to improve predictive accuracy and generalizability, bridging the gap between current algorithmic capabilities and complex industrial demands. The group will evaluate and refine OpenFold3 for predicting 3D structures of molecule complexes, focusing on small molecule-protein and antibody-antigen interactions. To accomplish this, the initiative leverages proprietary structural data from AbbVie and Johnson & Johnson in a confidentiality-preserving manner. This group will leverage Apheris' product to ensure secure and privacy-preserving collaborative AI model training, while keeping private data in its source environments. AWS cloud services further enhance scalability, reliability, and performance.

“Despite rapid advances in modeling of protein structure, capturing how proteins interact with drugs remains an open problem, even when using the most advanced machine learning tools available. The key bottleneck is data, which, fortunately, exists in large quantities in pharmaceutical repositories, but, unfortunately, cannot be readily shared due to intellectual property concerns. Federated learning, as exemplified by the AISB consortium, solves this problem by safeguarding private IP while enabling full utilization of available data across multiple companies, and ultimately, the whole industry.” – Mohammed AlQuraishi, Professor at Columbia University

“Through this consortium we can share data with other pharma partners, exploring the hypothesis that each of our internal data sets will be highly complementary when training AI models,” said John Karanicolas, Head of Computational Drug Discovery, AbbVie. “ The result could be transformative in how we advance AI-driven drug discovery to develop better medicines faster.

"In the life sciences space, data confidentiality and IP protection are non-negotiable. By enabling secure, federated learning with the Apheris’ product, AISB is setting a new standard for privacy-preserving AI training." – Robin Röhm, Co-founder and CEO, Apheris.

About the AISB Consortium

The AISB Consortium focuses on advancing AI in drug discovery. Within a secure, federated computing paradigm, its long-term vision and ultimate goal is to accelerate the application of AI in molecule design by achieving precision akin to X-ray crystallography.

About Apheris GmbH

Apheris powers federated life sciences data networks, addressing the critical challenge of accessing proprietary data locked in silos due to IP and privacy concerns. The Apheris product is a federated computing infrastructure with governance, security and privacy controls. For more information, visit: www.apheris.com

For additional information, please contact: press@apheris.com

AIDrugDiscovery
Pharma
News
Share blog post to Linked InTwitter

Insights delivered to your inbox monthly