The AISB Consortium is a novel collaboration aimed at helping to transform drug discovery and development.
The AISB Consortium brings together biopharmaceutical companies and cutting-edge federated learning technology to harness the collective expertise and data of industry leaders to push the boundaries of AI-driven drug discovery.
One major challenge in advancing AI algorithms for modern drug discovery is the limited availability of data on protein and ligand structures necessary for training AI models. Public databases do exist, yet they are commonly acknowledged as insufficient. Biopharmaceutical companies have made considerable investments in assembling extensive proprietary databases of these structures. However, the proprietary nature of this data presents challenges in terms of accessibility and collaboration.
This consortium is seeking to address this. For the first time, biopharmaceutical companies will allow AI models predicting the 3D structure of molecule complexes to be collaboratively trained on their proprietary protein structure data. Federated learning approaches will be leveraged to protect data confidentiality and prevent proprietary information from being shared across the partner companies.
Apheris provides the technology for secure, private and governed ML & analytics on distributed data. Leveraging federated learning and computational governance, this technology enables trust among parties and ensures the highest standard for data confidentiality and protection.
Biopharmaceutical company members, including AbbVie and Johnson & Johnson, provide structural biology expertise and proprietary data developed by each individual company. The combined expertise and data of the consortium members will allow AI models to be developed beyond the current limitations associated with training on publicly available and internal data only, while still preserving the confidentiality of the proprietary data.
Our collaborative vision
The AISB Consortium aims to leverage secure, federated learning to ensure both data privacy and AI model performance – with the ultimate goal of accelerating the application of AI in molecule design by achieving precision akin to X-ray crystallography through AI and machine learning in predicting protein complex structures. Our vision thrives on collaboration, combining our collective expertise to redefine drug discovery.
- Data available for training while preserving confidentiality/IP
- Infrastructure (Federated Learning, high performance compute)
- Leading edge AI algorithms
- Partnership between pharma and tech
Privacy-preserving Data Collaboration
Success of the AISB Consortium hinges on joint research projects where proprietary data of biopharmaceutical companies is used to collaboratively train ML models and unlock new data insights. Data confidentiality is always protected. Apheris’s technology ensures privacy and security, enhancing computational efficiency and secure model sharing.