In a major development for the field of AI in structural biology, Apheris and Johnson & Johnson have joined the OpenFold Consortium, an open science initiative focused on developing high-performance, open-source models for biological and pharmaceutical research.
The expansion of the consortium highlights a growing industry-wide commitment to collaborative, AI-driven drug discovery. OpenFold has rapidly become a central player in structural biology by releasing foundational models, such as OpenFold, OpenFold-SoloSeq, and OpenFold-Multimer, that are capable of predicting protein structures with remarkable accuracy. Recent efforts have extended into the modeling of DNA and small molecules, with all tools and datasets made publicly accessible under permissive open-source licenses.
Apheris, known for its work in secure, privacy-preserving machine learning, brings a critical capability to the OpenFold ecosystem, the ability to train and fine-tune AI models on proprietary datasets without ever exposing the data itself. This approach allows consortium members to evaluate and enhance OpenFold models using sensitive, high-value biomedical data, while ensuring compliance with data protection standards.
Alongside its participation in OpenFold, Apheris will continue to lead the AI Structural Biology Consortium (AISB), a complementary initiative aimed at fostering confidential AI research across institutions with proprietary datasets. This dual-track approach reflects a larger trend in the life sciences toward hybrid models of open innovation and protected data sharing.