Coursera Spring Sale
40% Off Coursera Plus Annual!
Grab it
Explore a comprehensive 52-minute conference talk presenting AtomWorks, a revolutionary data framework designed to accelerate the development of state-of-the-art biomolecular foundation models for diverse applications including structure prediction, generative protein design, and fixed backbone sequence design. Discover how this framework addresses the considerable challenges in developing and training new deep learning models for biomolecular structure prediction, building upon the transformative impact of protein structure databases in the field. Learn about RosettaFold-3 (RF3), a cutting-edge structure prediction network trained using AtomWorks that demonstrates the capability to predict arbitrary biomolecular complexes with enhanced chirality treatment, effectively narrowing the performance gap between closed-source AlphaFold3 and existing open-source implementations. Gain insights into the technical innovations that make RF3 a broadly useful structure prediction tool and understand how AtomWorks is positioned to accelerate the next generation of open-source biomolecular machine learning models. Access information about the open-source release of both the AtomWorks framework and RF3, including curated training data, code, and model weights, all made available under a permissive BSD license to support the AI for drug discovery community's collaborative research efforts.
