AlphaFold Expands Predictions to Protein Complexes and Molecular Interactions
Google DeepMind and partners have released large-scale predictions for protein complexes, including homomeric and heteromeric structures across thousands of proteomes. The update includes high-confidence structures for millions of interactions, accelerating research in biology and drug discovery.
TLDR
A new collaboration between EMBL-EBI, Google DeepMind, NVIDIA, and Seoul National University has added millions of AI-predicted protein complex structures to the AlphaFold Database. This includes high-confidence predictions for homodimers and other complexes, providing new insights into how proteins interact in cells.
The dataset prioritizes proteins important for human health and disease, with 1.7 million high-confidence homodimer predictions added, and more to come.
Key Advances
- Large-scale study of over 31M predicted homo- and heteromeric complexes.
- High-confidence structures for millions of protein interactions.
- Open data release to support global research in systems biology and drug discovery.
Early applications include better understanding of cellular machinery and potential drug targets.
Why this story matters
The ability to model not just individual proteins but their interactions at scale transforms systems biology. It enables faster hypothesis generation in drug discovery, synthetic biology, and understanding of diseases involving protein complexes. This expansion of AlphaFold brings the technology to the 'next level' for studying molecular interactions.
Sources
- EMBL-EBI: "Millions of protein complexes added to AlphaFold Database" (March 18, 2026)
- Nature: "AlphaFold database hits ‘next level’: the AI system now includes protein pairing" (March 17, 2026)
- bioRxiv preprint: "AlphaFold Database expands to proteome-scale prediction of protein complexes" (March 27, 2026, work from March)
- Google DeepMind and Isomorphic Labs updates on AlphaFold 3 and complexes (2026)