DeepMind, EMBL-EBI, NVIDIA and CEPI add predicted protein-complex structures for 2,800+ viruses to the AlphaFold Database
On Sept 24, 2026 a consortium of Google DeepMind, EMBL-EBI, NVIDIA, CEPI and universities in Korea, Switzerland and the UK released predicted viral protein-complex structures for more than 2,800 viruses in the AlphaFold Database. About 30% of the interactions are new to science. The predictions were made with AlphaFold2 on NVIDIA's BioNeMo inference stack, and the database now holds 260M+ predictions.
Key facts
- Complex structures for 2,800+ viruses; ~30% of the interactions are new to science
- Partners: Google DeepMind, EMBL-EBI, NVIDIA, CEPI, Seoul National University, Sungkyunkwan University, SIB, University of Glasgow
- Built with AlphaFold2 on the NVIDIA BioNeMo Inference Runtime; NVIDIA open-sourced its BioNeMo Structure Prediction Pipeline
- AlphaFold Database now holds 260M+ predictions
Science result
- Field
- biology / structural virology
- Problem
- Proteome-scale structures of virus protein–protein complexes
- Result
- Predicted complex structures for 2,800+ viruses released openly; ~30% of interactions previously unknown
- AI system
- AlphaFold2, NVIDIA BioNeMo
- Human role
- Human-led consortium using AI prediction at scale
- Verification
- Computational predictions (not experimentally validated at release)
- Status
- confirmed
What happened
An open data release applying structure prediction to viral protein complexes at pandemic-preparedness scale, coordinated with CEPI.
Why it matters
It gives vaccine and antiviral researchers structural hypotheses for thousands of viruses at once. The interactions are predictions and still need experimental checks.
Changelog
- 2026-09-29: created
Related events
- AlphaFold 2 solves protein structure prediction at CASP14 ★★★★★
- WHO AFRO, CEPI and DRC's INRB use Claude in the Bundibugyo Ebola outbreak response ★★★
Sources (1)
id: 2026-09-24-viral-protein-complex-structures-alphafold-db · updated 2026-09-29 · open in the interactive timeline