NVIDIA and DeepMind open protein structures for 2,800-plus viruses on AlphaFold

NVIDIA, Google DeepMind and EMBL-EBI announced on September 24, 2026 the open release of predicted 3D structures for protein complexes of more than 2,800 viruses in the AlphaFold Database. The work uses AlphaFold2, accelerated by the NVIDIA BioNeMo Inference Runtime, and prioritizes viral families known to infect humans.

What changed

Most proteins do not act alone; they form complexes that vaccines and drugs often target. According to the NVIDIA blog, about 30% of the newly published interactions are new to science and were absent from the experimental Protein Data Bank. The AlphaFold store now holds more than 260 million predictions.

NVIDIA also released the BioNeMo Structure Prediction Pipeline so other groups can predict structures from their own sequences.

Why it matters

Prior coronavirus research sped COVID-19 vaccines. The next pathogen may not offer that head start. The set ranges from common-cold viruses to threats such as mpox. Predictions carry confidence scores: they are structural hypotheses for experiments, not replacements for crystallography.

Partners include CEPI, Seoul National University, Sungkyunkwan, the Swiss Institute of Bioinformatics and the University of Glasgow. The release coincided with a UN General Assembly meeting on pandemic prevention on September 25. EMBL stresses access for labs with fewer resources.

Sources

Transparency: This content was created, edited or reviewed with the aid of artificial intelligence. Information was cross-checked with public posts on X and sources available on the internet. Consult the original sources for the full context.

Por GeekikiBot