NVIDIA and Google DeepMind Map Protein Complexes of Over 2,800 Viruses to Accelerate Pandemic Defense
A global coalition led by NVIDIA, Google DeepMind, and EMBL-EBI has published predicted 3D protein complex models for more than 2,800 viruses to aid drug and vaccine development. Accelerated by NVIDIA's BioNeMo runtime, roughly 30% of the newly mapped protein interactions were previously unknown to science.
Open AI Models Target Viral Threats Across 2,800 Pathogens
NVIDIA announced a partnership with Google DeepMind and the European Molecular Biology Laboratory’s European Bioinformatics Institute (EMBL-EBI) on September 24, 2026, to publish predicted 3D structures of protein complexes across more than 2,800 viruses. The dataset is freely accessible to the global scientific community through the AlphaFold Database's Pandemic Preparedness Portal.
The initiative brings together a broad international research consortium that includes the Coalition for Epidemic Preparedness Innovations (CEPI), the Medical Research Council-University of Glasgow Centre for Virus Research, the Swiss Institute of Bioinformatics, Seoul National University, and Sungkyunkwan University. Alongside the dataset, NVIDIA has made its GPU-accelerated BioNeMo Structure Prediction Pipeline publicly available, allowing researchers to convert custom protein sequences into 3D structural predictions.
Scaling AlphaFold2 via BioNeMo to Map Uncharted Interactions
To model complete viral proteomes at scale, the team ran Google DeepMind's AlphaFold2 using the NVIDIA BioNeMo Inference Runtime for GPU acceleration. While traditional experimental methods—such as protein crystallization combined with X-ray diffraction—can take years and run thousands of dollars per target, the optimized AlphaFold2 pipeline outputs predictions in minutes.
The project focused explicitly on protein complexes—clusters of interacting proteins that viruses use to execute core biological functions. These molecular assemblies frequently represent the exact target sites required for designing therapeutic drugs and preventative vaccines. The survey examined virus families capable of infecting humans, spanning benign common-cold strains to high-consequence pathogens like Mpox.
Significantly, about 30% of the mapped protein interactions represent shapes and arrangements that have never been documented in the Protein Data Bank, the historical open repository for experimentally verified protein structures. Each prediction in the database is tagged with an algorithmic confidence score to guide lab verification.
"This database is an engine for hypothesis generation," said Chris Dallago, applied research science team lead in digital biology at NVIDIA. "We’re enabling biologists and the AI community to investigate protein interactions, not just as single molecules but as complexes, so the whole field can move forward."
"When the next pandemic happens, there may be something that comes out of the blue, and we’ll be lacking the knowledge we had for COVID," added Joe Grove, a professor of molecular virology at the University of Glasgow who collaborated on the work. "What we’re trying to do is stockpile some of that knowledge ahead of time."
Preempting Future Outbreaks as AlphaFold Expands Past 260 Million Structures
The release coincided with a United Nations General Assembly meeting convened by the World Economic Forum in New York City addressing global pandemic prevention, preparedness, and response. The urgency is underscored by data from the Center for Global Development, which projects an estimated 50% probability that the world will encounter a pandemic on par with COVID-19 by 2050.
With this release, the AlphaFold Database expands past 260 million predicted proteins and protein complexes, representing nearly all cataloged proteins known to modern biology.
According to Jo McEntyre, interim director of EMBL-EBI, releasing the data without proprietary restrictions is crucial for frontline teams: "The dataset also covers lesser-studied viruses and lowers the barriers for scientists in low-resource settings who are confronting outbreaks firsthand."
