Global Research Coalition Releases Dataset of Predicted Virus Structures to Combat Future Pandemics

NVIDIA has joined forces with Google DeepMind and other global research organizations to release predicted 3D structures for over 2,800 viruses in the AlphaFold Database. This move aims to help scientists prepare for future pandemics by making foundational biology more accessible at scale.

The dataset was created using AlphaFold2, an AI model developed by Google DeepMind that predicts protein structures with high accuracy. The team used NVIDIA's BioNeMo Inference Runtime to optimize the process and enable bulk predictions of viral proteomes.

This collaboration comes as a warning from experts that there may be a roughly 50% chance of a pandemic as severe as COVID-19 by 2050. By stockpiling knowledge ahead of time, researchers hope to accelerate fundamental science and improve response times in case of future outbreaks.

The dataset covers over 30% new protein interactions never documented before, providing valuable insights for the scientific community. With this open data, biologists can investigate complex protein structures more efficiently, enabling faster development of treatments and vaccines.