Beyond COVID: New AI Project Targets 2,800 Potential Threat Viruses
What Happened
A global team led by NVIDIA, Google DeepMind and research partners has used AI to predict the 3D structures of proteins from more than 2,800 human-infecting viruses. The data has been added to the open-access AlphaFold Database, giving scientists a head start in studying potential future pandemic threats. The project combines DeepMind's AlphaFold2 with NVIDIA's BioNeMo platform to map viral proteins at scale. Researchers say this could help speed up the development of treatments and vaccines if a new virus emerges. About 30% of the protein interactions identified were previously unknown to science.
Key Takeaways
Unlike COVID-19, which benefited from decades of prior coronavirus research, the next pandemic pathogen may be largely unknown. By creating a vast AI-generated virus database now, scientists hope to accelerate future outbreak responses and drug discovery efforts.