NVIDIA and DeepMind Release 2,800 Viral Protein Complex Predictions

NVIDIA and Google DeepMind have released a dataset containing 2,800 predicted viral protein complexes, coinciding with a United Nations General Assembly meeting focused on pandemic preparedness. This dataset, part of the AlphaFold Database, expands the total number of predictions to 260 million, with 30% of these interactions being new to science. Joe Grove, a professor at the Medical Research Council-University of Glasgow, highlighted a 50% chance of a pandemic as severe as COVID-19 by 2050, emphasizing the need for robust data to prepare for unforeseen outbreaks. Chris Dallago from NVIDIA described the database as an 'engine for hypothesis generation,' while Jo McEntyre from EMBL-EBI stressed the importance of open data for advancing viral diagnostics and treatment development. This initiative aligns with the Coalition for Epidemic Preparedness Innovations' goals and involves collaboration with institutions like Seoul National University and the Swiss Institute of Bioinformatics. Risha Patel from Google DeepMind stated that the ambition of the AlphaFold Database is to democratize access to foundational biology at scale, providing essential resources for current Ph.D. researchers and accelerating fundamental science.