AI-designed bird flu vaccine outperforms stockpiled vaccines in animal trials

A University of Cambridge team and spinout DIOSynVax Ltd. have used artificial intelligence to design a broad-spectrum avian influenza vaccine that provided complete protection in animal trials against multiple globally circulating bird flu variants, according to a study published in Cell Reports Medicine. The approach contrasts with current government stockpiles of strain-specific vaccines, which protected only about half of animals exposed to the same set of viruses in comparative tests. The researchers, led by Professor Jonathan Heeney of the University of Cambridge’s Department of Veterinary Medicine, collected global genetic sequence data from wildlife, livestock and human avian flu infections, then used an AI model to integrate the sequences and design super-antigens containing features shared across variants. Candidates were screened in cell models and advanced to mice and ferrets, which were challenged with genetically distinct avian flu strains linked to human disease and pandemic risk. The experimental vaccine protected all treated animals, while the World Health Organization pandemic preparedness vaccine showed only partial protection. The work demonstrates a plug-and-play method for rapidly engineering vaccines against RNA viruses that mutate quickly and can spill over from animals to people. By targeting common antigenic features, the vaccine is intended to anticipate viral evolution rather than respond to it after strains emerge. The team said the method could support national biosecurity by enabling faster responses to epidemics and reducing repeated vaccine reformulation. The same platform has previously produced broad-spectrum coronavirus and influenza vaccine candidates, and development continues for Lassa fever, Marburg, Ebola and multi-season flu vaccines.