AI designs functioning viruses in lab study

US researchers have used AI to design complete viral genomes that became fully functional viruses in the laboratory. The novel viruses were bacteriophages, which infect bacteria rather than people, and were created as a potential step toward new treatments for antibiotic-resistant infections.

The Stanford team trained models called Evo1 and Evo2 on genetic code from viruses, bacteria, plants and people, then refined them to generate bacteriophage designs. Researchers selected 302 AI designs and synthesised them in the lab, where 16 proved effective at killing E. coli bacteria.

The work has been described as a major milestone for synthetic biology, with researchers saying AI may help design phages, drugs, enzymes and antibodies. The phage genome is around 5,400 base pairs long, compared with around 500,000 base pairs for the smallest genome of a living cell and three billion base pairs for the human genome.

The advance also raised biosafety and biosecurity concerns. Experts at the Center for Health Security at Johns Hopkins University warned that generative viral genome design now poses urgent questions, while the researchers said they excluded viruses that infect complex organisms, worked only on phage and conducted the research in a secure laboratory.