Stanford researchers use the Evo 2 genome language model to design ΦX174-based bacteriophages that kill E. coli

Stanford researchers led by Brian Hie and Samuel King used Evo 2, a generative AI model that writes whole genomes, to design novel bacteriophage genomes starting from ΦX174. They synthesized nearly 300 phages, tested them against E. coli and ended up with 16 exceptional ones; a cocktail of the 16 rapidly overcame resistance in E. coli that was immune to native ΦX174. The work is described in a new paper, and Evo 2 has been made openly and freely available.