GermanyResearchers at Stanford University and the Arc Institute have used artificial intelligence to design fully functional, living bacteriophage genomes for the first time. Some of these artificially created phages outperformed their natural counterparts by replicating faster, displacing natural variants, or being more effective at killing host bacteria. The study, published in the journal Science, includes an unusual warning about the potential risks of such technology. Bacteriophages are viruses that exclusively infect bacteria and have been considered a potential weapon against antibiotic-resistant microbes due to their ability to target bacteria specifically. However, bacteria can develop resistance mechanisms against phages, making it necessary for effective phage therapies to adapt quickly to new resistances, something AI could potentially assist with. The team led by lead author Samuel King utilized genome language models called Evo 1 and Evo 2, trained on vast amounts of genetic data, to generate phage genomes. Starting with the well-studied ΦX174 virus, they generated 302 candidate genomes, of which 285 were synthesized and tested in the lab. Sixteen of these synthetic phages, a
Bias read (Center): The article discusses scientific research involving AI-generated bacteriophages and does not present any political stance, controversy, or ideological framing. It focuses purely on the technical achievements and implications within the field of microbiology.
