Tiny viruses known as bacteriophages infect bacteria and replicate inside them. These phages outnumber other biological entities on Earth and have evolved over billions of years to target bacteria and bypass their protections. This variety offers potential for medicine, as some phages are being tested against infections resistant to antibiotics. However, selecting effective phages requires deeper insight into their infection mechanisms and performance inside the body. Researchers at Stanford University recently applied AI to create the full DNA sequence of a phage. The resulting synthetic genome produced a working virus in lab tests, showing AI can model biological systems accurately enough for real-world results. The study focused on a small, well-known phage called ΦX174 that targets E. coli. While this proves the concept, larger therapeutic phages are far more complex, with bigger genomes and advanced tools to identify hosts and evade defenses. Much about phage diversity remains unclear, yet AI tools like AlphaFold now predict protein shapes from genetic data, offering clues to unknown functions. At the University of Leicester, scientists collect diverse phages from nature that attack resistant bacteria. These collections could train AI to link DNA sequences, protein structures, and actual behaviors, helping identify traits for effective treatments in patients.
