AI-designed E. coli killer points toward new ways to fight antibiotic-resistant bacteria
In science, the suffix "phage" means to devour. Thus, the name bacteriophage might conjure images of tiny creatures that "eat" bacteria. While that conception isn't 100% scientifically accurate, bacteriophages are lethal killers of bacteria nonetheless, and scientists are excited about engineering phage DNA as a path to new antibiotics.
Scientists at Stanford University have developed a new method to combat antibiotic-resistant bacteria using engineered bacteriophages, or phages. Chemical engineer Brian Hie and bioengineering graduate student Samuel King created Evo 2, a generative AI model that designs DNA sequences for novel phages. By inputting the DNA of the ΦX174 phage, Evo 2 generated nearly 300 new phages, which were tested against E. coli bacteria.
The researchers narrowed down the list to 16 highly effective phages that showed greater fitness than the native ΦX174 phage.
The ability to target E. coli with multiple genetically distinct phages may help overcome bacterial resistance, a common weakness of antibiotics. As bacteria evolve immunity to single medications, a cocktail of diverse phages could make it harder for bacteria to develop resistance. The team envisions this approach could be used to create phages targeting other harmful bacteria, such as tuberculosis, MRSA, and Pseudomonas aeruginosa.
King developed a computational framework to evaluate the genomes generated by Evo 2, selecting the most promising candidates for further exploration. This framework considered traits based on ΦX174 and related phages, focusing on the most viable alternatives. While DNA synthesis is still costly, the team's approach helped narrow down the options.
Evo 2 has been made open source, allowing anyone to design new genomes themselves. Hie is optimistic about the tool's potential to revolutionize medical and biological sciences, providing a powerful advantage against pandemics and biological threats.
Written by urgent.news from Medical Xpress's reporting — not their text. Machine-written; read the original for the full account.
