The US-based startup Radical Numerics has administered AI-designed bacteriophages to a patient suffering from an antibiotic-resistant E. coli infection. The company utilized its newly unveiled genomic model, Omnii-1, to modify the T7 bacteriophage, specifically targeting the gp17 protein responsible for bacterial cell recognition.
The experimental treatment was authorized by US regulators under an expanded access program for patients with serious conditions lacking alternative therapies. Radical Numerics noted that the therapy is not approved for general use, and the case has not yet undergone independent peer review or publication in a scientific journal.
Laboratory results indicate that Omnii-1 successfully designed functional variants in 69.9% of cases, compared to 4.8% for random modifications. Out of 6,694 variants tested, 2,161 demonstrated higher efficacy than the original T7 phage, with some versions featuring up to 28 mutations.
CEO Eric Nguyen stated that the technology aims to reduce the time required to identify suitable phage therapy from months to just a few days. The startup was founded by former researchers from Stanford University and the Arc Institute, who previously worked on AI-driven viral genome design.
