Researchers have administered AI-designed, bacteria-targeting viruses to a person in what they claim is a first, but don’t press them for results: They’re not ready to say how it went quite yet. Radical Numerics, a San Francisco-based AI lab developing AI models for bioscience, announced the treatment this week, claiming it marks the first time a cocktail containing AI-designed bacteriophages has been administered to a human patient under a compassionate-use arrangement. These aren’t just any viruses, mind you: Omnii-1, the AI model publicly unveiled by Radical Numerics this week, was used to design viable bacteriophages, a type of virus that specifically infects bacteria. Therapeutic use of phages isn’t novel in and of itself, with the bacteria-targeting viruses being sought as an alternative to combat the rise of antimicrobial-resistant bacteria that are increasingly rendering antibiotics ineffective. Phages have a number of potential advantages over antibiotics, most notably their ability to target specific bacterial strains – a big plus when it comes to sparing beneficial gut bacteria that can be affected by broad-spectrum antibiotics. AI has already shown promise in speeding up phage engineering, as we reported last year when a group of researchers from Stanford University and the Arc Institute created AI-designed phages that targeted E. coli bacteria. The phages, designed using the genetic sequence of a well-studied natural E. coli phage, included several variants that outperformed their biological predecessor in laboratory tests. Similar to the Stanford work, which is referenced in Radical Numerics' write-up, the company used its own AI model to design bacteriophages, which it then tested against bacteria isolated from the antibiotic-resistant infection its patient was suffering from. “The cocktail containing Omnii-1-designed phage variants eliminated the patient's antibiotic-resistant isolate in vitro,” the company explained, so there’s at least a success on record there. “The investigational cocktail was subsequently administered to the patient under compassionate use.” Unfortunately, that’s where the trail of results ends, as Radical Numerics didn’t publish anything about the results of testing the phage in a human, citing the need to protect their privacy. The company also doesn’t want to give the impression that its AI phage therapy is generally safe or effective, noting that observations from a single patient don’t establish much of anything. It’s not even stated what infection the patient was suffering from, though the research write-up does mention testing Omnii-1’s ability to predict which phages could infect Pseudomonas aeruginosa, a bacterium that can develop antibiotic resistance and cause infections in patients with burns and other wounds, as well as persistent lung infections in cystic fibrosis patients. The AI model “performed competitively” against a specialized model at predicting phage-host interactions, the research write-up notes. There’s nothing in the paper to suggest the patient was suffering from a Pseudomonas aeruginosa infection. “This marks an important milestone in translating AI-designed biology from computational predictions and laboratory validation to clinical administration,” Radical Numerics said of its work. The treatment takes the technology beyond laboratory testing, even if it turns out not to fully eliminate the infection. Radical Numerics hopes its work will lead to a process for designing and building phages targeted to individual bacterial infections, and the need for alternatives is already pressing – antibiotic resistance is a growing global health threat, with drug-resistant infections already causing substantial mortality worldwide. Radical Numerics didn’t respond to questions for this story. ®
