AI-designed bacteriophages administered to patient in compassionate-use case

First compassionate-use administration

Radical Numerics, a San Francisco-based AI lab developing AI models for bioscience, announced this week that a cocktail containing AI-designed bacteriophages was administered to a human patient under a compassionate-use arrangement. The company claims it is the first time such a cocktail has been given to a person. Bacteriophages are viruses that specifically infect bacteria, and the company used its Omnii-1 AI model, publicly unveiled this week, to design viable phages.

The company said the cocktail containing Omnii-1-designed phage variants eliminated the patient's antibiotic-resistant isolate in laboratory testing. The investigational cocktail was subsequently administered to the patient under compassionate use. Radical Numerics did not publish results from human testing, citing the need to protect patient privacy. It also said observations from a single patient do not establish whether the therapy is generally safe or effective.

How the phages were designed and tested

Therapeutic use of phages is not new. The bacteria-targeting viruses are being sought as an alternative to antibiotics, which are increasingly rendered ineffective by antimicrobial-resistant bacteria. Phages can target specific bacterial strains, which may help spare beneficial gut bacteria that can be affected by broad-spectrum antibiotics. AI has also been used to speed phage engineering; researchers from Stanford University and the Arc Institute previously created AI-designed phages targeting E. coli, including variants that outperformed their biological predecessor in laboratory tests.

Radical Numerics tested its phages against bacteria isolated from the antibiotic-resistant infection the patient was suffering from. The company did not state what infection the patient had. Its research write-up mentions 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 model performed competitively against a specialized model at predicting phage-host interactions, but there is nothing in the paper to suggest the patient had a Pseudomonas aeruginosa infection.

What remains unknown

Radical Numerics called the treatment an important milestone in translating AI-designed biology from computational predictions and laboratory validation to clinical administration. The company hopes its work will lead to a process for designing and building phages targeted to individual bacterial infections. Antibiotic resistance remains a growing global health threat, with drug-resistant infections already causing substantial mortality worldwide. Radical Numerics did not respond to questions for the story.

Sources: The Register