← Back to Home

AI-designed phages administered to patient with drug-resistant infection; treatment outcomes not yet disclosed

Radical Numerics says AI-designed phages have been administered to a patient through compassionate use. Laboratory results showing bacterial killing provided a basis for the attempt, but publicly available information cannot yet answer whether the patient benefited.

By SURL BioNews

When antibiotics struggle to control an infection, another treatment approach is to find viruses that can attack the disease-causing bacteria. Yet finding bacteriophages that match a patient's bacterial strain is often a major challenge. Radical Numerics announced on October 20 that phages designed by its Omnii-1 model had been administered to a patient with a drug-resistant bacterial infection. The news of this use in a human also brings AI biological design to a stage where clinical evidence is needed to answer questions.

According to the company, a phage cocktail containing AI-designed variants first eliminated the patient's bacterial isolates in laboratory experiments, then was administered through an authorized, physician-led compassionate use pathway. The company calls this the world's first clinical use of AI-designed phages in a human, but this claim of priority and the details of the case currently come primarily from materials published by the company itself.

Bacteriophages infect bacteria, and some types can replicate inside them, ultimately causing the bacteria to lyse. Their highly specific ability to recognize hosts also makes matching particularly difficult: a phage that can infect one bacterial strain may not be able to target another strain of the same species. AI's role here is to help propose and screen potentially effective designs, which are then tested experimentally.

The company's publicly presented technical demonstration uses the T7 phage as an example, modifying sequences associated with gp17, a tail fiber protein involved in recognizing bacteria, rather than generating an entire viral genome from scratch each time. Omnii-1 learns from genomic data and data on phage–host interactions, then uses experimental results to help select candidate variants. However, tests of these design and matching capabilities cannot be taken as direct evidence of efficacy in a patient.

TechRadar reported on October 30 that the company's technical account had not yet undergone peer review, and that the patient's treatment outcomes, hospital, and treating physician had not been disclosed. The report quoted Harvard microbiologist Michael Baym, who was not involved in the work, expressing continued skepticism until he sees evidence of treatment efficacy and the novelty of the viruses. The company said it was withholding the patient's clinical course and outcomes to protect their privacy.

Compassionate use gives severely ill patients who lack suitable alternative therapies an opportunity to receive investigational treatment outside clinical trials; it does not mean a product has been approved for widespread use. The company also acknowledges that observations from a single patient cannot establish general safety or efficacy. This announcement shows AI design taking a step toward administration in humans. Whether it can control infection, what adverse reactions it may cause, and whether it can reliably serve more patients still require clinical data that can be examined.

References

  1. JournalArta
  2. Radical Numerics
  3. TechRadar