New superbug-killing antibiotic discovered using AI

Abstraction: AI accelerates discovery of novel narrow-spectrum antibiotic compounds

Key points:

• AI trained on thousands of known drug-bacterium interactions, then screened 6,680 unknown compounds in ~1.5 hours, shortlisting candidates
• Identified abaucin, highly potent against Acinetobacter baumannii — described by Dr. Jonathan Stokes (McMaster) as "public enemy number one" for antibiotic resistance
• Abaucin is unusually narrow-spectrum, killing only A. baumannii; researchers believe this precision will slow resistance emergence and reduce side effects
• Lab tests showed abaucin treated infected wounds in mice and killed patient-derived A. baumannii samples
• Clinical availability projected no earlier than 2030 after further lab refinement and trials
• Team plans to apply the same AI pipeline to Staphylococcus aureus and Pseudomonas aeruginosa next

Connections: Mcmaster University · Mit · Drug Discovery AI · Antibiotic Resistance

Source: https://www.bbc.com/news/health-65709834