A team at the Massachusetts Institute of Technology used generative artificial intelligence to search vast chemical space and propose new antibiotic compounds capable of tackling two high-profile drug-resistant infections. The researchers ´designed more than 36 million possible compounds and computationally screened them for antimicrobial properties´, according to MIT News. From that computational sweep they selected two leading candidates for laboratory and animal testing.
The two molecules, named NG1 and DN1, are structurally different from existing antibiotic classes and appear to act through novel mechanisms that disrupt bacterial cell membranes. In vitro assays and mouse models showed that NG1 cleared infections caused by gonorrhoea while DN1 worked against methicillin-resistant Staphylococcus aureus. Early results indicate these compounds bypass known chemical scaffolds, which could make them less vulnerable to the resistance pathways that compromise many current drugs.
Researchers framed the work as an exploration of underused areas of chemistry rather than an incremental tweak of existing antibiotics. Senior author professor James Collins described the effort as opening ´new possibilities´ for antibiotic development and enabling access to much larger chemical spaces. Lead author Aarti Krishnan said the team deliberately excluded features that would resemble existing antibiotics to ´help address the antimicrobial resistance crisis in a fundamentally different way´.
Despite the encouraging laboratory and mouse data, the compounds remain at an early stage of development. The study demonstrates that generative methods can produce biologically active molecules that traditional discovery approaches might miss, but human safety, dosing and efficacy remain to be established in clinical trials. If subsequent studies confirm safety and effectiveness, the approach could change how researchers hunt for antibiotics and offer new tools against superbugs that kill hundreds of thousands of people worldwide each year.
