AI Uncovers 12,000 Potential New Antibiotics in Ancient Microbes

Antibiotic resistance is one of humanity’s most significant health threats, killing an estimated 1.27 million people a year. Scientists warn it could get far worse. However, a new study suggests that the next lifesaving drug might come from one of Earth’s oldest and strangest life forms.

Using a deep-learning AI tool called APEX, researchers at the University of Pennsylvania scanned the proteins of 233 species of Archaea. These single-celled organisms aren’t bacteria, plants, animals, or fungi. They survive in extreme environments like scalding hot springs, deep-sea vents, and toxic salt flats, evolving unique biochemical defenses over billions of years.

The AI identified 12,623 molecules with potential antimicrobial activity. These compounds, called archaeasins, appear to work differently from most known antibiotics. Instead of attacking a bacterium’s outer defenses, they scramble its internal electrical signals, shutting the cell down from the inside.

Scientists synthesized 80 archaeasins and tested them against dangerous, drug-resistant bacteria like E. coli, Klebsiella pneumoniae, Pseudomonas aeruginosa, and Staphylococcus aureus. Impressively, 93% killed at least one strain. In mice infected with a hospital-acquired superbug, three archaeasins stopped the infection in just four days – and one worked as well as polymyxin B, a powerful “last-resort” antibiotic.

With antibiotic resistance rising and traditional drug pipelines drying up, searching in Archaea could open an entirely new frontier in medicine.

Read the study: M.D.T. Torres et al. Deep learning reveals antibiotics in the archaeal proteome. Nat Microbiol, published online August 12, 2025.