Antibiotic resistance poses a significant challenge to humanity, emphasizing the urgent need for new antibiotics. While the majority of antibiotics are derived from fungi and bacteria, Archaea presents largely untapped sources for discovering new antibiotics. In a recent study, researchers at the University of Pennsylvania employed deep learning techniques to investigate paleozoans. By analyzing the proteomes of 233 archaeal species, we discovered 12,623 potential antibacterial compounds.

“Previous attempts to discover new antibiotics have mainly focused on fungi, bacteria, and animals,” stated Dr. Cesar de la Fuente, a researcher at the University of Pennsylvania.

“Historically, we have leveraged AI models to identify antibiotic candidates across various sources, from the DNA of extinct organisms to the compounds found in animal venom.”

“We are now applying these methodologies to a new dataset: hundreds of proteins from ancient microorganisms.”

“There are undoubtedly other life forms waiting to be investigated.”

In contrast to bacteria and eukaryotes (which include plants, animals, and fungi), Archaea represents a distinct branch on the evolutionary tree.

While they may resemble bacteria under a microscope, archaeal organisms differ fundamentally in their genetics, cell membranes, and biochemistry.

These unique features enable them to thrive in some of Earth’s most extreme environments, such as heated seabed vents and scalding hot springs like Yellowstone National Park.

Archaea typically flourish in isolation from other organisms, and their biology has evolved in unprecedented ways, with limited evolutionary pressure, exposure to toxic substances, and extreme temperatures.

This presents a promising, yet largely uncharted, source for novel molecular tools, including compounds that may act like antibiotics but function differently than existing treatments.

“Our interest in archaeal organisms stems from their biochemical adaptations to extreme environments,” remarked Dr. Marcelo Torres from the University of Pennsylvania.

“We hypothesized that having survived billions of years under such conditions, they might possess unique strategies to fend off microbial rivals.”

To uncover antibiotic compounds within Archaea, the researchers utilized artificial intelligence.

They adapted an upgraded version of APEX, an AI tool initially designed to identify antibiotic candidates from ancient biological sources, including proteins from long-extinct animals like woolly mammoths.

With thousands of peptides (short amino acid chains) known for their antimicrobial properties, the AI can predict the likelihood that a given amino acid sequence will exhibit similar effects.

By re-calibrating APEX 1.1 to incorporate data from thousands of additional peptides and pathogenic bacteria, the scientists established tools to forecast which peptides in Archaea might inhibit bacterial proliferation.

Upon scanning 233 archaeal species, over 12,000 potential antibiotic candidates were identified.

The authors labeled these molecular compounds, and chemical analysis indicated they differ from known antimicrobial peptides (AMPs), notably in their charge distribution.

The team subsequently selected 80 archaeal compounds for further testing against live bacteria.

“Finding new antibiotic molecules individually feels like searching for needles in a haystack,” commented Famping Wang, a postdoctoral researcher at the University of Pennsylvania.

“AI accelerates the search by pinpointing the location of the needle.”

Antibiotics can function through various mechanisms. Some disrupt bacterial membranes, while others inhibit protein synthesis within the organism.

Notably, the researchers found that unlike many known AMPs that target the outer defenses of bacteria, Alcaeasen operates by disrupting internal electrical signals that are vital for cell survival.

Tests on drug-resistant bacteria revealed that 93% of the 80 alkadeins exhibited antibacterial activity against at least one bacterial strain.

The team chose three alkaiersins to evaluate in animal models.

Four days following a single dose, all three alkaiersins halted the spread of drug-resistant bacteria commonly acquired in hospitals.

One of these compounds exhibited activity on par with polymyxin B, an antibiotic often used as a last resort against drug-resistant infections.

“This study underscores the vast potential for discovering new antibiotics within Archaea,” stated Dr. De La Fuente.

“As the prevalence of antibiotic-resistant bacteria rises, exploring unconventional sources for new antibiotics is essential.”

A paper detailing the results was published today in Nature Microbiology.

____

MDT Torres et al. Deep learning reveals antibiotics in the archaeal proteome. Nat Microbiol. Published online on August 12, 2025. doi:10.1038/s41564-025-02061-0