Antibiotic resistance is one of the biggest threats facing humanity, making the need for new antibiotics more important than ever. While most antibiotics come from bacteria and fungi, Archaea offers almost undeveloped reservoirs for antibiotic discovery. In the new study, researchers at the University of Pennsylvania utilized deep learning to systematically explore paleozoans. By mining proteomes of 233 archaeal species, we identified 12,623 min of molecules with potential antibacterial activity.
Torres et al. 80 synthetic archiasins, 93% of which exhibited antibacterial activity in vitro Against acinetobacter baumannii, E. coli, Klebsiella pneumoniae, Pseudomonas aeruginosa, Staphylococcus aureus and Enterococcus spp. Image credits: Torres et al. , doi: 10.1038/s41564-025-02061-0.
“Previous efforts to find new antibiotics have primarily seen fungi, bacteria and animals,” said Dr. Cesar de la Fuente, a researcher at the University of Pennsylvania.
“In the past, we have used AI models to identify antibiotic candidates from all sources, from extinct organism DNA to animal venom chemicals.”
“We're now applying these tools to a new dataset: hundreds of ancient microorganism proteins.”
“There are absolutely other areas of life waiting to be explored.”
Unlike bacteria and eukaryotes (including plants, animals and fungi), Archaea occupies their own branches in the trees of life.
They resemble bacteria under a microscope, but archaeals are fundamentally different in genetics, cell membranes, and biochemistry.
These differences allow us to survive in some of the Earth's most extreme environments, from overheated seabed vents to ferocious hot springs like Yellowstone National Park.
Archaea often has few other organisms thriving, and biology has evolved in an unusual way, with little durability pressure, toxic chemicals and extreme temperatures.
This makes it a promising, but mostly unexplored source for new molecular tools, including compounds that function like antibiotics but may work differently to those currently in use.
“We were drawn to archaeals because they had to evolve biochemical defenses in unusual environments,” said Dr. Marcelo Torres of the University of Pennsylvania.
“We thought that if they had survived billions of years under these conditions, they might have developed a unique way to combat microbial competitors.
To uncover potential antibiotic compounds hidden in Archaea, researchers turned to artificial intelligence.
They leveraged an updated version of APEX, an AI tool originally developed to identify antibiotic candidates in ancient biology, including proteins from extinct animals such as wool mammoths.
With thousands of peptides (short chains of amino acids) with known antimicrobial properties, the vertices can predict the likelihood that a particular sequence of amino acids will have similar effects.
By re-adjusting Apex 1.1 for information on thousands of additional peptides and bacteria causing human disease, scientists have prepared tools to predict which peptides in archaeals may inhibit bacterial growth.
Scanning 233 archaeal species gave more than 12,000 antibiotic candidates.
The authors called these molecular molecules, and what chemical analysis revealed is different from known antimicrobial peptides (AMPs), particularly in charge distribution.
We then selected 80 archaea to test against real bacteria.
“Trying to find molecules one new antibiotic at a time is like searching for needles in a haystack,” says Famping Wang, a postdoctoral researcher at the University of Pennsylvania.
“AI speeds up the process by identifying where the needle is.”
Antibiotics work in a variety of ways. Some punch holes in the bacterial membrane, while others shut down the ability of the organism to make protein.
Researchers have discovered that unlike most known amplifiers that attack the outer defenses of bacteria, Alcaeasen pulls plugs from inside and scrambles electrical signals that keep the cells alive.
Tests for drug-resistant bacteria for various disease causes, 93% of the 80 alkadeins surveyed showed antibacterial activity against at least one bacteria.
The team then selected three alkaiersin to test in animal models.
Four days after one dose, Alkaiazen all arrested the spread of drug-resistant bacteria that are often acquired in hospitals.
One of the three compounds demonstrated activity comparable to polymyxin B, an antibiotic commonly used as the final line of protection against drug-resistant infections.
“This study shows that there are potentially many antibiotics waiting to be discovered in Archaea,” said Dr. De La Fuente.
“As more and more bacteria develop resistance to existing antibiotics, it's important to find new antibiotics in unconventional places that will replace them.”
The paper on the results was published in the journal today Natural Microbiology.
____
MDT Torres et al. Deep learning reveals antibiotics in the archaeal proteome. Nat MicrobiolPublished online on August 12, 2025. doi:10.1038/s41564-025-02061-0
