Scientists use AI to turn venom into life-saving medicine

In the ongoing fight against drug-resistant bacteria, an unlikely hero is emerging—venom. Once known only for its deadly sting, venom from snakes, scorpions, and spiders is now being explored as a powerful weapon against superbugs. A research team from the University of Pennsylvania has taken a bold step by combining artificial intelligence with molecular science to unlock venom's hidden potential.

Their work, recently published in Nature Communications, dives into a global dataset of venom proteins, pulling back the curtain on an overlooked source of promising new antibiotics. The results are not only exciting—they may offer a fresh approach to tackling one of the world’s deadliest medical threats.

Drug-resistant infections are a growing crisis. Each year, they contribute to roughly 5 million deaths. The situation is becoming more serious, especially as the development of new antibiotics has slowed to a near standstill. Traditional methods of discovering antibiotics are no longer enough to keep pace with the evolution of resistant bacteria.

One of the most dangerous types of these bacteria is known as gram-negative bacteria. These microbes have a tough outer shell that makes them particularly hard to kill with standard drugs. Scientists have had limited success finding effective treatments against them.

But nature may offer a surprising solution. Venoms have evolved over millions of years as powerful biological tools, packed with small proteins called peptides. These peptides are designed to kill or disable prey, often by attacking cells in ways that mirror how antibiotics work. Despite their potential, venom peptides have rarely been studied for their medical benefits—until now.

To uncover these hidden gems, the Penn researchers turned to a deep-learning system called APEX. This powerful algorithm was trained to scan and analyze the properties of venom proteins with precision and speed no human could match. In a matter of hours, APEX sorted through 16,123 venom proteins and broke them down into more than 40 million smaller peptide fragments, known as venom-encrypted peptides.

Out of this massive database, APEX selected 386 peptides that showed promising signs of being able to fight bacteria. What made them stand out was their structure. These peptides had a high net charge and strong hydrophobic properties, which are key features for penetrating bacterial membranes. Their flexible shapes allowed them to fold into α-helices, a common shape seen in many natural antimicrobial peptides that destroy bacteria by disrupting their protective layers.

“Venoms are evolutionary masterpieces, yet their antimicrobial potential has barely been explored,” said Dr. César de la Fuente, who led the study. “APEX lets us scan an immense chemical space in just hours and identify peptides with exceptional potential to fight the world’s most stubborn pathogens.”

The research didn’t stop with predictions. The team synthesized 58 of the top peptide candidates to test them in the lab. These synthetic venom peptides were put up against several strains of bacteria, including some known to resist current antibiotics—like Escherichia coli and Staphylococcus aureus.

The results were striking. Fifty-three of the peptides killed the bacteria effectively, all while leaving human red blood cells unharmed. This is a crucial sign that the peptides might be safe for use in medicine. “By pairing computational triage with traditional lab experimentation, we delivered one of the most comprehensive investigations of venom-derived antibiotics to date,” said Dr. Marcelo Torres, a co-author of the study.

Dr. Changge Guan, another co-author and a postdoctoral researcher in the De la Fuente Lab, explained that their platform didn’t just find effective peptides. It also mapped over 2,000 completely new antibacterial motifs—short chains of amino acids responsible for antimicrobial action.

To see how the peptides worked in living organisms, researchers tested several of them in mice infected with Acinetobacter baumannii, a dangerous gram-negative bacterium known for causing hospital-acquired infections. The peptides significantly reduced bacterial levels in the infected mice. Even better, the treatment showed no signs of toxicity, which is often a major concern when testing new drug candidates.

These results bring hope that venom peptides can be shaped into real, usable antibiotics for people. However, the journey isn’t over. The team is now refining the most promising candidates by adjusting their chemical structures to improve their stability and effectiveness—an approach known as medicinal chemistry.

This study stands out not just for what it discovered, but for how it did it. By blending AI with biology, the team has opened a new frontier in drug discovery. Instead of spending years combing through molecules by hand, researchers now have tools to quickly identify and test new options. And in the case of venom, those options seem especially rich.

These findings also highlight how nature can still surprise us. Venom, once feared for its deadly power, may become a life-saving tool. As antibiotic resistance continues to grow, exploring nature’s hidden chemical libraries could become key to survival. With further research, the next generation of antibiotics might not come from a lab dish or a soil sample—but from the fangs and stingers of creatures we once ran from.

Note: The article above provided above by The Brighter Side of News.

Like these kind of feel good stories? Get The Brighter Side of News' newsletter.