AI Uncovers Hundreds of Potential Antibiotics in Animal Venom, Offering New Hope Against Antibiotic Resistance

AI finds hundreds of potential antibiotics in snake and spider venom

Snake, scorpion, and spider venom are most frequently associated with poisonous bites, but with the help of artificial intelligence, they might be able to help fight antibiotic resistance, which contributes to more than one million deaths worldwide each year. In a study published in Nature Communications, researchers at the University of Pennsylvania used a deep-learning system called APEX to sift through a database of more than 40 million venom encrypted peptides (VEPs), tiny proteins evolved by animals for attack or as a defense mechanism. In a matter of hours, the algorithm flagged 386 compounds with the molecular hallmarks of next-generation antibiotics. "Venoms are evolutionary masterpieces, yet their antimicrobial potential has barely been explored," said senior author César de la Fuente, Ph.D., a Presidential Associate Professor of Psychiatry, Microbiology, Bioengineering, Chemical and Biomolecular Engineering, and Chemistry. "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." Combining emerging tech with established methods From the AI-selected shortlist, the team synthesized 58 venom peptides for laboratory testing. 53 killed drug-resistant bacteria -- including Escherichia coli and Staphylococcus aureus -- at doses that were harmless to human red blood cells. "By pairing computational triage with traditional lab experimentation, we delivered one of the most comprehensive investigations of venom derived antibiotics to date," added co-author Marcelo Torres, Ph.D., a research associate at Penn. Changge Guan, Ph.D., a postdoctoral researcher in the De la Fuente Lab and co-author, noted that the platform mapped more than 2,000 entirely new antibacterial motifs -- short, specific sequences of amino acids within a protein or peptide responsible for their ability to kill or inhibit bacterial growth. The team is now taking the top peptide candidates which could lead to new antibiotics and improving them through medicinal-chemistry tweaks.

AI found a way to turn venom into antibiotics - Earth.com

Antimicrobial resistance now kills at least 1.27 million people a year and is linked to almost five million deaths, making common infections progressively harder to treat. Meanwhile, the pipeline for new antibiotics has slowed, but animal venom offers new hope amid fears of a post‑antibiotic era. Using deep-learning tools, researchers mined 40 million venom‑encoded peptides collected from snakes, scorpions, spiders, and other animals. They pinpointed 386 molecules with the hallmarks of powerful antibiotics. The work was led by César de la Fuente at the University of Pennsylvania, who specializes in machine biology and antibiotic discovery. Decades ago, a peptide from the Brazilian pit viper inspired captopril, the first ACE inhibitor for hypertension, proving that toxins can be tamed for therapy. More recently, FDA‑approved ziconotide, derived from cone snail venom, has provided severe pain relief without opioids. Those successes hint at a vast pharmacological trove hidden in venoms, yet only a fraction of the peptides they contain have been examined. Digital screening can accelerate that search beyond what manual experiments allow. Venomous animals have evolved specialized chemistries for offense or defense over millions of years. These chemistries generate peptides that slice into nervous systems, coagulation pathways, and bacterial membranes. Because each species fine‑tunes its arsenal, the combined library covers many chemical shapes rarely found in conventional drug collections. The Penn team compiled proteins from four databases, representing cone snails, spiders, snakes, scorpions, and more. This diversity fed the AI with sequences spanning more than 16,000 proteins. The model then shredded these sequences into over 40 million candidate peptides. APEX turns each animal venom peptide's sequence into numbers that reflect traits like charge and water-repellence. Then, it predicts how much of each peptide is needed to stop the growth of 34 different bacterial strains. Sequences predicted to inhibit growth at 32 micromoles per liter or less advanced to a similarity filter that weeded out peptides too close to known antibiotics. The filter left 386 candidates that occupied unexplored corners of peptide sequence space. This ensured any hits would add new chemical diversity. Crucially, the whole triage ran in a few hours on standard graphics‑processing hardware, a pace impossible for wet‑lab screening. "Venoms are evolutionary masterpieces, yet their antimicrobial potential has barely been explored," said César de la Fuente. His lab's strategy turns that evolutionary advantage into a search engine for urgently needed drugs. The platform also found over 2,000 new short sequence patterns that help break bacterial membranes -useful building blocks for future antibiotics. The researchers synthesized 58 of the AI‑ranked peptides and challenged them against drug‑resistant strains of Escherichia coli and Staphylococcus aureus; 53 wiped out the bacteria at doses that left human red blood cells unharmed. "By pairing computational triage with traditional lab experimentation, we delivered one of the most comprehensive investigations of venom‑derived antibiotics to date," said Marcelo Torres, a research associate at Penn. Torres noted that the most potent peptides came from spiders, mirroring the predators' need to immobilize prey quickly. One standout candidate knocked down Acinetobacter baumannii infections in mice by up to 99 percent after a single topical dose. The animals showed no weight loss or other toxicity signs, an early but encouraging safety signal. Most classic antibiotics block enzymes, so single genetic tweaks can disarm them, but venom peptides that destabilize bacterial membranes face fewer resistance routes. Animal venom hits carried strong positive charges and balanced hydrophobic residues, a combination that latches onto the negatively charged bacterial surface and punches holes in it. Fluorescence assays confirmed that 26 peptides swiftly collapsed the cytoplasmic membrane potential in Pseudomonas aeruginosa. Outer membrane permeabilization played a smaller role, suggesting depolarization is their primary kill step. This mechanism works like human defensins but uses new sequences that bacteria haven't seen, making resistance unlikely - and because human cells have cholesterol and a neutral charge, they aren't harmed in the same way. Testing in human kidney cells found that most scorpion and cone snail peptides were benign at concentrations far above their antibacterial doses. A handful of spider peptides showed cytotoxicity, guiding the team to prioritize safer scaffolds. In the mouse skin‑infection model, the lead spider peptide cut bacterial counts by roughly three orders of magnitude without harming the animals. Pharmacokinetic tweaks such as end‑capping or non‑natural amino acids could further prolong its activity and reduce dosing frequency. Computational screens suggest 40 percent of candidates likely avoid potassium channels, a common off-target for cationic peptides. Electrophysiology studies now under way will test that prediction before systemic trials begin. Medicinal chemists are refining the top animal venom hits to sharpen selectivity, resist protease degradation and improve serum half‑life. Machine‑learning feedback loops will retrain APEX with every new data point, steadily raising prediction accuracy. Regulatory hurdles remain, yet the researchers argue that topical use in skin or burn infections could reach patients faster than systemic applications. Researchers can also pair peptide antibiotics with existing drugs to delay resistance evolution. The goal is a pipeline where venom data trains artificial intelligence, AI guides chemistry, and chemistry delivers antibiotics in months. If realized, that loop could turn the tide in the struggle against superbugs. Like what you read? Subscribe to our newsletter for engaging articles, exclusive content, and the latest updates.

AI Finds Hundreds of Potential Antibiotics in Snake and Spider Venom | Newswise

Newswise -- PHILADELPHIA - Snake, scorpion, and spider venom are most frequently associated with poisonous bites, but with the help of artificial intelligence, they might be able to help fight antibiotic resistance, which contributes to more than one million deaths worldwide each year. In a study published in Nature Communications, researchers at the University of Pennsylvania used a deep-learning system called APEX to sift through a database of more than 40 million venom encrypted peptides (VEPs), tiny proteins evolved by animals for attack or as a defense mechanism. In a matter of hours, the algorithm flagged 386 compounds with the molecular hallmarks of next-generation antibiotics. "Venoms are evolutionary masterpieces, yet their antimicrobial potential has barely been explored," said senior author César de la Fuente, PhD, a Presidential Associate Professor of Psychiatry, Microbiology, Bioengineering, Chemical and Biomolecular Engineering, and Chemistry. "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." From the AI-selected shortlist, the team synthesized 58 venom peptides for laboratory testing. 53 killed drug-resistant bacteria -- including Escherichia coli and Staphylococcus aureus -- at doses that were harmless to human red blood cells. "By pairing computational triage with traditional lab experimentation, we delivered one of the most comprehensive investigations of venom derived antibiotics to date," added co-author Marcelo Torres, PhD, a research associate at Penn. Changge Guan, PhD, a postdoctoral researcher in the De la Fuente Lab and co-author, noted that the platform mapped more than 2,000 entirely new antibacterial motifs -- short, specific sequences of amino acids within a protein or peptide responsible for their ability to kill or inhibit bacterial growth. The team is now taking the top peptide candidates which could lead to new antibiotics and improving them through medicinal-chemistry tweaks. Support included funding from the Procter & Gamble Company, United Therapeutics, a BBRF Young Investigator Grant, the Nemirovsky Prize, Penn Health-Tech Accelerator Award, Defense Threat Reduction Agency grants HDTRA11810041 and HDTRA1-23-1-0001, and the Dean's Innovation Fund from the Perelman School of Medicine at the University of Pennsylvania. Research reported in this publication was supported by the Langer Prize (AIChE Foundation), the NIH R35GM138201, and DTRA HDTRA1-21-1-0014. Cesar de la Fuente provides consulting services to Invaio Sciences and is a member of the Scientific Advisory Boards of Nowture S.L. and Phare Bio. The de la Fuente Lab has received research funding or in-kind donations from United Therapeutics, Strata Manufacturing PJSC, and Procter & Gamble, none of which were used in support of this work. An invention disclosure associated with this work has been filed. ### Penn Medicine is one of the world's leading academic medical centers, dedicated to the related missions of medical education, biomedical research, excellence in patient care, and community service. The organization consists of the University of Pennsylvania Health System and Penn's Raymond and Ruth Perelman School of Medicine, founded in 1765 as the nation's first medical school. The Perelman School of Medicine is consistently among the nation's top recipients of funding from the National Institutes of Health, with $580 million awarded in the 2023 fiscal year. Home to a proud history of "firsts" in medicine, Penn Medicine teams have pioneered discoveries and innovations that have shaped modern medicine, including recent breakthroughs such as CAR T cell therapy for cancer and the mRNA technology used in COVID-19 vaccines. The University of Pennsylvania Health System's patient care facilities stretch from the Susquehanna River in Pennsylvania to the New Jersey shore. These include the Hospital of the University of Pennsylvania, Penn Presbyterian Medical Center, Chester County Hospital, Doylestown Health, Lancaster General Health, Penn Medicine Princeton Health, and Pennsylvania Hospital -- the nation's first hospital, founded in 1751. Additional facilities and enterprises include Good Shepherd Penn Partners, Penn Medicine at Home, Lancaster Behavioral Health Hospital, and Princeton House Behavioral Health, among others. Penn Medicine is an $11.9 billion enterprise powered by more than 48,000 talented faculty and staff.