Deep learning reveals antimicrobial peptides within prions
Researchers from the University of Pennsylvania used a deep-learning platform to find antibiotic candidates in prion and prion-like proteins. These peptides, termed prionins, may offer a new way to combat drug-resistant pathogens. The study suggests proteins typically linked to neurodegenerative diseases also contain features for immune defense.
What changed
Deep learning has revealed 1,179 candidate antimicrobial peptides called prionins hidden within prion-related proteins.
Live updates
-
AI Identifies Antimicrobial Peptides Within Prion Proteins
confidence 100%Researchers from the University of Pennsylvania used a deep-learning platform to find antibiotic candidates in prion and prion-like proteins. These peptides, termed prionins, may offer a new way to combat drug-resistant pathogens. The study suggests proteins typically linked to neurodegenerative diseases also contain features for immune defense.
What's confirmed:
- Researchers used the APEX 1.1 deep-learning platform to scan 19.3 million protein fragments from 2,897 prion and prion-like proteins.
- The AI model identified 1,179 candidate antimicrobial peptides named prionins.
- Of 75 synthesized prionins, 59 inhibited bacterial pathogens and 53 perturbed membranes.
- Two prionin candidates reduced Acinetobacter baumannii infection burden in mice.
- The research was published in Nature Microbiology.