Penn researchers develop predictive AI model for antibiotic discovery
Penn researchers develop predictive AI model for antibiotic discovery
Publish Date: 2026-06-15 23:32:00
Source Domain: www.thedp.com
- Researchers at the School of Engineering and Applied Science have introduced an updated AI-powered model, ApexGO, to optimize the development of antibiotic candidates.
- ApexGO improves upon a previous model and can suggest modifications to existing molecules to enhance their effectiveness as antibiotics.
- Presidential Associate Professor César de la Fuente highlighted ApexGO’s ability to generate highly effective antimicrobial peptides, with 85% proving successful in real-world experiments and 72% demonstrating better performance against bacteria compared to their templates.
- Research assistant professor Marcelo Torres noted that ApexGO learns from biological sequences to prioritize candidates likely to possess antibacterial activity, focusing on promising but uncertain areas to improve search efficiency.
- Graduate student Natalie Maus emphasized that ApexGO helps narrow down the vast potential peptides for easier testing and demonstrated stronger antibacterial activity in mice than some FDA-approved last resort antibiotics.
- César de la Fuente stressed the urgency of developing new antibiotics due to the rising threat of antimicrobial-resistant infections, which cause millions of deaths globally and may escalate in the coming decades.
- De la Fuente, Maus, and Torres all underscored the potential of ApexGO to make antibiotics discovery more systematic and to apply similar methodologies to broader peptide and protein engineering challenges.
- Natalie Maus highlighted the importance of open collaboration within the scientific community to tackle the growing problem of antibiotic resistance.
