Penn Engineers Create AI Tool to Speed Antibiotic Discovery
The race against antibiotic resistance is on, and researchers at the University of Pennsylvania have just made a significant stride forward. They've developed ApexGO, an AI-powered tool that revolutionizes the way we discover and improve antibiotic candidates. This innovative approach could be a game-changer in the fight against antibiotic resistance.
A New Approach to Antibiotic Discovery
ApexGO takes a unique approach to antibiotic discovery, moving away from traditional methods that rely on screening large databases. Instead, it starts with a small number of imperfect candidates and improves them step by step. This iterative process is guided by a predictive algorithm that evaluates each modification and guides the next step, much like a skilled navigator charting a course through an uncharted territory.
"Antibiotic discovery is fundamentally a search problem across an enormous molecular space," says César de la Fuente, Presidential Associate Professor in Bioengineering and Chemical and Biomolecular Engineering. "ApexGO gives us a way to navigate that space with far more direction."
From Imperfect to Potent
ApexGO begins with a promising but imperfect peptide, a short string of amino acids. It then proposes precise edits, predicts whether those changes are likely to enhance antimicrobial activity, and guides the next round of edits. This process continues until versions that are more likely to work are created.
Laboratory tests against disease-causing bacteria supported ApexGO's predictions. 85% of the AI-generated molecules halted bacterial growth, and 72% outperformed the peptides from which they were derived. In mice, two antimicrobial peptides created by ApexGO reduced bacterial counts at levels comparable to polymyxin B, an FDA-approved antibiotic.
A Systematic Search
Until now, antibiotics have largely been found by accident, with penicillin being the most famous example. ApexGO points to a more systematic way forward, allowing researchers to search the vast space of all possible antimicrobial peptides computationally.
"We ran ApexGO for a few months and found hundreds of candidates," says Jacob R. Gardner, Assistant Professor in Computer and Information Science. "If we ran that process for a year, how many thousands of these could we find?"
Future Directions and Beyond Antibiotics
While the molecules proposed by ApexGO showed promising antibiotic activity, further optimization is needed before they can be used to treat infections in humans. However, the study suggests that AI can help researchers decide which molecules are worth making and testing.
"In this case, we wanted to optimize peptides for antimicrobial activity," says de la Fuente. "But you could imagine applying the same idea to peptides with other biological functions, like modulating the immune system or targeting tumors."
ApexGO is an important step toward a future where AI can help us move faster from an idea to a real therapeutic candidate, especially in the face of rising antibiotic resistance worldwide.
