McMaster University AI Model Synthesizes Novel Antibiotic "Synthecin" to Neutralize Drug-Resistant Staph Infections

McMaster University’s AI designs "synthecin," a novel antibiotic that kills resistant staph infections, exploring 46 billion compounds to find the cure.

By: AXL Media

Published: Apr 24, 2026, 4:27 AM EDT

Source: Information for this report was sourced from McMaster University

Accelerating Discovery in the Multidrug-Resistance Era

The global crisis of antimicrobial resistance has necessitated a radical shift in how new medicines are discovered, moving away from slow, manual laboratory screens toward high-speed computational design. Researchers at McMaster University’s Michael G. DeGroote Institute for Infectious Disease Research have met this challenge by developing SyntheMol-RL, a reinforcement learning framework. This generative AI model does not merely search existing databases but explores a massive theoretical chemical space of 46 billion compounds. By utilizing 150,000 molecular building blocks and 50 validated chemical reactions, the AI functions as a digital architect, configuring "molecular Lego blocks" into entirely new antibacterial structures.

Solving the Clinical Paradox of Solubility

One of the primary hurdles in drug development is the inherent conflict between a molecule’s antibacterial potency and its water solubility. Many compounds that kill bacteria in a petri dish fail in clinical trials because they cannot dissolve in the human body or are toxic to healthy cells. Assistant Professor Jon Stokes and his team, in collaboration with Stanford University, refined SyntheMol-RL to integrate these pharmacological constraints directly into the generative process. Instead of filtering out insoluble molecules after they are designed, the AI now prioritizes solubility as a foundational requirement, ensuring that the resulting candidates are both synthesizable in the lab and viable for human metabolism.

Validation Through the Discovery of Synthecin

To test the enhanced model, the researchers tasked it with designing water-soluble agents specifically targeting Staphylococcus aureus, the pathogen responsible for notoriously resistant staph infections. From a curated list of 79 AI-proposed candidates, the team identified a particularly promising, structurally unique molecule they named synthecin. Unlike traditional antibiotics discovered through soil samples or botanical extracts, synthecin is a "born-digital" drug, optimized by AI to meet specific molecular weight and reactivity profiles before a single physical sample was ever produced in the laboratory.