Innovative AI Method Reveals Antiviral Compounds Using Limited Data

Researchers at the Perelman School of Medicine at the University of Pennsylvania have developed a novel artificial intelligence (AI) approach that significantly accelerates the discovery of antiviral drugs, even when only limited experimental data are available. This breakthrough combines traditional laboratory techniques with advanced machine learning models to identify promising drug candidates against human enterovirus 71 (EV71), the primary cause of hand, foot, and mouth disease.

In their recent study published in Cell Reports Physical Science, the team demonstrated that their AI system could reliably predict which small molecules might inhibit EV71 by analyzing a small panel of just 36 compounds. The model assessed chemical structures and features to score each molecule's potential to block the virus. Remarkably, when the researchers tested the AI-selected compounds in laboratory cell experiments, five out of eight effectively slowed virus activity, yielding a hit rate roughly ten times higher than traditional screening methods.

"This approach drastically reduces the time and resources required for antiviral discovery," said lead researcher de la Fuente. "What once took months can now be accomplished in just days, especially valuable in urgent outbreak situations where data collection is constrained."

The confirmed antiviral compounds were further examined through computer simulations, revealing that they targeted specific sites on the virus. These insights could guide future efforts to prevent the virus from changing shape or entering host cells, thus impeding infection.

EV71 infection severity ranges from mild rashes and fever to serious neurological complications, primarily affecting young children and immunocompromised individuals. Currently, there are no FDA-approved antivirals targeting EV71, underscoring the importance of this rapid discovery method.

This innovative technique not only expedites antiviral drug development but also offers a template adaptable to other emerging viruses, including respiratory pathogens or reemerging diseases like polio. The collaboration involved industry leaders like Procter & Gamble and academic partners such as Cornell University, illustrating a successful synergy of multidisciplinary efforts.

As Angela Cesaro, Ph.D., a project co-author, stated, "Our AI-driven method can be a game-changer for rapid response to future viral threats. Even with limited data, machine learning can accelerate the development of effective therapies and improve preparedness."

This advancement highlights the growing potential of combining computational power with laboratory science to tackle urgent public health challenges. For more details, see the full study at source.