Researchers at the University of North Carolina at Charlotte used artificial intelligence to investigate how the H5N1 avian influenza virus evolves to avoid the immune system.
The H5N1 avian influenza virus is spreading to birds and mammals around the world, and the growing threat to human health is becoming a worrying prospect. As of June 2025, 70 people had been infected with H5N1 and one confirmed death in the US. Now, new research from researchers at the University of North Carolina, Charlotte (UNCC) highlights troubling evidence that the virus is evolving in ways that could make it even more dangerous to humans.
Mutant virus with potential for a pandemic
Researchers who analyzed the virus using artificial intelligence found that recent strains of H5N1 have been better able to avoid immune defenses.
“The virus certainly changed from what we saw a decade ago,” says Dr. Colby T. Ford, a computational biologist at UNCC and lead author of the study. “They don't look the same.”
The virus is certainly mutated from what we saw 10 years ago.
Ford warned that “this could be bad,” given that these changes could seriously undermine the effectiveness of existing vaccines and treatments. According to the team, vaccines developed 10 years ago may not provide sufficient protection against circulating strains of the virus today.
AI and antibody testing reveals changes in the virus
This study was published by BS, BS and Shirish Yasa of ASM Microbe 2025, the annual general meeting of the American Society of Microbiology in Los Angeles.
The team collected data on more than 1,800 H5N1 proteins and used AlphaFold 3, a cutting-edge AI protein folding tool, to predict the structure of these viral components. We then used physics-based modeling to simulate how 11 different antibodies, supplied from both humans and mice, bind to viral proteins.
“As we reach the new isolates we see, antibody performance is declining,” Ford said. As the virus developed, its ability to bind with protective antibodies decreased, weakening the natural defenses of the human immune system.
Tracking transmission paths using big data
Beyond antibody analysis, researchers have used large datasets to map how different viral clades spread. “We see that there are clear clades with very different paths when it comes to transmissions between hosts,” Ford explained.
One of the team's findings traced the recent death of H5N1 in Louisiana to a viral clade that can jump directly from birds to humans. This direct transmission shows a worrying development in the evolution of the virus.
Hope for high-tech solutions
Despite the results, researchers believe that an AI-driven approach offers new ways. Their preprint outlines how to use molecular data from new H5N1 strains to design targeted therapies that may further neutralize evolving viruses.
“Can we start generating new treatments based on these strains? The answer is yes and we can do that right away with the AI pipeline we built,” Ford said.
As H5N1 continues to adapt, integration of AI and molecular biology proves essential to responding to viruses.
