Researchers at the University at Buffalo are using artificial intelligence and advanced lab techniques to identify viruses that live harmlessly inside healthy people but remain largely unknown to science. A four-year project funded by a $1.6 million grant from the U.S. National Institutes of Health (NIH) aims to map the human virome, the collection of viruses that inhabit the body without causing disease.

What Is the Human Virome and Why Does It Matter?

Most people think of viruses as disease-causing invaders like influenza or COVID-19. But the reality is far more complex. Healthy humans carry trillions of viruses that coexist peacefully with their bodies, yet science knows surprisingly little about them. Unlike bacteria, which have been studied extensively, viruses are far more diverse and harder to detect using conventional methods.

"Healthy humans are full of trillions of viruses. Studying them is challenging because they are far more diverse than bacteria, and it can be hard to determine which viruses are present and what they are doing in the body," said Kathryn Kauffman, assistant professor in the Department of Oral Biology at the University at Buffalo School of Dental Medicine.

Kathryn Kauffman, Assistant Professor, Department of Oral Biology, University at Buffalo School of Dental Medicine

The project is part of the NIH Human Virome Program, a nationwide consortium of more than 300 researchers across the United States. The initiative has awarded more than $100 million to institutions studying the viruses that make up the healthy human virome and developing new tools to characterize them.

How Are Researchers Using AI to Detect These Hidden Viruses?

The key innovation lies in applying artificial intelligence models that work similarly to large language models (LLMs), the technology behind generative AI systems like ChatGPT. Instead of learning human language, these protein language models learn the "language" of viral protein sequences. This allows researchers to recognize patterns in viral proteins and identify viruses that conventional studies often miss.

The University at Buffalo team combines multiple cutting-edge approaches to maximize detection. Jun Qu, a professor in the Department of Pharmaceutical Sciences, uses next-generation mass spectrometry to detect viral proteins present in extremely small quantities. Yinyin Ye, an assistant professor in the Department of Civil, Structural and Environmental Engineering, contributes expertise in concentrating and purifying viruses from wastewater.

"By integrating cutting-edge proteomics with AI-driven bioinformatics approaches, we hope to uncover previously hidden viruses and gain new insights into their biological functions in healthy humans," said Jun Qu, professor in the Department of Pharmaceutical Sciences.

Jun Qu, Professor, Department of Pharmaceutical Sciences, University at Buffalo

Rather than relying solely on clinical samples from patients, the researchers use wastewater, which contains a diverse mix of viruses shed by healthy populations. The team plans to remove the dominant viruses that typically overwhelm laboratory analyses, allowing rarer viruses to be detected through DNA and RNA sequencing and protein analysis.

Steps to Identifying Unknown Viruses in the Human Virome

• Sample Collection: Researchers gather wastewater samples that contain viruses shed by healthy populations, providing a diverse viral landscape for study.
• Viral Concentration: The team removes dominant viruses that typically mask rarer ones, allowing hidden viruses to become detectable through advanced purification techniques.
• AI-Powered Analysis: Protein language models trained on viral sequences recognize patterns and identify previously unknown viruses that conventional methods would miss.
• Multi-Method Verification: Findings are confirmed using DNA and RNA sequencing, mass spectrometry, and protein analysis to ensure accuracy.

What Could These Discoveries Mean for Public Health?

The implications extend far beyond basic research. A better understanding of the viruses associated with healthy people could eventually improve wastewater-based surveillance systems, helping researchers identify unusual viral patterns that may signal emerging disease outbreaks.

Kauffman emphasized the scale of potential discovery. By the end of the four-year project, the team may identify hundreds of new types of viruses that scientists have never been able to detect before. This expanded knowledge of the human virome could reshape how public health officials monitor population health and prepare for emerging threats.

"By the end of this project, we may identify hundreds of new types of viruses that we haven't been able to detect before," said Kathryn Kauffman.

Kathryn Kauffman, Assistant Professor, Department of Oral Biology, University at Buffalo School of Dental Medicine

The project represents a broader shift in how researchers approach biological discovery. By combining AI models with traditional laboratory science, the team demonstrates that artificial intelligence is not replacing human expertise but rather amplifying it, allowing scientists to detect patterns and relationships that would be impossible to spot manually. As the field of AI-driven biology matures, projects like this one suggest that the next major breakthroughs in understanding human health may come from teaching machines to read the molecular language of life itself.