Researchers at Washington State University used AI and molecular simulations to identify a single critical contact within the viral fusion protein glycoprotein B (gB). Published in Nanoscale, lab alterations of that contact stopped herpesviruses from entering cells. The AI-driven approach ranked thousands of interactions to prioritize experiments, and the framework could be applied to other protein-driven diseases including Alzheimer’s.
AI Identifies a Single Molecular Weak Spot That Stops Viruses Before They Enter Cells
Researchers at Washington State University have combined artificial intelligence and molecular simulations to pinpoint a single molecular contact in a viral fusion protein that is essential for cell entry. In laboratory tests on herpesviruses, altering this contact prevented the virus from entering new cells — a potential new approach to stopping infections before they begin.
What the Study Found
Published in November in the journal Nanoscale, the study focused on viral entry, a stage of infection that is often poorly understood and difficult to disrupt. The team used machine learning and large-scale molecular simulations to evaluate and rank thousands of possible internal contacts within a key fusion protein, then tested the highest-priority candidates in the lab.
Target: Glycoprotein B (gB)
The researchers examined herpesviruses as a test case. These viruses rely on a surface fusion protein called glycoprotein B (gB) to drive the membrane fusion that allows viral genomes to enter cells. Although gB’s role in infection was known, its large size and complex architecture made it hard to identify which internal interactions are functionally critical.
"Viruses attack cells through thousands of interactions. Our research is to identify the most important one, and once we identify that interaction, we can figure out a way to prevent the virus from getting into the cell and stop the spread of disease," said Professor Jin Liu, a mechanical and materials engineering professor at Washington State University.
How AI Helped
Rather than relying on costly, time-consuming trial-and-error lab experiments, the team used simulations and machine-learning models to screen many potential interactions at once and rank them by predicted importance. This computational triage made it practical to focus experimental validation on a small set of high-value targets.
Leadership and Funding
The project was led by Professor Anthony Nicola (Veterinary Microbiology and Pathology) and Professor Jin Liu, and was supported by funding from the National Institutes of Health. The article notes broader U.S. investments in medical AI research, including a recent $50 million NIH initiative to apply AI to childhood cancer research.
Broader Implications
Beyond virology, the authors say the computational framework can be adapted to other diseases driven by altered protein interactions, such as neurodegenerative disorders like Alzheimer’s disease. By identifying the specific protein contacts that matter most, researchers can design strategies to weaken, strengthen, or block those interactions — accelerating therapeutic discovery.
Significance: The study demonstrates a practical workflow that pairs AI-driven ranking with targeted laboratory validation, offering a faster, cheaper route to find molecular targets that can block infection at the very first step.
