AI pinpoints single weakness that stops herpes viruses cold

Washington State University researchers have identified a precise molecular target that blocks herpes viruses from entering cells — a discovery that could reshape how we design antiviral drugs.

The insight came from an unlikely collaboration: computational scientists ran thousands of molecular simulations through machine learning algorithms, then handed off their findings to lab researchers who confirmed the theory with a single targeted mutation. In one stroke, they disabled the virus's ability to fuse with and infiltrate cells.

"Viruses are very smart," says Jin Liu, the study's corresponding author and a professor in the School of Mechanical and Materials Engineering. "The whole process of invading cells is very complex, and there are a lot of interactions. But not all of them matter equally — some are just background noise, while others are absolutely critical."

The Hidden Weak Point

Herpes viruses rely on a "fusion" protein that acts like a molecular grappling hook, allowing them to merge with cell membranes and slip inside. Scientists have long struggled to understand exactly how this large, complex protein changes shape to make entry possible — which is partly why vaccines for these common viruses remain elusive.

Prashanta Dutta and Liu took a different approach. They fed their AI system thousands of potential interactions within the fusion protein, asking it to identify which amino acids (the building blocks of proteins) were actually doing the heavy lifting. The algorithm sifted through the noise and flagged a single amino acid as critical to the viral entry process.

Then came the test: Anthony Nicola's lab team introduced a targeted mutation to that one amino acid. The result was immediate and decisive. The virus could no longer fuse with cells. The infection was stopped before it could begin.

Without the AI step, Liu estimates this discovery would have taken years of trial-and-error experimentation. "Testing even a single interaction experimentally can take months," he explains. "If we'd done this by brute force, it could have taken years to find. The combination of simulation and machine learning accelerated everything."

What Comes Next

The team still has work ahead. They've confirmed that mutating this one amino acid breaks the virus's entry mechanism, but they don't yet fully understand how that small change cascades through the entire protein structure. The next phase involves using simulations to map those ripple effects — to see how a tiny molecular adjustment reshapes the whole system.

That gap between what simulations predict and what happens in living cells is where the real challenge lies. But the pathway is clearer now: find the critical interactions, disable them, and watch the virus fall apart before it can cause harm.