A team of scientists at the University of Virginia has developed a breakthrough approach to drug design that could fundamentally change how new medicines are created. Researchers led by Nikolay V. Dokholyan, PhD, have created a suite of artificial intelligence-powered tools called YuelDesign, YuelPocket, and YuelBond that work together to transform drug discovery by accounting for something previous methods ignored: proteins don't stay still .

Why Does It Matter That Proteins Move?

For decades, drug designers have approached their work like a locksmith crafting a key for a lock that never changes shape. But that's not how biology works. When a drug molecule approaches its target protein in the human body, the protein shifts and flexes, a phenomenon scientists call "induced fit." Most existing AI tools treat proteins as frozen statues, ignoring this critical movement. This mismatch between computer predictions and biological reality is a major reason why nearly 90% of new drugs fail when they reach human testing .

"Think of it this way: Other methods try to design a key for a lock that's sitting perfectly still, but in your body, that lock is constantly jiggling and changing shape. Our AI designs the key while the lock is moving, so the fit is much more realistic," said Dokholyan, of UVA's Department of Neurology.

Nikolay V. Dokholyan, PhD, Department of Neurology, University of Virginia

The financial stakes are enormous. Developing a single new drug costs an estimated $2.6 billion or more, and the failure rate means most of that investment yields nothing. By designing drugs that account for protein flexibility, Dokholyan's team believes they can dramatically improve success rates and reduce costs .

How Do These New AI Tools Actually Work?

The three tools work in concert, each solving a different piece of the drug design puzzle. YuelDesign, the centerpiece, uses a cutting-edge form of artificial intelligence called diffusion models to generate drug molecules tailored to fit their protein targets. Unlike traditional methods, YuelDesign simultaneously generates both the protein pocket structure and the small molecule that will slot into it, allowing both to adapt to each other during the design process, just as they would in the human body .

YuelPocket, a companion tool, uses graph neural networks to identify precisely where on a protein a drug should bind, even when working with predicted protein structures from existing tools like AlphaFold. YuelBond ensures that the chemical bonds in designed molecules are accurate and stable. Together, these tools address what researchers describe as a critical bottleneck in modern drug development .

The team demonstrated the effectiveness of their approach by designing molecules for CDK2, a well-known cancer-related protein. Their results showed that only YuelDesign could capture the critical structural changes that occur when a drug binds to the protein, something other AI tools missed entirely .

Steps to Accelerate Drug Discovery With Flexible Protein Design

• Account for Protein Dynamics: Design drug molecules while treating proteins as flexible, dynamic structures rather than rigid snapshots, capturing how proteins actually behave in the human body.
• Identify Binding Sites Precisely: Use advanced neural networks to map exactly where on a protein a drug should attach, improving the likelihood that the drug will work as intended.
• Validate Chemical Bonds: Ensure that the chemical bonds in designed molecules are accurate and stable, preventing unwanted side effects from molecular instability.
• Democratize Access: Make drug discovery tools freely available to the scientific community worldwide, enabling researchers to tackle diseases that matter most to their patients.

What makes this development particularly significant is Dokholyan's commitment to accessibility. The research team has made all of their tools freely available to the scientific community, with no financial interest in the work. This democratization approach means researchers anywhere in the world can use YuelDesign, YuelPocket, and YuelBond to tackle diseases that matter most to their patients .

What Conditions Could Benefit Most From This Technology?

Dokholyan and his colleagues believe their approach could make a real difference for patients with cancer, neurological disorders, and many other conditions where current treatments hit dead ends. The ability to design drugs that target "wiggly proteins" more accurately opens doors to treating diseases that have resisted traditional drug development approaches .

The research has been published in peer-reviewed scientific journals including PNAS, JCIM, and Science Advances, with support from the National Institutes of Health, the National Science Foundation, the Huck Institutes of the Life Sciences, and the Passan Foundation. The team includes Wang, Dong Yan Zhang, Shreshty Budakoti, and Dokholyan .

While these tools represent a major advance in computational drug design, they're not a complete replacement for traditional drug development. Rather, they accelerate the early stages of discovery and design, potentially reducing the time and cost required to identify promising drug candidates before they move into laboratory and clinical testing. For patients waiting for treatments to diseases like cancer and neurological disorders, that acceleration could mean the difference between hope and despair .