On University at Buffalo Week: Speeding up drug discovery will have many benefits.
Tom Grant, assistant professor in the Department of Structural Biology in the Jacobs School of Medicine and Biomedical Sciences, looks into doing so.
Faculty Bio:
Thomas Grant is a structural biologist whose research focuses on developing innovative analytical methods to uncover the structure and dynamics of biological macromolecules.
Trained at the University at Buffalo, where he earned a BS in Mathematical Physics (2007) and a PhD in Structural and Computational Biology (2013), Grant’s research combines rigorous quantitative training with experimental expertise in X-ray crystallography, small-angle scattering, and X-ray free electron laser methodologies.
His work integrates advanced computational algorithm development with cutting-edge experimental tools to generate high-resolution structural models and time-resolved “molecular movies,” illuminating how proteins and nucleic acids function within the cell and informing rational drug design.
In 2025, he was awarded a $2.18 million grant from the National Institute of General Medical Sciences, part of the National Institutes of Health, to create a new artificial intelligence-powered tool that improves scientists’ understanding of how proteins move and change shape within the human body.
Transcript:
In the United States, it often takes 10 to 15 years to develop a new drug from its initial discovery to market approval. In my lab, we’re working on an artificial intelligence-driven project called SWAXSFold that we hope will dramatically speed up this process.
So let me explain what SWAXSFold is.
A few years ago, Google DeepMind released an algorithm called AlphaFold. It has had a major impact in biomedical research, winning the 2024 Nobel Prize in Chemistry. AlphaFold predicts 3D protein structures. By identifying these structures, scientists can target them with drugs that treat disease.
But the reality is that many proteins are dynamic. They move around and can have many different shapes – that’s how they function and operate. This limits AlphaFold’s usefulness for drug discovery.
That’s where our expertise in “SWAXS” comes in. SWAXS stands for small- and wide-angle X-ray scattering.
SWAXS uses X-rays to take snapshots of proteins as they move around and change shape. It can see the different conformations a protein adopts, which is exactly what AlphaFold can’t do. So we’re developing SWAXSFold to integrate SWAXS data directly into this structure prediction process.
As you can imagine, this requires intense computational resources.
Our team is utilizing New York State’s supercomputing facility, Empire AI, to train SWAXSFold with our database of nearly a half million protein structures.
As SWAXSFold learns from AI, it will give us unparalleled insight on how proteins change their shape, function and operate. In turn, this will improve our ability to predict protein structures and, thus, make it much easier and quicker to identify drug targets that treat everything from cancer to Alzheimer’s disease.
Read More:
[UBuffalo] - Researcher’s AI-powered drug discovery tool supported by Empire AI
