In the heart of New York, a quiet revolution is underway, one that is reshaping our fight against some of humanity's most daunting diseases.
Explore the DiscoveryThink of the human body as a city of billions of cells. Within each, proteins—the workforce of life—are constantly in motion, folding, shifting, and interacting in a complex dance. For decades, scientists could only capture blurry snapshots of these vital players. When drugs failed, it was often because they were designed for a static image, not the dynamic reality. A breakthrough at the New York State Center of Excellence in Bioinformatics & Life Sciences (CBLS) is changing that, offering a transformative new way to see this dance in action 1 .
The CBLS is not merely a facility; it is a unique collaboration of Western New York's most prominent research institutions. With the University at Buffalo as the lead, and partners like Roswell Park Comprehensive Cancer Center and the Hauptman-Woodward Medical Research Institute, the CBLS was established with a clear mission: to study the mechanistic processes of human disease and develop new diagnostic tools and therapeutic interventions 7 .
Research partner; focuses on translating bioinformatics discoveries into cancer treatments and personalized therapies 2 .
Research partner; contributes expertise in structural biology to understand disease mechanisms 7 .
This partnership creates a powerful ecosystem where fundamental biological research directly informs clinical application. The state-of-the-art, 135,600-square-foot facility serves as a physical hub, but the true work happens in the seamless flow of ideas and data between computational scientists, biologists, and clinicians 7 .
The core of the challenge lies with proteins. For years, tools like the Nobel Prize-winning AlphaFold have performed miracles, using AI to predict a protein's static structure from its amino acid sequence 1 . This was a monumental leap forward. However, it has a significant limitation.
"AlphaFold doesn't know which shape the protein actually has under your specific experimental conditions," explains Dr. Thomas Grant, a researcher at UB and CBLS. "Proteins are dynamic and can adopt many different shapes. Drugs need to bind to the actual shape the protein has in a person's body, not just any possible shape it could have." 1
This gap in understanding is a major reason why many promising drug candidates fail in later stages of development. Scientists were designing keys (drugs) for a lock (protein) they had only ever seen in a single, rigid photograph, when in reality, the lock was constantly changing form.
Traditional methods capture proteins in fixed positions, missing their natural movements and flexibility.
Proteins in the body are constantly moving and changing shape, requiring new approaches to visualization.
Dr. Grant and his team have pioneered a solution called SWAXSFold, an AI-powered tool supported by a $2.18 million grant and the immense computing power of the New York State-based Empire AI consortium 1 .
So, how does it work? The "SWAXS" part stands for small- and wide-angle X-ray scattering. Imagine shining a bright light through a protein solution. The way the light scatters creates a unique pattern, like a fingerprint that contains information about the protein's size, shape, and internal structure as it exists in its natural, fluid environment—no freezing or crystallizing required 1 .
The "Fold" part is the AI revolution. The team is integrating these experimental SWAXS patterns directly into the AI training process used by systems like AlphaFold. Instead of asking the AI, "What might this protein look like?" they give it the protein sequence plus the experimental data and ask, "What does this protein actually look like in solution?" 1 This integration of real-world data with powerful AI prediction is a first-of-its-kind approach.
Proteins are dissolved in a water-based solution at room temperature to mimic the protein's natural environment within human cells 1 .
X-rays are shone through the protein solution, and detectors measure the scattering patterns at small and wide angles to capture a fingerprint of the protein's dynamic, three-dimensional structure 1 .
The scattering data is fed directly into a modified AlphaFold AI algorithm alongside the protein's amino acid sequence to guide the AI to generate a structural model that perfectly fits the experimental data from the solution 1 .
The resulting model is analyzed for movement, flexibility, and potential drug-binding sites to understand the protein's functional form and identify precise targets for drug design 1 .
Static, single conformation
Limited understanding of dynamics
Higher drug failure rates
Dynamic, multiple conformations
Real-time structural changes
Higher precision drug targeting
A project of this scale relies on a sophisticated suite of computational tools and resources. The bioinformatics teams within the CBLS network leverage state-of-the-art infrastructure to turn raw data into discovery 2 .
| Tool / Resource | Category | Function in Research |
|---|---|---|
| High-Performance Computing (HPC) Clusters | Computing Infrastructure | Provides the massive processing power required for training AI models like SWAXSFold and analyzing large genomic datasets 1 2 . |
| AlphaFold | AI Software | Uses deep learning to predict potential protein structures from amino acid sequences; the foundation upon which SWAXSFold builds 1 . |
| Next-Generation Sequencing (NGS) | Data Generation | Technologies that rapidly sequence DNA and RNA, providing the genetic blueprint that codes for proteins 2 . |
| Bioconductor | Data Analysis | A statistical programming platform (based on R) used for the analysis and comprehension of high-throughput genomic data . |
| AutoDock Vina | Molecular Modeling | A widely cited software tool used for molecular docking and virtual screening, helping predict how small molecules, like drug candidates, bind to a protein target . |
| EMBOSS | Software Suite | A comprehensive collection of open-source analysis tools specially developed for the molecular biology and bioinformatics user community . |
High-performance CPUs and GPUs for complex calculations
Massive storage systems for genomic and structural data
High-speed connections for data transfer between institutions
Specialized hardware for training and running AI models
The implications of this research are profound. By revealing the true, dynamic shapes of proteins, SWAXSFold and related technologies at CBLS open the door to a new era of medicine.
For cancer patients, it could lead to drugs that effectively target cancer-causing proteins that were previously considered "undruggable" because of their shape-shifting nature 1 .
For those with neurological diseases like Alzheimer's, it provides a window into the protein misfolding that drives disease progression.
Furthermore, this approach is a critical step toward personalized medicine 1 .
"We're also developing tools that will help researchers understand how disease-causing mutations change protein structure," says Dr. Grant. "If we can see exactly how a mutation alters a protein's shape and function, we can design personalized therapies targeted to that specific change." 1
This work, happening at the intersection of biology, computing, and collaboration, is a powerful testament to the mission of the NYS Center of Excellence in Bioinformatics and Life Sciences. It is not just about building better tools; it is about building a healthier future for all.
First protein structures determined using X-ray crystallography
Development of NMR spectroscopy for studying proteins in solution
Advancements in cryo-electron microscopy for larger complexes
Rise of computational methods and molecular dynamics simulations
AlphaFold revolutionizes protein structure prediction
SWAXSFold integrates experimental data with AI for dynamic visualization