LOGAN, Utah — As humans, our eyes receive two-dimensional images, which our brains convert into a three-dimensional experience. This ability allows us to recognize our position in space, judge distance, have depth perception, and enjoy visually examining all kinds of objects and events.
But trying to imagine invisible structures and higher-dimensional processes that cannot be captured by human-designed scopes is challenging for data scientists and visualization professionals who rely on machine learning and AI tools to enhance visual exploration.
“Biological processes are an example of complex high-dimensional data,” says Kevin Moon, director of USU’s Data Science and Artificial Intelligence (DSAI) Center and associate professor in the Department of Mathematics and Statistics. “For example, one of the datasets we use to test our AI tools is clinical data measured from multiple sclerosis patients. These datasets contain hundreds of thousands of data points about disease progression at the cellular level, as well as treatments and clinical outcomes.”
Moon is the corresponding author of the paper “Gaining Biological Insights through Supervised Data Visualization,” published online June 30 in Nature Computational Science. The paper was co-published with lead author and USU alumnus Jake Rhodes (22nd year doctoral student, statistics), assistant professor at Brigham Young University, Moon’s USU colleagues Adele Cutler, professor emeritus in the Department of Mathematics and Statistics, and Yasuhiro Chou, professor in the Department of Biochemical Engineering. Along with USU alumnus and University of Utah researcher Wei Zhang (PhD’21, bioengineering).
The team’s research is supported by the National Institutes of Health and the IVADO Visiting Scholars Program, and includes additional national and international collaborators. *
“In this paper, we introduce RF-PHATE, an acronym for Random Forest-Potential of Heat-diffusion for Affinity-based Trajectory Imbedding,” Moon says. “It’s just a tidbit, but it’s a supervised data visualization technique that allows you to explore relationships between related data in multidimensional datasets.”
To understand this, he says it’s helpful to examine the capabilities of previously developed unsupervised and supervised data visualization techniques.
“Commonly used unsupervised methods such as PHATE, t-SNE, and UMAP, as well as existing supervised methods, can help visualize the structure of big datasets,” Moon says. “However, each has some weaknesses. Some tend to overemphasize the differences between groups of data, and some do not take into account how those groups relate to each other. RF-PHATE is much better at preserving the structure of how groups of data relate to each other.”
In their paper, the research team demonstrates the functionality of the model and documents how RF-PHATE provides evidence for previously suspected subtypes of multiple sclerosis.
“Identifying subtypes is important because MS affects each patient differently and knowing the specific type can inform treatment decisions,” Moon says.
Additional datasets used to investigate the RF-PHATE model included plasma data from COVID-19 patients and data from antioxidant-treated lung cancer cells, but Moon notes that the model is not limited to biological data.
“RF-PHATE can be applied to many other fields and can also be used to develop more interpretable AI models or analyze the models themselves,” he says. “This remains a very active area of research for our group.”
Moon says his group supports AI for Science. It is an international movement that encourages the use of artificial intelligence and machine learning to accelerate research, analyze large datasets, and simulate complex systems.
“Through interdisciplinary collaboration, we can develop and use AI tools to more effectively analyze scientific data and accelerate discovery,” he says.
*In addition to Utah State University, collaborating institutions on the June 30, 2026 Nature Computational Science paper include Brigham Young University, Université de Montréal, Mira-Québec AI Institute, University of Montreal Hospital Center, and University of California, San Francisco. University of Utah, Charles LeMoyne Hospital, University of Lausanne, McGill University.
/Open to the public. This material from the original organization/author may be of a contemporary nature and has been edited for clarity, style, and length. Mirage.News does not take any institutional position or position, and all views, positions, and conclusions expressed herein are solely those of the authors. Read the full text here.
