By Elisa Bryan
Dr. Marilyn Brandt, Research Professor at the University of the Virgin Islands (UVI), and her collaborators have achieved a milestone in coral reef science: the creation of the first machine learning model that can accurately distinguish between two of the most destructive coral diseases in the Caribbean—stony coral tissue loss disease (SCTLD) and white plague (WP).
At first glance, these two diseases appear almost identical, showing white lesions on coral tissue. But while their symptoms look similar, they differ greatly in how quickly they spread and how deadly they are to coral colonies. Until now, this similarity made it extremely difficult for scientists and reef managers to identify which disease was present on a reef, slowing down response efforts.
To solve this problem, samples from two disease-exposure experiments conducted by Dr. Brandt’s laboratory at UVI were processed by her colleagues at the University of Texas Arlington. They analyzed coral gene expression profiles from these samples and identified 463 unique biomarkers that act like molecular fingerprints of each disease. By training a machine learning model on these biomarkers, they built a diagnostic tool with >90% accuracy in distinguishing SCTLD from WP. The model was validated using coral samples collected directly from the field—including reefs in the U.S. Virgin Islands—demonstrating its real-world effectiveness.
This breakthrough goes far beyond academic research. For the U.S. Virgin Islands, it represents a possible new level of protection for our reefs. Coral reefs are the backbone of our marine ecosystems, supporting fisheries, fueling tourism, and protecting shorelines from storms. With these biomarkers a diagnostic tool can be created, and managers will be able to detect outbreaks faster, respond more effectively, and make better-informed decisions to safeguard these ecosystems.
The full publication may be found here: https://link.springer.com/article/10.1007/s42452-025-07382-7
