Towards predicting coral fate with a molecular biotechnology+machine-learning approach (NCEI Accession 0254274)

Given the widespread decline of coral reefs across the globe on account of climate change-induced rises in seawater temperature, a series of temperature-focused models have been generated to predict when and where bleaching events may occur (e.g., NOAA’s Coral Reef Watch). Although such algorithms are adept at forecasting the onset of periods of severe bleaching in many parts of the world, they suffer from poor predictive capacity in areas featuring high numbers of corals that have either adapted or acclimatized to life in marginalized environments, such as stress-hardened corals of the Florida Keys. In these areas, it may instead be superior to use physiological data from the corals themselves to make predictions about coral bleaching susceptibility. To that end, both field and laboratory analyses were undertaken with the massive Caribbean reef-builder Orbicella faveolata whereby, after elucidating the cellular pathways underlying both bleaching and high-temperature tolerance in diverse genotypes ex situ, the protein profiles of tagged field colonies were tracked across seasons. Neural networks trained with proteomic data from the laboratory specimens were then tested using proteomic data from bleaching-susceptible and bleaching-resistant field colonies, and the resulting artificial intelligence (AI) was capable of predicting with a high degree of confidence whether a coral colony would bleach. This ‘Omics+AI approach could be of potential use in delineating O. faveolata climate resilience elsewhere in the Florida Keys, and perhaps beyond. This dataset includes raw files from the mass spectrometer, as well as "distilled" mass spectrometry data that can be analyzed and interpreted by those with access to a personal computer and the Microsoft Office suite.