Award: OCE-1712240

The supported study, Immunity to Community: Can Quantifying Immune Traits Inform Reef Community Structure, had three main goals: 1) determine variability of coral immune traits in several common coral species found on Caribbean reefs, 2) determine the variability in resistance to white plague disease transmission in coral, and 3) develop a predictive model of coral community assemblage that incorporates immune traits. Collaborative PIs on this project, Dr. Laura Mydlarz (University of Texas Arlington) and Dr. Marilyn Brandt (University of the Virgin Islands) focused on goals 1 and 2 by conducting an immune stimulation experiment and a white plague transmission study on seven coral species from reefs in the US Virgin Islands to quantify immune traits. The award allocated to PI Erinn Muller focused on the third objective, incorporating the immune traits from the experiments into a predictive model to assess changes in coral community assemblages through time. Sara Williams, a Northeastern University Ph.D. candidate, was incorporated into the project as a modeler within PI Muller's Coral Health and Disease Program at Mote Marine Laboratory. Her efforts focused on conducting the hierarchical clustering, random forest analyses, and developing the trait-space model. All coding was completed in the open source program R and is freely available via GitHub. We used random forest analyses to determine which immunological traits most influenced the clustering of coral colonies based on susceptibility to white plague. The random forest analyses indicated that the gene for the protein Proteasome subunit alpha type 4 was the most important trait for determining the clusters based on white plague disease susceptibility. Additionally, the abundances of several bacteria within the coral microbiome, Endozoicomonas sp., Pseudomonas veronii, Arthobacter ramosus, Nautella italica, and Thalassobius mediterraneus, were important traits for cluster determination, as well as the concentration of immune-related protein Prophenyloxidase. To predict coral community assemblages through time, we applied a trait-space model with an environmental filter (i.e. a white plague outbreak) to a reef community and then used a hierarchical Bayesian model that incorporated intraspecific trait variation. First, an ideal trait distribution for disease resistance was created by taking the trait averages of the most disease-resistant species. This ideal target trait distribution acted as the environmental filter. The model produced post-outbreak relative abundances of the coral species based on the resistant trait targets. Initial results of the trait-space model suggest that the resistant species, Montastraea cavernosa, will be the most abundant species, followed by the tolerant species, Porites asteroides, P. porites, and Siderastraea siderea. Our model was validated by comparing the trait-space model predictions to a known coral reef assemblage post-disease outbreak in the USVI. The funded project provided educational and experiential training for a graduate student at Northeastern University as well as a technician within PI Muller?s Program. Project highlights have been communicated to the public through various activities including a Science Cafe, a Garden Club of America Annual meeting, as well as up to 100 other public engagement activities that the Mote Coral Health and Disease Program participated in over the last two years, including two Congressional Briefings at the US Capitol. Last Modified: 06/28/2020 Submitted by: Erinn M Muller