The data were imported into R v4.0.5 and converted into a phyloseq object [70 ]. ASVs were removed if they were present less than four times in 20% of the samples. The filtered count table was transformed using centered log-ratio (CLR) with the package microbiome [71 ]. Beta-diversity was analyzed with the package VEGAN 2.5.4 [72 (link)] and the filtered CLR-transformed table. The function vegdist was used to generate dissimilarity indices with a Euclidean distance. To identify significant differences among groups, a Permutational Multivariate Analysis of Variance (PERMANOVA) was used with the function adonis2 with 999 permutations, using a Euclidean distance. The function betadisper was used to calculate group dispersion, which was then tested for significance with the function Permutest.
Differences in beta-diversity for field samples were evaluated in apparently healthy (AH) corals across three zones (vulnerable, endemic, and epidemic). In addition, pairwise group comparison was assessed from betadisper output using the Tukey’s HSD function. The PERMANOVA output was also tested for pairwise comparisons with the function pairwise.adonis and adjusted with a Bonferroni correction [73 ]. Furthermore, all samples (including Acropora spp., sediment, and seawater) were also evaluated for beta-diversity differences in primers, year of collection, biome (field and aquaria), studies, coral species, and sample type (seawater, mucus, tissue slurry, tissue slurry and skeleton, and sediment). These factors were also correlated to principal components (PCs) using the R package PCAtools 2.5.15, and the functions pca and eigencorplot were used to remove the lowest 10% of the variance and to correlate the data and test for significance, respectively.
SCTLD-susceptible coral samples (i.e., without Acropora spp., sediment, and seawater) were also evaluated for beta-diversity. Both biomes (aquaria or field) were examined together and also separately. The matrices were generated with QIIME2-2021.11 with the plugin DEICODE, which runs a robust Aitchison Distance—a method that is not influenced by zeros in the data [74 ]. Pairwise comparisons of dispersion and differences in microbial composition between groups were evaluated using the QIIME2-2021.11 diversity beta-group-significance function using either the permdisp or PERMANOVA method, respectively. DEICODE was also applied to the data without the two most prevalent corals species, Orbicella faveolata (OFAV) and Montastraea cavernosa (MCAV), to see if the same pattern was evident in disease states with and without these coral species.
Differences in beta-diversity for field samples were evaluated in apparently healthy (AH) corals across three zones (vulnerable, endemic, and epidemic). In addition, pairwise group comparison was assessed from betadisper output using the Tukey’s HSD function. The PERMANOVA output was also tested for pairwise comparisons with the function pairwise.adonis and adjusted with a Bonferroni correction [73 ]. Furthermore, all samples (including Acropora spp., sediment, and seawater) were also evaluated for beta-diversity differences in primers, year of collection, biome (field and aquaria), studies, coral species, and sample type (seawater, mucus, tissue slurry, tissue slurry and skeleton, and sediment). These factors were also correlated to principal components (PCs) using the R package PCAtools 2.5.15, and the functions pca and eigencorplot were used to remove the lowest 10% of the variance and to correlate the data and test for significance, respectively.
SCTLD-susceptible coral samples (i.e., without Acropora spp., sediment, and seawater) were also evaluated for beta-diversity. Both biomes (aquaria or field) were examined together and also separately. The matrices were generated with QIIME2-2021.11 with the plugin DEICODE, which runs a robust Aitchison Distance—a method that is not influenced by zeros in the data [74 ]. Pairwise comparisons of dispersion and differences in microbial composition between groups were evaluated using the QIIME2-2021.11 diversity beta-group-significance function using either the permdisp or PERMANOVA method, respectively. DEICODE was also applied to the data without the two most prevalent corals species, Orbicella faveolata (OFAV) and Montastraea cavernosa (MCAV), to see if the same pattern was evident in disease states with and without these coral species.
Free full text: Click here
