Statistical Analysis
Statistical analyses were conducted using PRIMER v7 plus PERMANOVA, Statistical Analyses of Metagenomic profiles (STAMP) software (Parks et al. 2014), and R (R Project for Statistical Computing). Significant differences in the relative abundances of microbial genera and functions in the coral microbial communities sampled from inner and outer reefs were identified by permutational multivariate analysis of variance (PERMANOVA) using Bray-Curtis distances of normalized relative abundance obtained using a square-root transformation. A cluster analysis was created to visualize the similarity of the coral microbiome between temperature treatments. We also used PRIMER to calculate Pielou’s evenness index (J’) and Shannon’s diversity index (H’) of microbial genera. Principal Component Analyses (PCA) and the multiple comparisons of either taxa or functions across the three groups of metagenomes (i.e., pre-treatment, ambient, and heat) were conducted in STAMP using ANOVA/Tukey-Kramer, Welch’s pairwise comparisons, and Benjamini-Hochberg FDR corrections. We used R to test parametric assumptions of normality (Shapiro-Wilk’s test) and homoscedasticity (Bartlett’s test), and differences in the percent changes of maximum photochemical efficiency of photosystem II (Fv/Fm) and productivity to respiration ratios (GP: R) between temperature treatments (Student’s T-Test).