Fig. 2: Single-polyp statistical analyses.
From: Single-polyp metabolomics reveals biochemical structuring of the coral holobiont at multiple scales
a Principal component plot and (b) canonical plot from discriminant analysis supervised by colony using all metabolites. Color legend for genotypes is the same for (a, b). c Box plots showing the mean standard deviation of the abundance of every compound in all samples (“across colonies”), between branches within a colony (“between branches”), and between polyps within branches (“within branches”); ***p < 0.05. Boxplots are median with quartiles and whiskers extending 1.5 IQR beyond quartiles. d Canonical plot from discriminant analysis supervised by polyp number (i.e., position on the branch with lower numbers being closer to the base of the branch). e Pairwise Bray–Curtis dissimilarity matrix for polyps on the same branch. Shared letters represent no statistically significant differences between groups (pairwise Kruskal–Wallace; α = 0.95). f Heat map of ClassyFire SubClass chemical groups that were significantly correlated with the distance from the base of the branch with Pearson’s r values shown on the right of the map.
