Figure 3
Microbiota analysis of separates samples based upon where they live in the community. Hierarchical average clustering based on the Spearman/Pearson correlation coefficients of the proportion of all taxons, filtered for taxonomic variables with over 50% percent zeroes. Non-informative variables are characterized using the coefficient of variation (mean/standard deviation). Data normalization between samples was conducted by Sum, and those between taxons Pareto Scaling (mean-centered and divided by the square root of standard deviation of each variable). In total, 35 taxonomic features are shown in the heatmap, with statistical significance among four types of samples based on T-test/ANOVA analyses.