Fig. 4: Within-group and between-group community dissimilarity.

Bacterial community dissimilarity in the micropollutant treatments vs. the controls (i.e., “Con.”) over time (A) and the dynamic regularity within treatments (B). Significant differences in the Bray-Curtis dissimilarity for each pairwise comparison of the treatment and control (B1 microcosms vs. later microcosms in A were determined in a Wilcoxon test: ***, P < 0.001; **, 0.001 < P < 0.01; *, 0.01 < P < 0.05. The dissimilarity-overlap analysis within each treatment is shown in B; out of total 210 sample pairs, only the sample pairs that had median overlap above cutoff = 0.913 were shown in the plots. The gray lines indicate the linear regression with the slope (m), and the P value (the fraction of bootstrap realizations in which the slope was negative). Gray shaded areas indicate the 95% confidence intervals of the linear regression. Significance level was determined at P < 0.05. “B1–B7” represent the number of microcosm batch that corresponds to every 5-day inoculation.