Fig. 2

Differential occurrence of disinfectants and antibiotic resistance determinants. Reads within samples were classified using the CARD RGI tool and grouped to the drug class that they gave resistance to. A Principal Component Analysis (PCoA) showing samples grouped by Robust Aitchison’s distance (i.e. Euclidean distance of robust Centered-Log Ratio transformed counts). The variation accounted for by each principal component is indicated in parentheses on the axes. Ellipses show the mean plus three standard deviations for each treatment, and the box shows the results of ANOSIM and PERMANOVA tests between the treatments. B Chao1 richness or Simpson’s diversity index in each of the three replicates for each treatment (top). The results of ANOVA tests for differences between treatments are shown in boxes within the axes, while p-values for post-hoc Tukey’s honestly significant difference (HSD) between treatments are shown (highlighted in red are significant values; p ≤ 0.05; bottom). Both (A) and (B) show results for reads classified as ARGs. C The number of reads classified (in reads per kilobase per million, RPKM; blue colour scale) and the number of ARGs identified within each sample (red colour scale). The main heatmap (blue to yellow colour scale) shows the abundance of ARGs giving resistance to different drug classes with the number of genes detected within each drug class shown on the right. Numbers within cells indicate RPKM, while the colour shows the proportion of the maximum for that drug class