Search

Limiting microbial growth during drinking water distribution is achieved either by maintaining a disinfectant residual or through nutrient limitation without the use of a disinfectant. The impact of these contrasting approaches on the drinking water microbiome is not systematically understood. We utilized genome-resolved metagenomics to compare the structure, metabolic traits, and population genomes of drinking water microbiomes across multiple full-scale drinking water systems utilizing these two-distinct microbial growth control strategies. Microbial communities cluster together at the structural- and functional potential-level based on the presence or absence of a disinfectant residual. Disinfectant residual concentrations alone explained 17 and 6.5% of the variance in structure and functional potential of the drinking water microbiome, respectively, despite including samples from multiple drinking water systems with variable source waters and source water communities, treatment strategies, and chemical compositions. The drinking water microbiome is structurally and functionally less diverse and less variable across disinfected systems as compared to non-disinfected systems. While bacteria were the most abundant domain, archaea and eukaryota were more abundant in non-disinfected and disinfected systems, respectively. Community-level differences in functional potential were driven by enrichment of genes associated with carbon and nitrogen fixation in non-disinfected systems and γ-aminobutyrate metabolism in disinfected systems which may be associated with the recycling of amino acids. Metagenome-assembled genome-level analyses for a subset of phylogenetically related microorganisms suggests that disinfection may select for microorganisms capable of using fatty acids, presumably from microbial decay products, via the glyoxylate cycle. Overall, we find that disinfection exhibits systematic and consistent selective pressures on the drinking water microbiome and may select for microorganisms able to utilize microbial decay products originating from disinfection inactivated microorganisms.

Additional file 1: Table S1. Summary of water quality parameters measured for the samples collected as part of this study. (*NM = not measured). Table S2. BioEnv analyses in vegan package to determine the subset of variables significantly correlated with community similarities. This determines the Spearman’s correlation between Euclidean distances of scaled environmental variables with the Mash distances estimated using metagenomic reads. Table S3. Distance based Redundancy Analysis using Mash distance matrix generated using pairwise Mash distances between samples estimated using metagenomic reads. Table S4. Variance Partition analysis using water chemistry/environmental parameters identified as significant being significantly associated with read-based Mash distances by dbRDA analyses. Table S5. Number of predicted open reading frames (ORFs) for each metagenome co-assembly and number annotated against the KEGG database using Kofamscan. Table S6. BioEnv analyses in vegan package to determine the subset of variables significantly correlated with similarities in functional potential of community estimates using KEGG annotation. This determines the Spearman’s correlation between Euclidean distances of scaled environmental variables with Bray Curtis distance matrix generated from RPKM of KO detected in samples. Table S7. Distance based Redundancy Analysis using pairwise Bray-Curtis distances between samples estimated using from RPKM of KO detected in samples. Table S8. Variance Partition analysis using water chemistry/environmental parameters identified as significant being significantly associated with KO Bray-Curtis distance matrix by dbRDA analyses. Table S9. Summary of modules that were significantly higher abundance in disinfected systems as compared to non-disinfected systems. Table S10. Summary of modules that were significantly higher abundance in non-disinfected systems as compared to disinfected systems.

Additional file 1: Table S1. Summary of water quality parameters measured for the samples collected as part of this study. (*NM = not measured). Table S2. BioEnv analyses in vegan package to determine the subset of variables significantly correlated with community similarities. This determines the Spearman’s correlation between Euclidean distances of scaled environmental variables with the Mash distances estimated using metagenomic reads. Table S3. Distance based Redundancy Analysis using Mash distance matrix generated using pairwise Mash distances between samples estimated using metagenomic reads. Table S4. Variance Partition analysis using water chemistry/environmental parameters identified as significant being significantly associated with read-based Mash distances by dbRDA analyses. Table S5. Number of predicted open reading frames (ORFs) for each metagenome co-assembly and number annotated against the KEGG database using Kofamscan. Table S6. BioEnv analyses in vegan package to determine the subset of variables significantly correlated with similarities in functional potential of community estimates using KEGG annotation. This determines the Spearman’s correlation between Euclidean distances of scaled environmental variables with Bray Curtis distance matrix generated from RPKM of KO detected in samples. Table S7. Distance based Redundancy Analysis using pairwise Bray-Curtis distances between samples estimated using from RPKM of KO detected in samples. Table S8. Variance Partition analysis using water chemistry/environmental parameters identified as significant being significantly associated with KO Bray-Curtis distance matrix by dbRDA analyses. Table S9. Summary of modules that were significantly higher abundance in disinfected systems as compared to non-disinfected systems. Table S10. Summary of modules that were significantly higher abundance in non-disinfected systems as compared to disinfected systems.