Ecological Genomic Indicators for Urban Water Quality

This study addressed the questions regarding microbial communities in the Kranji Reservoir Catchment (1) Does the observed bacterial community composition in samples vary as a function of site, land-use category, and measured levels of E. coli and HF183 marker? And (2) To what extent do sequences previously associated with sewage samples (sewage-associated taxa; SAT) correlate to the abundance of observed FIB? (3) Finally we sought to determine the relationship betweenFIB and potential bacterial risk agents by examining whether sequences with genus-or species-level similarity to human pathogens (pathogen-like sequences genus/species; PLSg or PLSs) varied as a function of land use and/or FIB.

(1) Bacterial community composition (BCC) as sequenced by next-generation IlluminaMiSeq 16S rRNA gene varied as a function of both land-use and site in Kranji Reservoir and catchment.

A. Land-use: Classification of sequences at taxonomic levels revealed clustering of two major sample groups; those from Kranji Reservoir versus from other land use categories as distinguished by the first & second PCO axes. PERMANOVA analysis confirmed the variation of BCC with land use and not months or a combination of both. Genus group responsible for differences in BCC between land use categories was arcobacterand acinetobacterwith the exception of Kranji Reservoir bacterial groups that also included Flexibacterand Microcystis.

B. Sites: Sites sampled on different dates tended to harbor bacterial communities that were more similar to each other than to other samples based on Bray Curtis similarity indexes (0-100) of the OTU distribution.

C. FIB: E. coli abundance explained more variation in BCC than the human specific marker HF183. E. coli also explained more significant varianceamong bacterial community structure of SAT, PLSg and PLSs.

(2) Sewage-associated taxa (SAT) were recovered at high proportion in the horticultural area and were moderately correlated with measured FIB supporting the use of FIB as potential proxy for exposure to sewage in Kranji catchment.

(3) The proportion of pathogen-like sequences (PLSg and PLSs) recovered were moderately correlated to FIB and similarly, varied as a function of land use with a greater proportion in the horticultural area. Profiling BCC using next-generation sequencing provides insights into the proportion and occurrence of SAT and bacterial PLS in surface waters and a better understanding of their relationship to measured levels of FIB in tropical environments.