Answers are grounded in this paper and cite their sources.
Drinking water treatment strongly reshapes source water bacterial communities, yet what determines the community persisting to the tap remains unresolved. We surveyed four full-scale Korean drinking water treatment plants differing in source water (river or lake) and treatment level (conventional or advanced). Seventy-eight samples spanning raw water (RW) to tap water (TP), collected on three occasions in 2025, were analysed by 16S rRNA metabarcoding together with cellular ATP (cATP). Observed amplicon sequence variants declined from 1914 (RW) to 381 in treated water (TW) and 283 in tap water (TP). Shannon diversity differed significantly among these sampling points (Friedman test, χ2(2) = 18.67, p < 0.001). Community structure also differed among sampling points (PERMANOVA, R2 = 0.175, p = 0.001), and between-plant Bray–Curtis dissimilarity increased from 0.768 (RW) to 0.876 (TW), indicating divergence rather than convergence. Residual chlorine was the only variable with significant marginal effects in the common-point and multiple plant-level dbRDA models. Phreatobacter increased at every plant and round, reaching 34.3% (TP), although relative abundance combined with cATP remained close to RW levels. Legionella and Mycobacterium sequences persisted in TW and TP, yet all 36 samples examined by Legionella culture were negative, showing that gene detection alone cannot establish viability.
No takes yet. Share an insight, caveat, or question.
Jeong et al. (2026) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: