Inferring trophic conditions in managed aquifer recharge systems from metagenomic data

Abstract

Humans are increasingly dependent on engineered landscapes to minimize negative health impacts of water consumption. Managed aquifer recharge (MAR) systems, such as river and lake bank filtration, surface spreading or direct injection into the aquifer have been used for decades for water treatment and storage. Microbial and sorptive processes in these systems are effective for the attenuation of many emerging contaminants including trace organic chemicals such as pharmaceuticals and personal care products. Recent studies showed a superior efficiency of trace organic chemical biotransformation by incumbent communities of microorganisms under oxic and carbon-limited (oligotrophic) conditions. This study sought to identify features of bacterial genomes that are predictive of trophic strategy in this water management context. Samples from a pilot scale managed aquifer recharge system with regions of high and low carbon concentration, were used to generate a culture collection from which oligotrophic and copiotrophic bacteria were categorized. Genomic markers linked to either trophic strategy were used to develop a Bayesian network model that can infer prevailing carbon conditions in MAR systems from metagenomic data.

Keywords: Bacteria; Bayesian network; Genomic markers; Managed aquifer recharge; Trace organic chemical; Trophic strategy.