Fecal material in the environment is a primary source of pathogens that cause waterborne diseases and affect over a billion people worldwide. Microbial source tracking (MST) assays based on single genes (e.g., 16S rRNA) do not always provide the resolution needed to attribute fecal contamination sources. In this work, we used dialysis bag mesocosms simulating a freshwater habitat that were spiked separately with cow, pig, or human feces to monitor the decay of host-specific fecal signals over time with metagenomics, traditional qPCR, and culture-based methods. Sequencing of the host fecal communities used as inocula recovered 79 non-redundant metagenome-assembled genomes (MAGs) whose abundance patterns showed that the majority of the fecal community signal was not detectable in the mesocosm metagenomes after four days. Several MAGs showed high host specificity, and thus are promising candidates for biomarkers for their respective host type. Traditional qPCR methods varied in their correlation with MAG decay kinetics. Notably, the human-specific Bacteroides assay, HF183/BFDRev, consistently under-estimated fecal pollution due to not being present in all hosts and/or primer mismatches. This work provides new insights on the persistence and decay kinetics of host-specific gut microbes in the environment and identifies several MAGs as putative biomarkers for improved MST.

Little is known about the genomic diversity of raw municipal wastewater (sewage) microbial communities, including to what extent sewage-specific populations exist and how they can be used to improve source attribution and partitioning in sewage-contaminated waters. Herein, we used the influent of three wastewater treatment plants in Atlanta, Georgia (USA) as inoculum in multiple controlled laboratory mesocosms to simulate sewage contamination events and followed these perturbed freshwater microbial communities with metagenomics over a 7-day observational period. We describe 15 abundant non-redundant bacterial metagenome-assembled genomes (MAGs) ubiquitous within all sewage inoculum yet absent from the unperturbed freshwater control at our analytical limit of detection. Tracking the dynamics of populations represented by these MAGs revealed varied decay kinetics, depending on (inferred) phenotypes, e.g., anaerobes decayed faster under the well-aerated incubation conditions. Notably, a portion of these populations show decay patterns similar to common markers, Enterococcus and HF183. Comparisons against MAGs from different sources such as human and animal feces, revealed low cross-reactivity, indicating how genomic collections could be used to sensitively identify sewage contamination and partition signal among multiple sources. Overall, our results indicate the usefulness of metagenomic approaches for assessing sewage contamination in waterbodies and provides needed methodologies for doing so.