First Longitudinal Untargeted Metagenomic Survey Reveals Spatial and Temporal Structuring of the Microbial Community in Selected Tunisian Wastewater Treatment Plants

Wastewater-based metagenomics offers a comprehensive, population-level, and unbiased window into circulating microorganisms, complementing clinical surveillance by capturing signals across an entire connected population rather than only individuals seeking care. In Tunisia, previous wastewater investigations have relied primarily on targeted, single-site approaches, leaving the spatial and seasonal dynamics of wastewater microbiomes largely uncharacterized. This study presents the first longitudinal, multi-site shotgun metagenomic characterization of raw influent wastewater in Tunisia, sampling three treatment plants across the Greater Tunis area, serving populations ranging from approximately 60,000 to 450,000 inhabitants, over eight sampling events spanning winter and summer. Across 24 samples, Bacteria dominated the classified reads, with Proteobacteria as the dominant phylum and Aliarcobacter as the single most abundant genus. Genus-level richness varied significantly among treatment plants, forming a gradient that paralleled estimated population served, while community evenness remained comparatively stable across treatment plants and showed only a modest difference between the two sampling campaigns. Community structure differed significantly by both treatment plant and sampling campaign (winter vs. summer), and an exploratory genus-level differential abundance analysis identified largely distinct sets of taxa associated with each factor, with viral genera modestly overrepresented among campaign-associated taxa. Targeted screening detected clinically relevant bacterial genera and a standard fecal contamination marker in all 24 samples. Among genera previously reported as antimicrobial resistance (AMR) gene carriers, Acidovorax was significantly most abundant at one treatment plant. Dedicated screening of the assembled contigs against the CARD and ResFinder databases identified 129 non-redundant annotation events at 113 independent genomic loci (122 distinct database gene labels), and sample-resolved read mapping indicated that resistome composition differed between treatment plants and between the two sampling campaigns. These findings establish a population-level metagenomic baseline for wastewater microbiomes in Tunisia, demonstrating how longitudinal, multi-site sequencing can resolve spatial and between-campaign temporal structure relevant to environmental surveillance to support data-driven pathogen prioritization and pave the way for compliance with a One Health framework.