The research, led by Mustafa Karatas and Jelle Matthijnssens of the Laboratory of Clinical and Epidemiological Virology at KU Leuven, applied hybrid-capture viral metagenomics to 127 weekly 24-hour composite influent samples collected from a single sentinel wastewater treatment plant in Leuven, Belgium, over a span of 31 months. Rather than amplifying a predefined genetic target, the team used panels of synthetic probes designed to bind viral genetic material across a broad range of virus families. These probes capture fragments from the complex wastewater mixture, enriching them relative to the overwhelming background of bacterial, fungal, plant and human DNA and RNA, and thereby allowing deep sequencing to recover near-complete viral genomes from the sample.

The technical payoff was substantial. Across the 31-month monitoring window, the approach recovered a genome-resolved, target-enriched virome spanning 339 viral species and 715 distinct strains. This depth of resolution matters because public health decisions increasingly depend on knowing not just whether a virus is present, but which variant or genotype is circulating. For RNA viruses with high mutation rates, such as SARS-CoV-2 or norovirus, strain-level identification can signal the arrival of a new variant with altered transmissibility or immune escape properties long before clinical laboratories compile comparable data.

To validate the biological meaning of the wastewater signals, the researchers linked their time series to national clinical surveillance indicators for multiple pathogens. The concordance analysis showed that wastewater and clinical data tracked each other closely for several medically important viruses, including rotavirus A, SARS-CoV-2 and human mastadenovirus. For rotavirus A, a leading cause of severe childhood gastroenteritis, viral genetic material in sewage typically preceded clinical surveillance signals by approximately one week. That lead time, consistent with theoretical models of wastewater-based epidemiology, reflects the interval between infection, shedding of virus into sewage, and the delay before symptomatic patients seek care and are formally tested and reported.

Beyond confirming known pathogens, the longitudinal design revealed signals that routine surveillance programmes missed entirely. The most striking example was a surge of parvovirus B19 beginning in April 2023, roughly one year before the national public health alert was issued in Belgium. Parvovirus B19, which causes fifth disease in children and can produce severe complications in pregnant people and immunocompromised patients, is not part of standard wastewater monitoring panels. Its detection in the Leuven sewage time series illustrates the core advantage of a broad hybrid-capture approach: because the probe panel targets many virus families at once, the method can flag unusual increases in pathogens that no one thought to monitor, without requiring a new assay to be designed and deployed.

The strain-resolved dimension of the study also extended to rotavirus genotyping. Rotavirus A carries a segmented double-stranded RNA genome, and its genotype constellation, defined by the combination of G-type and P-type segments, can shift through reassortment and interspecies transmission. The wastewater data captured the temporal dynamics of rotavirus genotypes and identified unusual genotype constellations circulating in the community, including equine-like G3P[8] strains whose post-pandemic dominance in Belgium had previously been documented through clinical sampling. Recovering this information from sewage, without culturing or clinical specimen collection, shows that hybrid-capture metagenomics can serve as a genuine molecular epidemiology tool rather than a mere presence-absence detector.

Seasonal community turnover was another feature the multi-year dataset made visible. The viral composition of wastewater shifted predictably with the seasons, with enteric viruses peaking in winter months and respiratory viruses following their own epidemic rhythms. Because the samples were collected weekly from the same treatment plant serving a stable population, the researchers could distinguish genuine epidemiological trends from sampling noise, and the principal-coordinate analysis of viral community composition revealed coherent clustering by season and epidemic phase. This longitudinal perspective is precisely what episodic studies, which sample during outbreaks and stop afterwards, cannot provide.

The study also addresses a practical question facing surveillance programmes worldwide: what happens to wastewater monitoring after the COVID-19 emergency? Many programmes built during the pandemic have struggled to justify continued funding for single-pathogen assays. The Leuven results suggest an answer in the form of economies of scope. A single hybrid-capture workflow, applied consistently, generates data on hundreds of viruses at once, including SARS-CoV-2 variants, noroviruses, enteroviruses, adenoviruses, astroviruses and rotaviruses, along with less expected agents such as parvovirus B19. The marginal cost of watching an additional pathogen within the same pipeline is far lower than launching a separate targeted programme for each new threat, a consideration that becomes critical when novel viruses such as avian influenza H5N1 raise concerns about zoonotic spillover.

The researchers have made their analytical resources available to the community, including curated rotavirus A genotyping databases and scripts to reproduce the figures, hosted on GitHub, and the raw sequencing data, with human-associated reads removed for privacy, deposited in the Sequence Read Archive under accession PRJNA1391037. This openness lowers the barrier for other laboratories to adopt the approach and benchmark their own sentinel sites against the Leuven time series.

Limitations remain, as the authors and commentators in the field acknowledge. Hybrid-capture enrichment depends on probe design, so truly novel viruses divergent from panel targets may evade capture, and quantitative interpretation of wastewater read counts requires careful normalization against population size, shedding rates and sewer hydrology. Nonetheless, the study establishes a benchmark for what longitudinal, multi-pathogen, strain-resolved wastewater surveillance can achieve. By demonstrating a full year of early warning for parvovirus B19, a consistent one-week lead over clinical reporting for rotavirus A, and genome-resolved tracking of 715 viral strains from a single treatment plant, the work makes a compelling case that the future of wastewater-based epidemiology lies not in narrower tests, but in broader, deeper and longer views of the viruses circulating silently through urban populations.

Subject of Research: Hybrid-capture metagenomic sequencing of wastewater for longitudinal, strain-resolved surveillance of human-associated viruses

Article Title: Hybrid-capture-enabled longitudinal metagenomics allows strain-resolved human-associated virus surveillance in wastewater

Article References: Karatas, M., Bloemen, M., Swinnen, J., Close, L., Van Ranst, M., Wollants, E., & Matthijnssens, J. (2026). Hybrid-capture-enabled longitudinal metagenomics allows strain-resolved human-associated virus surveillance in wastewater. Nature Water, 4(9), 1128-1138. https://doi.org/10.1038/s44221-026-00693-y

Image Credits: AI Generated

DOI: 10.1038/s44221-026-00693-y

Keywords: wastewater surveillance, hybrid capture, viral metagenomics, strain-resolved, rotavirus A, SARS-CoV-2, parvovirus B19, adenovirus, public health, epidemiology, KU Leuven, Nature Water