Livestock aggregated samples for monitoring viruses infecting animals and potentially zoonotic viral pathogens

Abstract

Active surveillance of livestock pathogens is essential to prevent animal health losses and zoonotic spillover. This study evaluted aggregated environmental sampling as a non-invasive approach for monitoring swine- and cattle-associated viruses across farms and slaughterhouses, bridging the gap between agricultural biosecurity and public health. Over eleven months, 105 samples, including swine slurry, cattle manure, farm air, and slaughterhouse wastewater, were collected and analyzed using pathogen-specific (RT) qPCR and targeted viral metagenomics.

Seasonal and sample patterns were evident, with higher detection of rotavirus A (RoV-A) and bovine coronavirus (BCoV) in slurry and wastewater during winter, and porcine reproductive and respiratory syndrome virus (PRRSV), porcine epidemic diarrhea virus (PEDV) and transmissible gastroenteritis virus (TGEV) sporadicly in slaughterhouse wastewater. Farm slurry or manure were optimal for enteric viruses such as RoV-A or hepatitis E virus (HEV), and farm air proved valuable for respiratory viruses like BCoV.

Targeted sequencing identified a broader viral community, revealing up to 80% of total detected viral species in slaughterhouse wastewater alone. Frequent detection of porcine bocavirus, circoviruses and astrovirus, alongside zoonotic viruses such as HEV and porcine bufavirus (PBuV), underscored the environmental transmission risk at the human-animal interface. Sequencing also uncovered viruses of unclear pathogenicity, including kobuvirus and copiparvovirus, underscoring the complexity of the livestock virome and the potential for emerging viral threats. Slaughterhouse wastewater consistently captured the highest viral richness, integrating inputs from multiple farms and regions, while farm air samples yielded lower diversity but detected respiratory (astrovirus, caliciviruses) and persistent viruses (papillomaviruses, polyomaviruses). Aggregated sampling proved particularly efficient in swine systems, while cattle surveillance may require adapted strategies due to lower stocking densities and greater ventilation.

This work demonstrates the novelty and value of aggregated environmental samples, collected at different points in the production chain, as strategic One Health sentinels. This scalable, practical approach supports early warning and control of animal and zoonotic diseases, directly contributing to One Health surveillance.

Graphical abstract

Highlights

• •Aggregated samples enable large-scale livestock virus surveillance.
• •Slaughterhouse wastewater captures up to 80% of total viral diversity detected.
• •Matrix–virus matching improves detection efficiency in One Health monitoring.
• •Farm slurry/manure optimal for enteric viruses, air for respiratory pathogens.
• •Approach supports early warning and targeted zoonotic disease control.

1. Introduction

Global meat production has been steadily increasing to meet the growing demand for animal protein [1], leading to closer interactions between humans, animals, and shared environments. Whether on farms or at rendering facilities like slaughterhouses, these environments create ideal conditions for pathogen transmission between animals and between humans and animals. While many zoonotic pathogens, including influenza virus and hepatitis E virus (HEV) have their origins in livestock populations [2], [3], active surveillance of these pathogens at the human-animal interface is often insufficient. Detection efforts are often only initiated after outbreaks cause severe impacts on animal or human health [4].

Monitoring zoonotic and animal-specific pathogens presents challenges, largely due to industry concerns about biosecurity and animal welfare. Traditional surveillance methods, which require direct handling of animals, can introduce risks to both livestock and researchers, discouraging collaboration with industry stakeholders. However, growing evidence suggests that indirect, noninvasive sampling of aggregated materials, such as fecal slurry/manure [5] or air samples from farms [6], [7] and slaughterhouse wastewater [8], [9], [10], can serve as sentinel points for one-health surveillance of specific viruses [11].

Aggregated samples, representing a group of pigs within a population, have been explored as a tool for pathogen monitoring in the swine industry. The use of oral fluids collected from hanging cotton ropes for routine surveillance of swine pathogens, such as porcine reproductive and respiratory syndrome virus (PRRSV), porcine circovirus type 2 (PCV2) or influenza A virus (IAV), has proven to be more efficient and practical alternative to individual pig sampling [12], [13]. In contrast, while cattle production is associated with a range of infectious agents that affect the digestive system and can be especially harmful to calves [14], similar aggregated sampling approaches have yet to be implemented in the cattle industry. Existing research on pathogen monitoring in cattle has largely focused on detecting specific pathogens, such as bovine coronavirus (BCoV), bovine respiratory syncytial virus (BRSV), astrovirus (AstV) or rotavirus (RoV) in isolated cases [15]. This gap in surveillance is particularly concerning as cattle can act as reservoirs for pathogens like foot and mouth disease virus [16] or highly pathogenic avian influenza H5N1 [17], raising the risk of transmission to other animals, wildlife, and humans, and underscoring the need for improved tools to control potential spillover events.

In this study, we aimed to evaluate the use of aggregated sampling as a surveillance tool for monitoring zoonotic and animal pathogens within the cattle and swine industries. By leveraging pooled samples from farms (including pig slurry, cattle manure and farm air) and rendering facilities (such as slaughterhouse wastewater), we sought to assess the excretion levels and the prevalence of key pathogens, as well as the viral indicators specific to swine and bovine populations (porcine adenovirus (PAdV) and bovine poliomavirus (BPyV)). Additionally, we aimed to investigate the composition of the viral communities present in these aggregated samples.

2. Materials and methods

2.1. Sampling sites and sample collection

A total of 105 samples were collected for the study. The swine farms (F1 and F2) and the first cattle farm (F3) were located 50 km northeast of Barcelona and the second dairy cattle farm (F4) was located 40 km west of Barcelona city in Catalonia, Spain. On a month basis, from September 2023 to July 2024, and using gloves and 100 ml sterile containers, slurry and manure lixiviate samples were collected from 2 swine and 2 dairy cattle farms, with the explicit consent of their owners to participate in the study. Slurry samples from swine barns containing finishing pigs were collected approximately 5–10 cm below the surface of pits. On each sampling event, paired air samples (n: 34) were collected from stables with adult animals using a Coriolis μ air sampler (Bertin Technologies, France). This sampler, designed to capture viruses and particles ranging from 0.5 to 20 μm, was operated at a flow rate of 300 l/min for 60 min. To maintain a constant liquid volume and compensate for evaporation, the long-monitoring accessory provided by the manufacturer was used. Air samples were collected in sterile collection cones, pre-filled with 15 ml of a saline phosphate buffer solution (PBS). Table 1 provides a summary of the number of animals managed in each sampling site and the number of samples collected.

Table 1.

Number of aggregate samples from each sampling site (F: farms, E: slaughterhouses) and the approximate number of animals managed/handled.

		Number of animals	Sample type	Number of samples
Swine	F1	800	Slurry	8
			Air	8
	F2	2000	Slurry	8
			Air	8
	E1	6500	Wastewater	8
	E2	4000	Wastewater	7
Cattle	F3	200	Manure	11
			Air	9
	F4	280	Manure	11
			Air	9
	E3	500	Wastewater	9
	E4	800	Wastewater	9
				105

Four slaughterhouses located in the province of Barcelona, 2 processing swine (E1 and E2) and 2 processing ruminants (E3 and E4), were also selected for the study. A total of 33 wastewater samples were collected from the effluent of the slaughterhouse process (Table 1). The swine slaughterhouses receive pigs from other provinces within Spain as well as from other countries, while E3 and E4 receive cattle and sheep from local farms within the region. The sampling days were chosen according to the schedule of each slaughterhouse, ensuring that on the sampling day, only cows were slaughtered.

All wastewater samples were collected using 100 ml sterile containers, transported at 4 °C to the laboratory, and processed on the same day of collection. All activities were carried out in the laboratories of the University of Barcelona adhering to established biosafety protocols.

This study did not require formal institutional animal ethics approval, as no live animals were handled or subjected to experimental procedures. All samples consisted of environmental waste (slurry, manure, and wastewater) or air, collected using non-invasive methods. Access to the facilities was granted by the respective owners, and all data have been anonymized to protect the identity of the participating production sites.

2.2. Virus concentration, nucleic acid extraction and (RT)qPCR detection

To ensure maximum viral recovery, viral particles were eluted from the solid and organic fractions of the samples. This elution step, based on chemical desorption, is essential to release viruses that are strongly adsorbed to particles due to their electronegative charge, transferring them into the liquid phase for subsequent concentration. Specifically, 5 ml slurry or manure lixiviate and 80 ml of slaughterhouse wastewater samples were mixed with glycine buffer (pH 9.5, 0.25 N) at 1:1 ratio, shaken on ice for 30 min and then centrifuged (8000 xg) for 15 min at 4 °C to remove the remaining organic matter and bacteria. The supernatants were concentrated by ultrafiltration using the automatic Concentration Pipette (CP-Select™) with 150 KDa tips (InnovaPrep) into a final volume of 300 μl. The concentration of viral particles from aerosol samples was performed by ultrafiltering the volume from the collection cones using the automatic Concentration Pipette (CP-Select™) as described above. The viral nucleic acids (NA) were then extracted from 280 μl using the QIAamp® Viral RNA Mini Kit (Qiagen). The extracted NA were eluted in 80 μl, diluted 1:2 with elution buffer to have enough volume for all the analysis. Eluates were stored at −80 °C for further analysis by specific (RT)qPCR or targeted massive sequencing analysis.

The presence of 6 swine (HEV, rotavirus (RoV), transmissible gastroenteritis virus (TGEV), porcine epidemic diarrhea virus (PEDV), porcine reproductive and respiratory syndrome virus (PRRSV) and influenza A virus (IAV)) and 5 cattle viral pathogens (RoV, bovine coronavirus (BCoV), bovine respiratory syncytial virus (BRSV), HEV and IAV) were tested and quantified using specific (RT)qPCR assays commercially available (BCoV and BRSV from Eurovet Veterinaria; IAV from Exopol; and TGEV, PEDV and PRRSV from Life Technologies) or previously described in literature [18], [19]. Additionally, the presence of 2 viral indicators (porcine adenovirus (PAdV) and bovine polyomavirus (BPyV) were tested using previously described protocols [20], [21]. Except for the commercial kits, the (RT)qPCR for DNA and RNA viruses were performed as previously described, using TaqMan® Environmental Master Mix for DNA viruses and the RNA UltraSense™ One-Step (RT)q-PCR System for RNA viruses (Invitrogen). All (RT)qPCR standards were prepared using 3 synthetic gBlocks® Gene fragments (IDT), which were quantified with a Qubit 3.0 dsDNA HS Assay Kit (Invitrogen) and serially diluted from 10⁰ to 10⁷ copies per reaction. All (RT)qPCR assays were performed using the QuantStudio™ Real-Time PCR System from ThermoFisher Scientific. Both undiluted and 10-fold diluted NA extractions were analyzed, and non-template controls were included in each of the assays.

2.3. Viral enrichment, library preparation, capture, sequencing and bioinformatic processing

Nucleic acids (NAs) were pooled by season and included four pooled slurry and swine air samples, four manure lixiviates, four pooled cattle air samples, and four pooled wastewater effluents from swine and cattle slaughterhouses. To ensure the absence of contamination throughout the process, a negative control consisting of molecular-grade water was included from the start of the SISPA protocol and used as the starting material for library preparation as described in Fernandez-Cassi et al., (2018) [22]. Reverse transcription and tagging of NAs were carried out using SuperScript IV (Invitrogen) primer A random nonamer (5’ GTTTCCCAGTCACGATANNNNNNNNN’-3), primer B (5’ GTTTCCCAGTCACGATA’-3) and Sequenase 2.0 (Applied Biosystems). To generate sufficient cDNA for subsequent steps, amplification was performed through 25 PCR cycles using AmpliTaq Gold DNA polymerase (Applied Biosystems). The amplified cDNA was then purified with the Zymo DNA Clean & Concentration Kit (Zymo Research) and quantified using the Qubit dsDNA HS Assay Kit (Invitrogen).

Amplified SISPA products were normalized individually at 100 ng. Enzymatic fragmentation was performed at 37 °C for 40 min, followed by end repair and dA-tailing using the Twist EF Library Prep 2.0 kit, according to the manufacturer's instructions. Adapters with unique dual barcodes (Twist (HT) Universal Adapter System) were ligated using the Twist Library Preparation EF Kit 2.0 (Twist Bioscience). The libraries were subsequently purified using DNA beads from the Twist Library Preparation Kit 2, and amplification was carried out with 7 PCR cycles to increase sequencing yield. After amplification, the libraries were purified again and pooled into sets of 8 samples with a final DNA amount of 1.5 μg for hybridization with the Twist Comprehensive Virus Probe Panel (Twist Bioscience). The hybridization process involved a 16-h incubation at 70 °C, followed by multiple washing steps using Twist Dry Down beads and Twist Wash buffers. The captured fragments underwent an additional 14 cycles of amplification using the Equinox Library Amp Mix (2×). The final libraries were purified using AmpureXP Beads, then quantified with the Qubit (Thermo Fisher) and analyzed for fragment size using the Fragment Analyzer (Agilent). Sequencing was performed on an Illumina NextSeq500, generating approximately 1 million 150 bp paired end reads per sample, following the manufacturer's protocols (Illumina Inc.).

The generated paired-end FASTQ files were analyzed using the open-source ID-seq tool (IDseq Portal). Illumina adapters, duplicates, low-quality reads, and low-complexity reads were removed using Trimmomatic [23] and the CD-HITDUP tool v4.6.8 (CD-HIT, PRID:SCR 007105) [24]. Paired reads were processed using the Paired-Reads Interactive Contig Extension (PRICE) computational package (PRICE, RRID:SCR 013063) [25] and the Lempel–Ziv–Welch (LZW) compression score. Assembly-based alignment to the NCBI nucleotide and non-redundant protein databases (NCBI, 2024) was performed using GSNAPL [26] and RAPsearch2 [27]. A minimum identity of 70% and a read length of >100 nt were required for analysis. For human viruses, species were only considered if the alignments showed >90% identity with the NCBI database.

3. Results & discussion

3.1. Surveillance of swine pathogens and indicators in aggregated samples from farms and slaughterhouse wastewater

3.1.1. Specific detection of swine pathogens and indicators

The percentages of detection and quantifications of the porcine fecal indicator (PAdV) and various pathogens identified in swine (farms F1 and F2 and slaughterhouses E1 and E2) are shown in Fig. 1, Fig. 2, Fig. 3. Porcine indicators and pathogens were detected in almost all aggregated samples collected from pig facilities, with a few exceptions, such as the absence of the porcine fecal indicator, PAdV, in one air sample in winter out of the nine samples analyzed (see Supplementary 1). Of the six pathogenic viruses tested (HEV, RoV-A, TGEV, PEDV, PRRSV and IAV) only RoV-A and HEV were detected by (RT)qPCR. RoV-A was identified more frequently in slurry samples and at higher concentrations than HEV, with quantifications ranging from 1E+03 to 1E+06 genome copies/ml across both farms (Fig. 2). RoV-A primarily affects young piglets, particularly breeding pigs aged 3–4 weeks, causing acute gastroenteritis [28]. However, by the time pigs reach slaughter age, RoV-A is less likely to be present in their gastrointestinal tract, as the infection is typically cleared during growth. This trend was reflected in slaughterhouse samples, which represent nearly four times more animals, showing variable RoV-A detection rates throughout the year (Fig. 2). Despite variability in detection rates for RoV-A, when present, it exhibited the highest viral loads among all pathogens tested in the study (Fig. 3). The cold seasons (from September to March) showed higher positivity rates for RoV-A detection (49% vs. 25%) and were the most prevalent periods for its detection. However, RoV-A concentrations during this time were like those observed in slurry samples for the same period (Fig. 2, Fig. 3).

Fig. 1.

Percentage of detection of the different viral families identified by targetted sequencing (TES) as well as viral indicators and pathogens of interest in swine (F1 and F2) and cattle (F3 and F4) farms and swine (E1 and E2) and cattle (E3 and E4) rendering facilities.

Fig. 2.

Quantification (Genome Copies/ml) of the viral indicators and pathogens of interest in swine and cattle farming facilities (F1, F2, F3 and F4) over different sampling seasons (Wi: winter, Sp: spring, Su: Summer, Au: autumn). Swine slurry and air from swine farms are shown in dark and light pink respetively. Quantifications in dairy manure and air are shown in dark and light blue bars. (*) indicates that this virus was not tested in air samples. No PEDV, TGEV, PRRSV or IAV was detected in manure or air from swine farms (F1 and F2) and no HEV and IAV was detected in manure or air from bovine farms (F3 and F4). (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

Fig. 3.

Percentage of detection (%) over the cold season (Autumn and Winter) and warm season (Spring and Summer) and box plots representing mean values (Genome Copies/ml of sample) of the different viral pathogens and fecal indicators (in bold) detected in swine (pink) and bovine (blue-green) slaughterhouse wastewater. Mean temperatures are indicated in brackets. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

By contrast, HEV infections in pigs usually occur during the rearing phase, typically affecting piglets aged 9–14 months [29]. Although HEV is not considered a porcine pathogen, it is highly relevant to the swine industry due to its zoonotic potential, particularly through the consumption of undercooked pork products. In Spain, HEV is widely distributed among swine farms, with a reported herd prevalence of around 18% [30]. In this study, HEV was detected at varying percentages of detection (Fig. 1) and concentrations (Fig. 2) in the 2 farms. However, the consistent detection of HEV in slaughterhouse wastewater (Fig. 3) suggests that herd prevalence may be underestimated, and that sampling wastewater from slaughterhouses could provide a more reliable surveillance method, as it captures a broader representation of animals from diverse geographical areas.

3.1.2. Multipathogen detection in swine aggregated samples

Fig. 1 displays the relative abundances of the viral families detected in swine farms (F1 and F2) and slaughterhouses (E1 and E2), with seasonal distributions provided in Fig. 4A. The sequencing output by viral species is detailed in Fig. 5A.

Fig. 4.

Stacked bar plots showing the relative abundance of viral families across swine (A) and bovine (B) aggregated samples and sampling seasons (Wi: winter, Sp: spring, Su: Summer, Au: autumn).

Fig. 5.

List of viral species identified by targeted sequencing in swine (A) and cattle (B) aggregated samples. Bars indicate the total number of sequencing reads (Log10) and dots represent the intersections between the different aggregated samples type (slurry or manure, air and slaughterhouse wastewater (WW)). Viral species with (*) have been related with infection in different animal hosts and do not primarily infect swine (A) or cattle (B).

PAdV, RoV-A and HEV detected using qPCR and belonging to the Adenoviridae, Sedoreoviridae and Hepeviridae families, were also identified using targeted massive sequencing (Fig. 1). This multipathogen approach revealed seasonal variability in detection: for instance, RoV (specifically RoV-A, -B and -C, see Fig. 5A) were detected more frequently in autumn, with relative read counts below 15% in slurry and below 1.75% in slaughterhouse wastewater (Fig. 4A). Reads assigned to HEV, classified under the Paslahepevirus balayani species, reached up to four log₁₀ in both slurry and slaughterhouse wastewater samples (Fig. 5A). However, their relative abundance was low (<0.4%) compared to other viral families (Fig. 4A). Slaughterhouse wastewater revealed the presence of additional pathogens previously tested using (RT)qPCR, including TGEV, PEDV, and PRRSV, all showing seasonal variation in prevalence. Notably, PEDV was detected in 100% of samples during autumn and winter, which aligns with the seasonal peak of diarrhea outbreaks caused by these viruses in Spain [31]. The discrepancy between detection rates using (RT)qPCR and targeted metaviromics may be due to the reduced sensitivity of the selected assay, which is multiplexed with other pathogens such as TGEV and PRRSV, or due to decreased assay sensitivity caused by primer dropout resulting from natural mutation of PDEV. Similar sensitivity assays could be faced by the other viruses which were tested in a multiplex (RT)qPCR (i.e. PRRSV and TGEV). PRRSV showed a higher occurrence in summer. This virus poses a significant challenge to pig farms due to its association with increased medication costs, reproductive failures such as abortions, and elevated mortality rates among newly born piglets. While PRRSV survival and seasonality are influenced by environmental conditions, particularly air temperature and humidity [32], and that wave crests consistently occur during the winter months, here we found consistent results in both slaughterhouses included in the study. Interestingly, the two swine farms monitored did not show evidence of PRRSV during the sampling period. Instead of replacing traditional, resource-intensive on-farm surveillance programmes, the analysis of PRRSV in slaughterhouse wastewater should be considered as an additional passive surveillance tool. The detection of PRRSV in slaughterhouse wastewater could help veterinary authorities identify farms where active surveillance should be pursued. Moreover, finding a pathogen like PRRSV in this matrix should not be considered a “false positive” to be ignored, but rather a “sentinel” signal. Because the slaughterhouse integrates biological inputs from multiple farms and regions, a detection there acts as a geographic red flag, directing authorities to where they should prioritize resources for more intensive individual testing and localized interventions.

A total of 56 viral species belonging to 13 families were identified in aggregated swine samples using target sequencing. The most abundant viral families were Parvoviridae, Circoviridae, Astroviridae and Picornaviridae, regardless of sample type or sampling season (Fig. 4A). Twenty-one species out of 56 were detected in all types of aggregated samples (Fig. 5A), 12 in both slurry and slaughterhouse wastewater, 3 in slurry and air and 11 exclusively in slaughterhouse wastewater. Of these 11, four were human-specific (aichivirus A, enterovirus A, human associated cyclovirus, and mamastrovirus 4), one was rat-specific (rat kobuvirus), and 5 were pig-specific or have been directly related to swine pathologies (PEDV, porcine norovirus, porcine parvovirus, tetraparvovirus and tottorivirus sp.). These findings are consistent with previous studies reporting a similar host distribution of enteric viruses in urban environments. For instance, Chauhan and coworkers [33] observed a high prevalence of pig-related viruses, such as PEDV and porcine parvoviruses, in wastewater samples collected near livestock facilities. The presence of human-associated viruses suggests potential fecal contamination, likely from slaughterhouse workers. Still, this type of aggregated sample offers a valuable opportunity to detect specific swine pathogens that are often overlooked due to co-infections in standard veterinary diagnostics.

Within the Parvoviridae family, Porcine bocavirus (PBoV) accounted for most reads (Fig. 5A). Other prevalent species included astrovirus (PAstV), Enterovirus geswini (formerly enterovirus G (EV-G)), and porcine circovirus (PCV). However, the likelihood of detecting them varied depending on the setting (farms or slaughterhouses), likely due to the acute nature of the infections they cause. In literature, PCV has also been found in clinically healthy pigs through metagenomic analyses, suggesting that it may persist subclinically within swine populations. This persistence could contribute to co-infections with other pathogens and potentially influence broader disease dynamics [34], [35], [36]. In contrast, PBoV and PAstV have shown a clearer association with diarrhea outbreaks, particularly in piglets raised on high-density farms. Several studies have reported co-infections of PBoV with key swine pathogens, including RoV-A, PEDV, PRRSV, and even Classical Swine Fever Virus [37], [38], indicating the relevance of parvoviruses in multi-pathogen disease scenarios. Additionally, high percentages of PBuV and porcine parvoviruses reads were identified from air and wastewater samples (Fig. 5A). Bufaviruses have been detected in a wide range of mammals and have been considered a putative human gastroenteritis pathogen [39], [40]. Since bufaviruses are considered potential causes of human gastroenteritis, their presence in environmental samples could indicate a possible route for virus transmission to humans, raising public health concerns.

Porcine astroviruses (PAstV) are also highly prevalent in pig populations with nearly 80% of healthy finisher pigs harboring PAstV in their intestines at slaughter. While human AstV infections are mostly reported in winter, no data currently exists on the seasonal incidence of PAstV infections in pigs. Nevertheless, a recent study on diarrhea outbreaks in Spanish swine farms identified PAstV as frequently present, particularly PAstV2, PAstV4, PAstV5 with higher detection rates in postweaning and fattening pigs [31]. Our findings show widespread prevalence of astroviruses in pig populations over the year, which could have implications for public health and zoonotic transmission (Fig. 4A and 5A). The zoonotic potential of AstVs remains unclear, although porcine-human AstV recombinants have been documented, suggesting the possibility of human-to-pig transmission [41].

Porcine bocavirus was detected in all aggregated swine samples, underscoring its widespread presence across different environmental matrices. This consistent detection suggests it may serve as a reliable indicator of porcine fecal contamination, warranting further investigation into its potential role in environmental surveillance and farm hygiene monitoring.

The analysis of slaughterhouse wastewater alone accounted for the detection of 80% (41/51) of the total viral species identified in all aggregated samples from swine farms and rendering facilities. However, a key pathogen like PRRSV was only detected using (RT)qPCR, as the capture panel used to enrich sequencing in this study (Twist comprehensive panel) did not include specific probes for viruses within the Arteriviridae family. As previously reported [42], the use of more sensitive and targeted tools, such as panels specifically designed for the detection of veterinary pathogens or focused on the detection of viral families with pandemic potential, may increase the representation of the viral families that are less commonly observed.

3.2. Surveillance of cattle pathogens and indicators in aggregated samples from farms and slaughterhouse wastewater

3.2.1. Specific detection of cattle pathogens and indicators

In bovine farms and rendering facilities, the situation was somewhat different. Cattle are typically housed in larger, more ventilated spaces or outdoor areas, with a lower density of animals per farm which might have resulted in lower pathogen detection. Although at lower percentages of detection than swine facilities, both bovine aggregated sample types (manure and wastewater) proved useful for detecting 2 out of the 4 cattle pathogens tested BCoV and RoV-A (Fig. 1, Fig. 2, Fig. 3). Both pathogens were occasionally detected throughout the year in both farms and slaughterhouses, and BCoV was also identified in the air at farm 4 during winter at mean concentrations of 3.71E+01GC/ml (Fig. 1, Fig. 2). Both the farm and the slaughterhouse surveillance sites exhibited clear seasonal patterns of infection, with peaks occurring during the winter months. This is similar to the observed trends for RoV-A and BCoV, which are frequently detected in diarrhoeic calves in Spain, particularly during the winter [43], [44]. It should be noted that dairy cows are typically vaccinated against RoV-A and BCoV before parturition to stimulate the production of specific antibodies in the colostrum. This ensures that the first milk the calves receive is rich in antibodies, providing them with passive immunity during the early stages of life [45]. Interestingly, the bovine urine indicator BPyV, previously reported at high prevalences (90%) in ruminant slaughterhouse wastewater [20], was barely detected in 11 out of 18 (60%) of the slaughterhouse aggregated samples (Fig. 3). Like human polyomaviruses, which infect adults and are excreted via urine, BPyV may not be present in the current slaughterhouse processes, which now primarily involve slaughtering 9 calves (<1 year of age) for every adult cow or bull slaughtered [46]. Additionally, changes in the slaughterhouse procedures at rendering facilities (e.g., 24 h of fasting in holding pens with bedding such as straw or sawdust) may prevent urine from reaching wastewater effluents, which could help explain the observed differences in BPyV detection rates in the current study.

3.2.2. Multipathogen detection in cattle aggregated samples

Bovine aggregated samples revealed a total of 63 viral species belonging to 12 viral families (Figs. 4B and 5B). Eight species were found exclusively in manure, 4 solely in air, and 23 only in wastewater. Additionally, 8 species were detected in both manure and air, 13 in manure and wastewater, and 2 in air and wastewater. Only 8 species were present across all three sample types. In total, wastewater samples contributed to the detection of 44 out of 66 species, highlighting their central role in capturing the viruses excreted by cattle in the environment.

The most prevalent viruses in both manure and slaughterhouse wastewater were members of the Circoviridae family, comprising over 84% of total sequencing reads across all wastewater samples throughout the year, and between 33% and 66% of the reads in manure samples. During the colder seasons, 1637 reads from cattle manure samples were aligned to betacoronavirus-1 (48% coverage of the hemagglutinin esterase gene, 619pb), coinciding with the seasonal infection peak and confirmed by (RT)qPCR analysis (Section 3.2.1.). Similarly, genomic fragments corresponding to RoV-A and -C were detected across all three sample types, in agreement with virus detection by (RT)qPCR. No detection of IAV (Influenza A virus) or BRSV (Bovine respiratory syncytial virus) was observed in any of the samples analyzed.

Kobuviruses have been identified in both diarrheic and asymptomatic calves. Although their exact pathogenic role remains unclear, some studies have reported intestinal lesions associated with them, such as necrotizing enteritis [14]. Similarly, enteroviruses, particularly Enterovirus eibovi and Enterovirus fitauri (formerly enterovirus species E and F), are often present without causing symptoms, but they may contribute to intestinal inflammation or act synergistically with other pathogens to exacerbate disease severity [15]. The detection of these viruses, alongside bovine astrovirus (Fig. 4B), (which has also been associated with enteric and neurological conditions [14]), highlights the complexity of viral enteric disease in cattle. These findings reinforce the growing understanding that diarrhea in livestock is usually caused by viral coinfections influenced by animal age and management conditions, rather than by a single pathogen. Therefore, comprehensive multipathogen approach, such as the one provided by virome surveillance, particularly during high-risk periods such as the colder seasons, is essential to better understand the dynamics of enteric disease and to inform more effective prevention and management strategies in cattle production systems.

Interestingly, other pathogens known to be involved in co-infections, such as copiparvovirus, bocaparvovirus, and norovirus, were detected exclusively in wastewater samples, whereas aichivirus B, kobuvirus, and torovirus were present in both manure from farms and slaughterhouse wastewater (Fig. 5B). Sequencing approaches like the one applied in this study are essential for advancing our understanding of the complex viral landscape in cattle, particularly the role of co-infections, seasonal dynamics, and management-related factors in shaping disease outcomes.

Moreover, the detection of only 22 of the 63 identified viral species in air samples (Fig. 5B) highlights the more limited diversity of the airborne virome, likely due to lower viral load or reduced environmental persistence. Notably, reads corresponding to viruses associated with persistent infections, including bovine papillomaviruses and polyomaviruses, were present in the air, raising important questions about their transmission potential and the role of aerosols in the dissemination of chronic viral infections within cattle facilities.

Together, these findings underscore the importance of a holistic, compartment-based surveillance strategy to monitor viral populations in cattle environments. Integrating data from manure, wastewater, and air provides a more complete picture of viral circulation, enhancing early detection capabilities and informing more targeted and timely interventions for disease prevention and control.

Several viral species typically associated with ovine hosts were detected in cattle-related samples. Ovine papillomavirus was present in both air and manure, while ovine adenovirus appeared in air, manure, and wastewater. Other ovine viruses such as ovine astrovirus, ovine enterovirus, and ovine hokovirus were detected in specific matrices only. Importantly, these detections occurred on days when no sheep slaughtering was scheduled, though low-level residual contamination from previous days remains a possibility. Although most of these viruses exhibit narrow host specificity, it has been reported that bovine viruses like delta-papillomaviruses have been detected in sheep blood when sheep co-grazed with cattle, suggesting possible spill-over or reservoir dynamics [47].

3.3. Metagenomic analysis of aggregated samples as a surveillance tool for zoonotic viruses

A total of 87% (14 out of 16) of swine slurry samples showed evidence of at least one zoonotic virus (HEV or RoV). This finding is in line with previous results from North Carolina, USA, where HEV and RoV-A were detected in 105 swine slurry samples collected from two barns on a single farm) [5]. IAV was not detected in any of the samples tested from either swine or cattle surveillance sites, including farms and slaughterhouses, nor in any of the aggregate sample types (slurry, manure, air or wastewater). And neither HEV and BRSV were present at any of the bovine aggregated matrices.

Despite Spain being among the top five pork-producing countries in Europe, housing approximately 69% of the EU pig herd and contributing significantly to the EU's pork output, no serology-based nor (RT)qPCR-based surveillance for HEV, RoV-A, or IAV is currently in place. The 18,000 sequencing reads classified as HEV in slaughterhouse aggregated samples (n = 5) demonstrate its widespread presence in the swine production environment and support the last ECDC surveillance findings [48], which reported 520 HEV human cases in ten EU/EEA countries in January 2024 alone, with unusual increases in Belgium, Czechia, and Finland, and highlighted the consumption of processed pork products, such as mettwurst and salami, as possible vehicles of transmission. The increase in HEV incidence may be linked to improved awareness of the disease and advancements in diagnostic tools at the clinical level. The lack of systematic national monitoring makes it difficult to assess their true prevalence in pig populations and evaluate the associated risks to animal and public health. As highlighted by Rotolo and colleagues [49], herd-level infectious disease surveillance using aggregate samples (such as bulk tank milk and pen-based oral fluids) has proven effective for monitoring livestock populations. Similarly, our study demonstrates that slaughterhouse wastewater, a readily accessible aggregate matrix, can be leveraged for molecular surveillance to capture the diversity of zoonotic viruses circulating in production systems. Consistent with this, a 2009 study [50] reported that, in addition to wastewater, other aggregated samples such as sludge can reveal a wide diversity of HEV variants; while HEV genotype 3 was detected in individual animal samples, using Sanger sequencing of a 101-nt ORF2 fragment, slaughterhouse sludge exhibited a markedly higher strain diversity than animal-derived samples.

The intensive scale of livestock farming, combined with the frequent movement of animals, personnel and materials between production sites, creates an environment that is highly conducive to the spread of infectious diseases, both within and across farms. In this context, metagenomic surveillance is a powerful approach for tracking viral pathogens, monitoring fecal contamination sources and improving early warning systems, thereby helping to mitigate disease outbreaks and zoonotic spillover.

4. Limitations

This study has several limitations that must be considered when interpreting the results. Firstly, the relatively small sample size and aggregation of samples across different spatial and temporal contexts may have limited the detection of less prevalent or sporadically shed viruses. In addition, the current farming and slaughtering industry is very complex, with animals reared in one region but slaughtered far away from the production sites. This diminishes the potential of tracing the origin of detected pathogens in slaughterhouse wastewater back to the original farm where the disease might be circulating. Second, the reliance on a capture-based sequencing panel biased towards known human pathogens may have limited the detection of highly divergent or uncharacterized animal viruses. For example, although PRRSV was identified through (RT)qPCR, it was not captured by the sequencing workflow, most likely due to the absence of probes targeting the Arteriviridae family. This emphasizes the importance of using targeted tools specifically designed for veterinary pathogens to improve sensitivity and comprehensiveness.

The choice of viral concentration and detection methods probably affected the diversity and abundance of the viruses that were identified. For instance, the use of glycine release buffer has been shown to improve the recovery of enveloped viruses from wastewater samples [51].

Moreover, seasonal fluctuations in viral prevalence and environmental persistence, particularly among enteric and respiratory viruses such as RoV, HEV, PEDV and PRRSV, made comparisons across sites and sample types more difficult. The substantial variation observed between farms and slaughterhouses suggests that environmental, management-related and biological factors have a strong impact on viral circulation patterns and sampling outcomes. Although it is challenging to fully standardize wastewater surveillance protocols due to these contextual differences, it would be beneficial to implement specific indicators, such as porcine adenoviruses or parvoviruses and bovine polyomavirus or circoviruses, to validate each step of the analytical workflow, from viral concentration and nucleic acid extraction to detection and sequencing. Such indicators could improve the reliability of results, particularly when comparing studies or implementing longitudinal surveillance. This is particularly relevant given that comparisons of different experimental designs [52] and target sequencing strategies [53] have demonstrated substantial variability in sequencing efficiency depending on viral characteristics, such as the presence or absence of an envelope, and the type of viral genome (RNA or DNA, single or double strand). While full standardization of wastewater-based surveillance methods may be difficult due to the heterogeneity of sample types, facility practices, and viral targets, the use of robust internal indicators (parvoviruses, circoviruses, polyomaviruses and adenoviruses) can significantly enhance the reliability and comparability of results. These indicators can help validate each step in the workflow, from sample processing to sequencing and data interpretation, thereby mitigating some of the variability introduced by methodological and environmental factors.

5. Conclusions

These aggregated sampling approaches, especially when applied across multiple points in the animal production chain (farms and slaughterhouses) offer a practical and scalable solution for early detection of pathogens. Seasonal differences in virus circulation were clearly observed when using specific pathogen detection ((RT)-qPCR), with gastrointestinal pathogens such as rotavirus A (RoV-A), porcine epidemic diarrhea virus (PEDV), and bovine coronavirus (BCoV) showing higher prevalence during autumn and winter. In contrast, PRRSV was primarily detected in summer in swine slaughterhouse wastewater, despite traditionally peaking in colder months. Respiratory viruses were more often detected in aerosol and wastewater matrices, while enteric viruses predominated in slurry from pigs and manure from cows. When using a multipathogen approach (target sequencing) the highest viral richness and detection rates were consistently observed in both animal slaughterhouse wastewater, which captured up to 80% of the total viral diversity, suggesting its value as a key sentinel point for territorial surveillance. Nevertheless, while swine production allowed for consistent and high-yield viral detection due to higher animal density and closed housing, cattle surveillance required tailored strategies, as lower animal densities and more ventilated settings yielded more diluted viral signals.

Our findings underscore the value of aggregated sampling for monitoring viral pathogens in intensive livestock systems, while minimizing biosecurity risks. Aggregated wastewater from slaughterhouses, which integrates biological inputs from multiple farms, proved particularly effective for tracking virus circulation across regions, offering a strategic tool for early warning and targeted disease control. Multipathogen sequencing revealed a complex virome that included not only established pathogens but also viruses with unclear pathogenic roles, such as kobuvirus, copiparvovirus, and some astrovirus, many of which are commonly involved in co-infections and may influence disease outcomes. The consistent detection of porcine bocavirus and astrovirus across all sample types and seasons supports their use as internal indicators for environmental surveillance. However, further data on their distribution, epidemiology, and host specificity, particularly from diverse geographical regions, would be valuable to assess their applicability as global indicators. Furthermore, the identification of zoonotic viruses such as HEV and PBuV, potential agents of human gastroenteritis, raises concern about possible spillover routes. Their presence in slurry and wastewater samples demonstrates the potential of this approach for monitoring environmental transmission risks.

While single-pathogen (RT)qPCR detection, effectively targeted known swine and cattle viruses, the application of multipathogen sequencing in aggregated samples enabled the detection of viruses with unknown or emerging zoonotic potential. These findings emphasize the importance of integrating wastewater-based epidemiological surveillance across livestock systems to better understand the role of co-infections and environmental persistence in disease dynamics, and to strengthen early warning systems for emerging threats.

This methodology has already been applied beyond the current study, supporting the environmental surveillance of bluetongue virus in Catalonia and contributing to the characterization of HEV subtypes circulating in the region, with manuscripts currently in preparation. These applications further demonstrate the practical value and transferability of aggregated sample-based surveillance approaches for real-time monitoring of animal and zoonotic viruses.

CRediT authorship contribution statement

Marta Rusiñol: Writing – original draft, Visualization, Validation, Supervision, Methodology, Formal analysis, Data curation, Conceptualization. Sandra Martínez-Puchol: Writing – review & editing, Methodology, Formal analysis, Data curation. Diana Ribeiro: Methodology, Formal analysis. Júlia Verdaguer: Methodology, Formal analysis. Ona Torrejón-Llorens: Methodology, Formal analysis. Marta Itarte: Writing – review & editing, Methodology, Formal analysis. Ignasi Estarlich-Landajo: Writing – review & editing, Methodology. Cristina Mejías-Molina: Methodology. Gisela Juliachs-Torroella: Methodology. Rosina Girones: Writing – review & editing, Funding acquisition, Conceptualization. Gustavo A. Ramírez: Conceptualization. Jordi Baliellas: Methodology, Conceptualization. Silvia Bofill-Mas: Writing – review & editing, Validation, Supervision, Resources, Project administration, Methodology, Conceptualization. Xavier Fernández-Cassi: Writing – review & editing, Validation, Supervision, Methodology, Conceptualization.

Declaration of generative AI and AI-assisted technologies in the writing process

During the preparation of this work the authors used ChatGPT to improve language and readability. After using this tool/service, the authors reviewed and edited the content as needed and take full responsibility for the content of the publication.

Declaration of competing interest

The authors declare the following financial interests/personal relationships which may be considered as potential competing interests:

Appendix A. Supplementary data

Supplementary material: Viral quantifications in manure and air from 2 swine and 2 bovine farms and 4 slaughterhouses.

Data availability

The raw sequencing data generated during the current study are available in Zenodo under the DOI number https://doi.org/10.5281/zenodo.16752236