Integrated approaches for pathogen monitoring and shotgun metagenomic analysis in Atlantic salmon farming

Integrated approaches for pathogen monitoring and shotgun metagenomic analysis in Atlantic salmon farming | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Article Integrated approaches for pathogen monitoring and shotgun metagenomic analysis in Atlantic salmon farming Ottavia Benedicenti, David A. Strand, Saima Nasrin Mohammad, Snorre Gulla, and 3 more This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-8373998/v1 This work is licensed under a CC BY 4.0 License Status: Published Journal Publication published 14 Apr, 2026 Read the published version in Scientific Reports → Version 1 posted 10 You are reading this latest preprint version Abstract Specific tools for detecting waterborne pathogens are essential for limiting disease spread in aquaculture. We evaluated a field-deployable workflow combining filtration of eDNA/eRNA with targeted (RT-)qPCR and complementary shotgun metagenomics to monitor pathogens and microbial community dynamics across an Atlantic salmon production cycle, from hatchery to offshore cages. The primary aim was to assess workflow feasibility and performance under real farm conditions, while secondarily examining whether metagenomic profiles could contextualise microbial shifts associated with pathogen presence. ISAV was consistently detected in hatchery water at ~ 4×10³–9×10³ copies per litre, whereas PRV1 was detected only inside sea pens from August onward (~ 4×10²–1.5×10⁴ copies per litre) and increased by more than two orders of magnitude after wellboat delousing. Shotgun metagenomics yielded a median of ~ 1.5×10⁵ reads per sample (mean read length ~ 2.5 kb; N50 > 2 kb), enabling broad taxonomic screening. PRV1-positive seawater samples showed modest decreases in richness and shifts in viral taxa, though patterns were subtle and should be interpreted cautiously given low pathogen loads. The workflow was practical for trained farm personnel, and this integrated approach offers a scalable system for routine pathogen surveillance and supports earlier, evidence-based biosecurity actions, providing broader microbial information than qPCR alone. Biological sciences/Biological techniques Biological sciences/Biotechnology Earth and environmental sciences/Environmental sciences Biological sciences/Microbiology Atlantic salmon pathogens environmental DNA non-invasive sampling shotgun metagenomic sequencing Figures Figure 1 Figure 2 Figure 3 Figure 4 Figure 5 Figure 6 Figure 7 Introduction The production of Atlantic salmon, Salmo salar (Linnaeus), was around 2.02 million tonnes during January–September 2024, a slight 1% decline compared to 2023, and the main producer, Norway, accounted for ca. 1.5 million tonnes in the same period (–1.3% year-on-year) 1 . However, infectious diseases continue to pose major challenges for the salmon farming industry, affecting fish health, welfare, and productivity 2 . Therefore, specific and sensitive tools for pathogen detection are crucial for the development of efficient preventive/ mitigative strategies and to limit the spread of pathogens in fish farms. In recent years, environmental DNA/RNA (eDNA/eRNA) sampling has become an increasingly adopted approach for pathogen surveillance in aquaculture 3 – 9 . Moreover, a good screening method could not only be used to detect pathogens shed from infected fish but also to find pathogens circulating within the farm environment. This could allow for early implementation of preventive measures in response to elevated pathogen levels and in advance of clinical symptoms emerging. For this reason, monitoring fish pathogens in ambient farm water (tanks or cages) offers a less invasive and more scalable alternative to direct fish sampling and may potentially provide earlier insights into pathogen presence during infection. This approach also aligns with a 3R (Replacement, Reduction, Refinement) strategy 10 and supports the One Health framework by promoting integrated biosecurity monitoring that benefits aquatic animal welfare, environmental sustainability, and human health 11 . There are recent studies that have explored metagenomics and metabarcoding in freshwater aquaculture systems, particularly focusing on microbial communities in Recirculating Aquaculture Systems (RAS), as they rely heavily on microbial communities (e.g., biofilters, water, and biofilms) for nutrient cycling, water quality and fish health 12 – 14 . A recent study aimed to compare how different sequencing approaches affect the characterisation of microbial communities in RAS 12 . Here, while 16S rRNA gene sequencing was reliable for identifying community structure and spatiotemporal patterns, long-read sequencing was less effective for quantitative pattern analysis but useful for identifying functional roles and pathogens, and shotgun metagenomics provided broader insights, including detection of fungi, viruses, and bacteriophages, enabling exploration of inter-domain interactions 12 . Oxford Nanopore Technology was also used to analyse microbial communities and detect bacterial pathogens in an Atlantic salmon commercial freshwater RAS performing long-read 16S rRNA gene sequencing 14 . This technology enables real-time, high-resolution monitoring of microbial communities in water systems 14 , 15 . However, the choice of sequencing methods and reference databases might constitute a bottleneck in some studies. For instance, the diversity of RNA viruses in Lake Needwood (USA), explored using metagenomic sequencing, did not match any known viruses, indicating a large pool of novel viral diversity 16 . Another study reviewed and emphasised the potential role of metagenomics, coupled with bioinformatics tools and databases, as a non-invasive, rapid tool for environmental monitoring and disease prevention 13 . Although metagenomics is increasingly used in recirculating aquaculture systems (RAS), its application in tanks and cage-based systems remains limited, despite evidence showing its effectiveness in detecting pathogens and monitoring environmental impacts such as eutrophication 17 . Because each sequencing method has inherent strengths and weaknesses, microbial community studies must clearly document methodological choices (e.g., filtration, extraction protocols, sequencing depth), as these significantly affect inter-study comparability. In this context, a combined qPCR–shotgun metagenomics approach may support early mitigation strategies in aquaculture by improving pathogen detection and providing ecological context relevant to fish health. For this reason, this work explored an innovative, non-invasive approach to monitor the presence of pathogens in aquaculture environments using a previously established method for filtering, storing, and extracting eDNA/eRNA from water samples 18 . In addition to targeted detection, shotgun metagenomics analysis was employed to characterise broader microbial community dynamics, providing insights into potential environmental shifts associated with pathogen presence or other indicators of disturbance (e.g., mucus shedding, stress, and pathogen occurrence). In this study, we used a previously published procedure for detecting various targeted pathogens 18 in both hatchery and seawater sites, where we followed the same Atlantic salmon population. Pathogen surveillance during stressful operational events, such as wellboat delousing, revealed in our study a marked increase in pathogen concentration post-treatment in the water. Moreover, we explored the feasibility of on-site microbial screening testing portable sequencing technologies, such as the Oxford Nanopore MinION, which could support rapid field assessments. In fact, preliminary microbiome analyses, based on metagenomic profiling, offered insights into microbial community dynamics. These included signs of decreased richness and shifts in the virome, which may reflect environmental changes potentially associated with the presence of a salmon pathogen in the water. Moreover, custom pathogen databases were also used to validate outputs from publicly available resources, highlighting the importance of curated reference data in improving the accuracy and relevance of metagenomic pathogen surveillance in aquaculture. The overarching aim of this study was therefore to evaluate the feasibility and performance of a practical, field-deployable workflow for pathogen surveillance in aquaculture, combining onsite filtration with targeted qPCR detection and complementary metagenomic profiling. While the central focus is methodological, we also assessed whether shotgun metagenomics provided additional ecological context — such as microbial community shifts — relevant to pathogen occurrence. This dual focus allowed us to determine both the practical usability of the workflow under real farm conditions and its potential to support broader health-monitoring frameworks. Results 2.1 qPCR / RT-qPCR assays The same Atlantic salmon population was monitored from the hatchery stage (freshwater and smoltification process) through transfer to the seawater site. qPCR screening across all sampling points detected two targeted pathogens: infectious salmon anaemia virus (ISAV) in hatchery inlet-, tank-, and outlet waters, and piscine orthoreovirus genotype 1 (PRV1) in seawater cage samples from August 2023 onwards (Table 1 & Table S2 ). No pathogens were detected in offshore samples (1 km). PRV1 was consistently detected only inside the pens, except during the final sampling event, when low-level PRV1 signals were also observed 200 m from the cage. Additional sampling was performed before and after delousing treatments on wellboats. PRV1 concentrations increased following treatment, likely reflecting enhanced viral shedding associated with fish stress during handling procedures (Table 1 & Table S2 ). Table 1 Cq values and RT-qPCR assays estimated copy numbers per L for ISAV (hatchery) and PRV1 (seawater sites and wellboat). Results are shown as mean ± standard deviation from three pooled biological replicates and two technical replicates. Samples Group Cq Mean ± St. Dev. Copy numbers/ L ± St. Dev. Outlet water (Feb/23) Hatchery 35.39 ± 0.37 3952.00 ± 1012.58 Tank water (Feb/23) Hatchery 35.42 ± 0.09 3824.20 ± 254.28 Inlet water (Feb/23) Hatchery 34.68 ± 0.23 6414.00 ± 1009.75 Control (Feb/23) Hatchery N/A N/A Outlet water (May/23) Hatchery 34.30 ± 0.08 8370.00 ± 444.06 Tank water (May/23) Hatchery 34.25 ± 0.30 8700.00 ± 1753.62 Inlet water (May/23) Hatchery 34.38 ± 0.12 7884.00 ± 661.85 Control (May/23) Hatchery N/A N/A 1km from cage − 1m depth (Apr/23) Seawater sites N/A N/A 1km from cage − 20m depth (Apr/23) Seawater sites N/A N/A Seawater cage (Apr/23) Seawater sites N/A N/A Control (Apr/23) Seawater sites N/A N/A 1km from cage − 1m depth (May/23) Seawater sites N/A N/A 1km from cage − 20m depth (May/23) Seawater sites N/A N/A Seawater cage (May/23) Seawater sites N/A N/A Control (May/23) Seawater sites N/A N/A Seawater cage (Jun/23) Seawater sites N/A N/A Seawater cage (Jul/23) Seawater sites N/A N/A Seawater cage (Aug/23) Seawater sites 38.00 ± 0.66 209.94 ± 91.78 Control (Aug/23) Seawater sites N/A N/A Seawater cage (Sep/23) Seawater sites 31.68 ± 0.06 14826.00 ± 557.20 1km from cage − 1m depth (Oct/23) Seawater sites N/A N/A 1km from cage − 20m depth (Oct/23) Seawater sites N/A N/A Seawater cage (Oct/23) Seawater sites 34.12 ± 0.08 2800.80 ± 162.92 Control (Oct/23) Seawater sites N/A N/A Wellboat before treatment (Nov/23) Wellboat 36.81 ± 0.77 481.78 ± 242.99 Wellboat after treatment (Nov/23) Wellboat 29.93 ± 0.07 48760.00 ± 2319.31 Seawater cage (Dec/23) Seawater sites 32.15 ± 0.02 10770.00 ± 132.94 1km from cage − 1m depth (Feb/24) Seawater sites N/A N/A 1km from cage − 20m depth (Feb/24) Seawater sites N/A N/A Seawater cage (Feb/24) Seawater sites 39.20 ± 1.46 111.30 ± 95.69 Control (Feb/24) Seawater sites N/A N/A 200m from cage − 1m depth (Feb/24) Seawater sites 40.09 ± 0.01 47.80 ± 0.45 200m from cage − 20m depth (Feb/24) Seawater sites N/A N/A Although some detections occurred at high Cq values (> 35), these signals fell within the expected behaviour of low-concentration environmental samples approaching the assay’s limits of detection (LOD). Importantly, low-positive results were reproducible across biological (n = 3) and technical (n = 2) replicates and supported by consistent standard curve–based quantification (Table S2 ), indicating that these detections reflect true low-level environmental shedding rather than analytical noise. 2.2 Sequencing pipeline output and quality assessment From the output of the Oxford Nanopore raw ligation sequencing data processed using the NEPAL pipeline, we evaluated three processing strategies—simplex, duplex, and nanofilt-filtered reads—to assess differences in read quality and downstream performance. Five key metrics were analysed for each run (runs 1–4) using boxplots (Fig. S1 & Table S3 ): 1) average read length, which varied between runs, with duplex reads generally shorter due to trimming during consensus generation; 2) average quality scores (range: Q15–Q21), which differed significantly among workflows, with duplex reads exhibited the highest average quality, while nanofilt results depended on the filtering threshold; 3) GC content (%), which remained relatively stable across workflows and runs (35–50%, as expected for environmental samples); 4) N50 (bp) values used as an indicator of assembly performance, with values exceeding 2,000 bp; and 5) Q30 (%), where duplex reads showed lower proportions of high-confidence bases compared to nanofilt-filtered reads. Based on these results, we proceeded with nanofilt-filtered reads for taxonomic profiling using the taxprofiler pipeline with the Kraken2 database. Although filtering may remove some informative reads, Kraken2 benefits from longer reads and higher Q30 values; therefore, nanofilt-filtered reads were considered more robust for taxonomic classification, offering improved sensitivity and specificity. The taxprofiler output was assessed using standard quality metrics (Fig. S2 & Table S4 ). As expected, control samples exhibited a lower number of reads compared to experimental samples (Fig. S2 & Table S4 ). Most processed reads had a median length between 1,749 and 2,749 bp, which is appropriate for long-read sequencing, especially for environmental or low-input samples (Fig. S2 & Table S4 ). GC content was consistent across all samples (≈ 40–50%), supporting the overall uniformity of library preparation (Fig. S2 & Table S4 ). However, Homo sapiens sequences were detected in several samples, suggesting contamination during sampling rather than during library prep, despite the use of sterile materials (e.g., gloves, sterile filter cups, etc.) (Fig. S2 & Table S4 ). 2.3 Taxonomic classification We first analysed the overall relative abundance (%) among the top six classified phyla, excluding the S. salar genome (taxprofiler pipeline) and H. sapiens contamination. Samples were grouped by environment (hatchery, seawater sites, and wellboat) and controls (Fig. S3 ). Hatchery samples exhibited the lowest proportion of unclassified reads (12.82 ± 6.37%) and minimal variability in “other” taxa (1.88 ± 0.07%) (Fig. S3 & Table S5 ). In contrast, seawater sites showed higher unclassified percentages (23.99 ± 4.71%) and greater variability in “other” taxa (5.03 ± 1.25%) (Fig. S3 & Table S5 ). The wellboat group displayed the greatest variability in taxonomic classification, driven by the second sample collected after treatment, which had an exceptionally high proportion of unclassified reads (43.5%) (Fig. S3 & Table S5 ). Transient increases in the proportion of unclassified reads are common in environmental metagenomics, particularly when community composition shifts toward taxa that are poorly represented in reference databases or that require de novo assembly for accurate annotation 19 . This suggests that treatment may have altered the microbial community or introduced sequences that were more difficult to classify. The Zymo microbial standard positive control was almost fully classified (99.99%), with negligible unclassified or “other” taxa (0.01%) (Fig. S3 & Table S5 ), as expected for a well-characterised mock community. Similarly, the positive control from the spiked pilot experiment 18 was also nearly fully classified (97.56%), with only a small fraction unclassified (1.53%) (Fig. S3 & Table S5 ). These results confirm that the pipeline performed well on known or control samples. A general taxonomic composition profile showed a shift from Actinomycetota (phylum of Gram-positive bacteria with high GC content) and Bacteroidota (Gram-negative bacteria) dominance in the hatchery, to more diverse communities in seawater sites reflecting environmental changes (Fig. S3 ). Hatchery water is nutrient-rich and controlled, favouring the enrichment of heterotrophic degraders typical of recirculating systems 20 , 21 , while seawater sites showed: Chlorophyta peaks during spring/algal bloom periods, indicating seasonal eutrophication and/or phytoplankton proliferation; Nitrososphaerota presence, which suggest active nitrification, likely linked to nitrogen cycling in the marine environment 22 ; and Uroviricota and Nucleocytoviricota , which could be correlated to viral regulation of microbial and algal populations, influencing bloom dynamics and nutrient turnover 23 , 24 (Fig. S3 ). For subsequent comparisons across environments, Archaea and Eukaryota were excluded from the analyses because their high abundance in marine sites, often driven by seasonal factors such as algal blooms, could mask bacterial and viral taxa. It was thus considered that focusing on bacteria, viruses, and bacteriophages would provide a more informative perspective on microbial dynamics throughout the salmon production cycle, from hatchery to seawater sites and ultimately slaughter. Therefore, we showed the relative abundance (%) of the top six genera per sample after nanofilt processing, including “others” (Fig. 1 & Table S6 ). Hatchery samples showed the highest percentage for “other” (ca. 22.08%), but relatively low unclassified reads, while seawater sites had higher unclassified reads (ca. 23.99%), reflecting marine complexity with taxonomic gaps (Table S7 ). The extreme unclassified proportion (43.50%) was driven by the wellboat post-treatment sample, while controls behaved as expected with low unclassified reads except for minor reagent contamination (negative control) (Table S7 ). Hatchery water microbiota is dominated by Mycolicibacterium ( Actinomycetota ), Mycobacterium ( Actinomycetota ) and Flavobacterium ( Bacteroidota ) (Fig. 1 & Table S6 ). In contrast, seawater sites exhibited seasonal shifts that strongly influenced microbial composition: spring blooms (April–May) favoured polysaccharide degraders ( Polaribacter 25 ), along with oligotrophic specialists such as “ Candidatus Pelagibacter” (SAR11 26 ), summer (July–August) was characterised by viral–algal interactions ( Prasinovirus 27 ), autumn (September–October) introduced potential opportunistic pathogens ( Aliivibrio 28 , Pseudalteromonas 29 ), and winter (December–February) was dominated by “ Candidatus Pelagibacter” (SAR11), together with cold-adapted genera like Colwellia 30 and bloom-termination-associated genera such as Kordia 31 . The negative control showed a high proportion of “other” genera (59.94%), likely reflecting low-level contamination commonly reported in extraction kits and reagents. Therefore, potential contaminant taxa should be considered when interpreting low-biomass samples, particularly in the taxonomic composition of wellboat samples after treatment. The pilot experiment 18 was dominated by Yersinia (99.99%), confirming successful sequencing and pipeline accuracy (Fig. 1 & Table S6 ). Moreover, the Zymo microbial standard positive control 32 was characterised by Staphylococcus , Bacillus , Listeria , and Enterococcus in balanced proportions (12–18%), further validating pipeline accuracy (Fig. 1 & Table S6 ). 2.3 Alpha and beta diversities Alpha and beta diversity analyses between hatchery and seawater sites were not significantly different at the phylum or species level; therefore, subsequent analyses focused on the genus level, as this was considered the lowest phylogenetic rank providing reliable information based on the selected standard Kraken2 database. Alpha diversity showed significant differences in Pielou’s evenness, Shannon, and Simpson indices, while richness did not differ significantly between environments (Wilcoxon test; Fig. 2 a). Beta diversity based on Bray–Curtis dissimilarity 33 revealed a clear separation between hatchery and seawater samples (PERMANOVA, R² = 0.323, p = 0.001; Fig. 2 b). Differential abundance analysis of the six most abundant genera using centred log-ratio (CLR) transformation highlighted taxa enriched in hatchery versus seawater sites: hatchery showed higher relative abundance in Aurantimicrobium and Mycolicibacterium , while marine sites in “ Candidatus Pseudothioglobus”, Lyrvirus , Pelagivirus , and Siovirus (Fig. 2 c). Volcano plots of CLR effect sizes (Wilcoxon test 34 ) and false discovery rate (FDR)-adjusted p-values 35 indicated a consistent set of genera significantly different between environments (FDR ≤ 0.1 and |effect size| ≥ 0.5; Fig. 2 d). Effect sizes (CLR) are reported alongside FDR-adjusted p-values in order to highlight not only statistically significant differences, but also the magnitude of compositional shifts between environments. The use of an effect size threshold (|effect size| ≥ 0.5) ensures that only taxa showing biologically meaningful changes are retained for interpretation. We then analysed alpha and beta diversity within each environment, but no significant differences were observed among water sources within the same environment for any metric (Figs. 3 & 4 ). Since marine site samples were initially negative for PRV1 and later tested positive, we investigated whether PRV1 status (positive vs negative) was associated with changes in alpha and beta diversity or differentially abundant taxa in seawater cage samples at the genus level. Alpha diversity did not show significant differences, although a trend toward decreased richness was observed when PRV1 was detected in water samples (Fig. 5 a). Principal Coordinates Analysis (PCoA) based on Bray–Curtis dissimilarity 33 illustrated beta diversity patterns between PRV1 groups, but PERMANOVA results (R² = 0.162, p = 0.14) indicated no significant differences (Fig. 5 b). Differential abundance analysis of the top six genera suggested a potential increase in bacteriophage-associated taxa in PRV1-positive samples (Fig. 5 c), although the volcano plot showed no taxa meeting the significance threshold (FDR ≤ 0.1 and |effect size| ≥ 0.5) (Fig. 5 d). The lack of taxa meeting the significance threshold likely reflects limited effect magnitudes and reduced statistical power, rather than the absence of true biological differences. This is in line with recommendations for microbiome studies, which emphasise the need to interpret effect sizes together with p-values due to the high variability and constrained sample sizes typical of environmental sequencing datasets 36 . 2.4 Databases’ comparison We compared the taxonomic classification using Kraken2 with standard vs custom pathogen reference databases at the species level. The custom pathogen reference database reflects the presence of strain and species-specific Atlantic salmon pathogenic taxa, highlighting the importance of using databases tailored to address the questions at hand. This comparison showed the total absence of Francisella noatunensis , Moritella viscosa and Tenacibaculum piscium in hatchery and seawater site samples by the standard database, while Renibacterium salmoninarum and Tenacibaculum maritimum were identified only from the standard database (Figs. 6 a, 6 b & 6 c). In the wellboat water samples, the standard databases failed to detect most of the pathogenic species, especially after the treatment (Figs. 6 a & 6 c). Therefore, the custom reference database improved the detection of pathogenic species in Atlantic salmon, compared to the standard database. Discussion Infectious diseases represent a major challenge for the Atlantic salmon industry, and pathogen surveillance is a key preventive measure to limit pathogen spread and enable fish farm personnel to act before clinical symptoms appear in the population. Early detection is particularly important when pathogens are shed into the farm environment at high concentrations and persist over time. This study primarily aimed to assess the practicality and performance of a combined qPCR–shotgun metagenomics workflow for routine pathogen surveillance in aquaculture, using farm-deployable filtration and portable sequencing technologies. In addition to demonstrating the feasibility of implementing this approach under real production conditions, we also evaluated whether metagenomic data could provide complementary ecological context, including microbial community patterns potentially associated with pathogen presence. This methodological emphasis guided the design and interpretation of the study. We used a previously established method for filtering, storing, and extracting eDNA/eRNA from water samples 18 and tested it under real farm conditions, with the support of trained fish farm personnel who evaluated the practicality of the filtration procedure for surveillance purposes. This approach was applied while monitoring the same Atlantic salmon population from the hatchery stage through seawater transfer until slaughter. We used specific targeted qPCR assays, and we were able to detect ISAV in all water sources (outlet-, tank- and inlet waters) at similar concentrations during the smoltification process, while PRV1 was detected in water inside the pens from August 2023. PRV1 is highly prevalent in Norwegian aquaculture: the Norwegian Fish Health Report 2024 documents that the virus was detected d at 336 sites, which is a significant increase compared to numbers from 2023 (267), confirming that PRV1 circulates widely in both freshwater and seawater production environments 2 . Surveillance data also show recurrent detections in wild and escaped salmon. Thus, the low-level PRV1 signals identified in our water samples are arguably consistent with what might be expected in terms of background viral prevalence in Norwegian production systems. According to the farm’s fish health records, no notifiable diseases were reported throughout the full production cycle. Routine diagnostics identified low‑level ISAV (ISAV-HRP0) presence in the hatchery and PRV1 at the seawater site, but neither was associated with increased morbidity or mortality. No additional pathogens or infectious events were reported during the entire production cycle. Furthermore, pathogen surveillance during wellboat delousing revealed a marked increase in PRV1 concentration in water samples post-treatment. Environmental pathogen surveillance inherently involves low target concentrations, which frequently produce high Cq values near the LOD. These signals must be interpreted with caution; however, when reproducible across biological replicates and consistent with expected shedding patterns, they remain informative for early detection purposes. None of the targeted pathogens were detected outside the pens, highlighting the importance of sampling location for accurate pathogen surveillance in open-water cages. Hydrodynamic transport between the sea pens and surrounding waters is expected in open water aquaculture systems. Although current data for the site were not available, the low pathogen loads detected suggest that any water exchange did not obscure the differentiation between pen-adjacent and offshore samples. Future studies incorporating site-specific current measurements could help refine the spatial interpretation of waterborne pathogen signals. We also analysed the same water samples for microbial community composition in the laboratory to assess the feasibility of using portable sequencing technologies, such as the Oxford Nanopore MinION, to evaluate whether indicators of balanced or disrupted health in fish farms could support targeted pathogen-level screening. The hatchery exhibited a predominance of phyla associated with nutrient-rich environments 20 , 21 , whereas seawater showed an expected seasonality in microbial community composition 22 , 23 . Across taxonomic levels, seawater cage sites consistently displayed higher diversity and evenness, particularly at the order, family, and genus levels, compared to the hatchery environment. Richness remained similar between environments at the genus level, suggesting that observed differences were driven by community structure rather than species count. However, alpha and beta diversity showed no significant differences among water sources within the same environment. PRV1-positive samples exhibited trends such as decreased microbial richness and alterations in the virome. Marine environments harbour a substantially higher abundance and diversity of bacteriophages, and their presence varies according to season, temperature and photic shifts, and according to targeted hosts, such as the “ Candidatus Pelagibacter” (SAR11) in cold water habitats 37 . Bacteriophages play a pivotal role in marine microbial food webs by regulating bacterial mortality, recycling nutrients through the viral shunt, and influencing biogeochemical cycles, particularly carbon and nitrogen fluxes 38 . In our study, the differential abundance of Llyrvirus , Pelagivirus and Siovirus was significantly higher in seawater sites than in the hatchery, while PRV1-positive samples showed an increased differential abundance of Igirivirus , Powvirus , and Sednavirus . All these viruses are marine phages belonging to the class Caudoviricetes (tailed dsDNA phages) 39 . These observations suggest possible early indications of microbiome and virome shifts, which might be associated with pathogen presence. However, given the low pathogen loads (high Cq values), the stable health status of the fish population, and the absence of clinical disease outbreaks, these patterns should be interpreted as preliminary and potentially influenced by seasonal environmental variability rather than as definitive signs of dysbiosis. The key limitation of our study was that the selected fish groups exhibited good health status, which significantly constrained our ability to establish meaningful correlations between pathogen presence and indicators of health or dysbiosis based on microbiome composition. This uncertainty highlights the need for future studies conducted under more variable or compromised health conditions, where stronger pathogen signals or disease events may allow clearer interpretation of community-level ecological changes. Such studies will be essential to determine whether the subtle trends observed here represent biologically relevant responses or fall within the expected range of natural variability. While low-level human DNA contamination was detected in some samples, this signal was taxonomically restricted and completely removed through explicit filtering of all human-associated reads. Importantly, no evidence of cross-contamination with any of the microorganisms analysed in this study was observed: all negative controls remained free of target pathogens, and microbial community profiles showed coherent ecological patterns rather than signatures of random contamination. Therefore, contamination is unlikely to have influenced the prokaryotic or viral fractions retained for downstream analysis. Although shotgun metagenomics can also be used for antimicrobial resistance (AMR) and virulence gene profiling, this was not included in our analyses because sequencing depth and filtering parameters were optimised for taxonomic classification rather than functional annotation. In addition, the presence of human-associated contamination in some samples could confound AMR interpretation. Future studies with higher sequencing depth and stricter assembly criteria will be required to robustly integrate AMR screening into this workflow. Moreover, although Streptomyces was detected in some samples, no antibiotic treatments were applied during the production cycle. Streptomyces is commonly found in aquatic and soil-derived environmental microbiota, and its presence likely reflects natural background communities rather than the use of antimicrobial agents. Furthermore, no high-quality metagenome‑assembled genomes (MAGs) were reconstructed in this study. The sequencing depth achievable with portable field‑deployable Nanopore runs, together with the high complexity and viral load of marine water samples, was insufficient for robust genome binning. Higher‑coverage sequencing and longer runs will be required to recover MAGs in future applications. Despite all limitations, we successfully established a protocol to extract sufficient DNA from water samples for shotgun metagenomic sequencing, where pipeline quality metrics and taxonomic classification of positive controls confirmed the robustness of the procedure. Moreover, our study emphasised two critical aspects: (i) the persistent issue of human contamination in field samples and its associated microbiota, despite precautions to maintain sterile conditions, and (ii) the importance of reference database selection for achieving comprehensive and accurate taxonomic profiling. Most published studies rely on 16S rRNA gene sequencing 12 – 14 , which provides only a bacterial community snapshot and does not capture viral components. In contrast, our shotgun metagenomics approach revealed shifts in the virome, highlighting the ecological relevance of viruses and bacteriophages in farm water, where they interact closely with fish and their mucosal surfaces. Furthermore, we demonstrate that this workflow can be integrated into targeted DNA-based pathogen screening by using a curated reference database containing high-quality genomes of Atlantic salmon pathogens. These curated entries were added to the Kraken2 library to improve classification accuracy. However, sequence matches alone should not be interpreted as definitive evidence of pathogen presence, as short regions of high similarity may reflect closely related environmental taxa rather than the target pathogen. This was evident when comparing taxonomic classifications using Kraken2 with the standard PlusPFP database vs our custom reference database. This comparison has important implications for practical pathogen surveillance in aquaculture. Standard reference databases such as PlusPFP are highly comprehensive but may occasionally produce false positives for high-consequence pathogens — for example, R. salmoninarum — due to short conserved genomic regions shared with related environmental taxa rather than true pathogen presence. In contrast, curated pathogen-specific databases substantially reduce this risk by restricting the search space to well-annotated, biologically relevant genomes. Our custom database was assembled using stringent criteria, including the selection of complete or near-complete assemblies from NCBI, the exclusion of poorly annotated or fragmented genomes, and the retention only of taxa with confirmed pathogenic relevance for Atlantic salmon. This targeted approach enhances classification accuracy, reduces taxonomic ambiguity, and provides a more reliable foundation for DNA-based pathogen surveillance in open water farming systems. Taken together, our findings suggest that this integrated qPCR–metagenomics workflow is likely to be most informative under conditions where pathogen activity or fish stress is elevated — for example, during known or suspected outbreaks, after events such as delousing or grading, or when monitoring multiple cohorts or farms simultaneously. Additional validation across diverse production systems, health statuses, and environmental conditions will be essential to determine the sensitivity, robustness, and operational value of this approach in routine farm-level decision making. As a conclusion, our study establishes a baseline for future applications of continuous pathogen monitoring using water samples throughout the production cycle via qPCR, combined with microbiome profiling and adaptive sampling pipelines supported by custom pathogen reference databases. While shotgun metagenomics can reveal ecological patterns relevant for understanding disease risk, its application in low-biomass water matrices requires careful contamination control and conservative interpretation. In contrast, targeted qPCR provides robust and sensitive pathogen detection even under these conditions and remains the most appropriate tool for routine surveillance. Methods 4.1 Sampling workflow Water samples were collected from two distinct environments associated with the same Atlantic salmon population (from ca. 250 g to 4–6 kg): a hatchery (including inlet-, tank-, and outlet waters) and the marine grow-out site. In the marine environment, routine monitoring samples were collected from within the sea pen and from two offshore reference locations situated 1 km away from the pen, at depths of 1 m and 20 m. Additionally, during the final sampling, two further offshore samples were collected at 200 m from the pen, also at 1 m and 20 m depth, in order to increase the spatial resolution around the farm during this time point (Fig. 7 ). In the hatchery, tank‑associated water was collected from a fixed sampling point located immediately adjacent to the tanks, corresponding to the outflow water that had passed through the tank system. At the marine site, water was collected from the dead‑fish collection point, which draws water directly from inside the sea pen and is routinely used by the farm for monitoring purposes, or offshore from a service boat. These sampling points provided water representative of the environment immediately surrounding the fish. Environmental variables (e.g., temperature, salinity, dissolved oxygen) were not available from the farm and could therefore not be incorporated. Given the nature of the sampling procedure and the focus of the study on pathogen detection rather than environmental drivers, this limitation does not influence the interpretation of the results. Additional samples were taken on the wellboat immediately before and after delousing. For each sampling event and location, three biological replicates of 500 mL were collected and filtered on site, together with a 500 mL sterile water control (autoclaved and 0.22 µm-filtered laboratory-grade water). This control was included to detect any contamination introduced during field filtration or laboratory processing (filtered using the same equipment, filters, and workflow as the samples). The choice of 500 mL per sample was based on previous method development work, where this volume provided the best balance between sensitivity and filtration efficiency 18 . Larger volumes increased the risk of membrane clogging and did not improve detection probability, whereas 500 mL was sufficient to detect even low amounts of viral and bacterial pathogens in field conditions. All replicates were processed using the same filtration setup, on-site and based on the results from a previous work 18 : a combination of the Mixed Cellulose Ester membrane filter (MF-Millipore® Membrane Filters, 0.45 µm pore size, 47 mm diameter, hydrophilic, Millipore, USA, provided by Merck Life Science AS, Norway) and Glass Fibre Filter (2.0 µm pore size, hydrophilic glass fibre with a binder resin, and 47 mm diameter, Millipore, USA, provided by Merck Life Science AS, Norway) were used in place of the original Nalgene™ filters within the Single Use Analytical Filter Funnels (Thermo Fisher Scientific, USA) (Fig. 7 ). 500 mL of sampled water and 500 mL of sterile water (control) were filtered with a flow rate setting of 3.8 to 4.0 L/min with the EZ-Stream vacuum pump (Millipore, USA) (Fig. 7 ). All three biological replicates were analysed by targeted qPCR and (RT)qPCR in duplicate reactions, whereas shotgun metagenomics was performed on one pooled extract per sampling location, generated by combining the three biological replicates to ensure sufficient DNA yield for sequencing. These technical duplicates should not be interpreted as independent biological replicates, which were instead represented by the three separately collected and processed 500 mL water samples per location. This approach provided complete qPCR coverage across all environments and time points, while delivering representative metagenomic profiles for hatchery samples, all seawater sampling events, and both wellboat samples. 4.2 eDNA/eRNA extraction and qPCR / RT-qPCR assays Automated DNA and viral RNA extractions were performed on a MagNA Pure 96 instrument (Roche) with the MagNA Pure 96 DNA and Viral NA Large Volume Kit (Roche, France), using Pathogen Universal LV protocol with a sample input volume of 1000 µL and an elution volume of 100 µL, per sample, respectively. The pathogen panel screened at the hatchery and seawater sites was defined by the collaboration agreement with the farm. Only pathogens included in the farm’s authorised diagnostic panel could be tested, and the authors were not permitted to add additional targets. As a result, the hatchery screening did not include some freshwater-specific bacterial pathogens. This study followed the farm’s established surveillance procedures, in line with its methodological focus on evaluating a field-deployable qPCR–metagenomics workflow rather than conducting a full epidemiological assessment. In the hatchery, screening included ISAV, PRV1, piscine myocarditis virus (PMCV), infectious pancreatic necrosis virus (IPNV), and Yersinia ruckeri , while at marine sites the panel comprised ISAV, PRV1, PMCV, IPNV, salmonid alphavirus (SAV), and Paramoeba perurans . RT-qPCR was carried out for ISAV 40 , PRV1 41 , PMCV 42 and SAV 43 with the Brilliant III Ultra-Fast QRT-PCR Master Mix (Agilent Technologies, USA) using the following protocol: 10 µL of 2× QRT-PCR master mix, 0.2 µL of 100 mM DTT, 1 µL of RT/RNase block, 1 µL of the 20× viral assay (6 µM of assay probe, 10 µM of each forward and reverse primers), and 5 µL extracted viral RNA. Thermocycling was performed on a CFX384 Bio-Rad and CFX-manager (software version 3.1, Bio-Rad, USA) under the following conditions: 10 min at 50°C, 3 min at 95°C, 45 cycles of 5 s at 95°C and 10 s at 60°C. Moreover, IPNV 44 RT-qPCR was performed with the QIAGEN OneStep RT-PCR Kit (Qiagen, USA) using the following protocol: 4 µL of 5× QIAGEN OneStep RT-PCR Buffer, 0.8 µL of dNTP Mix (containing 10 mM of each dNTP), 1 µL of MgCl 2 (25 mM), 1 µL of the 20× viral assay (6 µM of assay probe, 10 µM of each forward and reverse primers), 0.1 µL of RNaseOUT™ Recombinant Ribonuclease Inhibitor (40 U/µL, Invitrogen™, USA), 0.8 µL of QIAGEN OneStep RT-PCR Enzyme Mix, 7.3 µL of nuclease-free water, and 5 µL extracted viral RNA. Thermocycling was performed on a CFX384 Bio-Rad and CFX-manager (software version 3.1, Bio-Rad, USA) under the following conditions: 30 min at 50°C, 15 min at 95°C, 45 cycles of 30 s at 94°C and 60 s at 60°C. qPCR was carried out for Y. ruckeri 45 with TaqMan™ Fast Advanced Master Mix (Applied Biosystems™, USA) using the following protocol: 10 µL of 2× TaqMan Fast Advanced Master Mix, 5 µL of the bacterial assay (10 µM of assay probe and each forward and reverse primers), and 5 µL extracted DNA. Thermocycling was performed on a CFX384 Bio-Rad and CFX-manager (software version 3.1, Bio-Rad, USA) under the following conditions: 2 min at 50°C, 20 s at 95°C, 50 cycles of 3 s each at 95°C, and 20 s at 62°C. Specific qPCR assay for P. perurans was carried out with 25 µL reactions consisting of 12.5 µL TaqPath™ qPCR Master Mix, CG (Applied Biosystems™, USA), 500 nM of each primer and 250 nM of probe, nuclease-free water and 5 µL DNA sample 46 . The following qPCR cycling conditions were used: an initial denaturation at 95°C for 20 s, followed by 50 cycles of denaturation at 95°C for 3 s and annealing at 60°C for 30 s 46 . All qPCR assays were analysed in duplicate, and a no-template control (H 2 O) and water sample control from the MagNA Pure 96 extraction protocol were included on each plate as negative controls. Standard curves of serial dilution of the relevant target sequences (CFX-manager software version 3.1, Bio-Rad, USA) were used for the calculation of the concentration used for the experiments. Primers and probes are shown in the supplementary (Supplementary Table S1 ). Full standard curves, including amplification efficiencies, R² values, slope, and y-intercept, are provided in the Supplementary Material (Table S2 ). These curves were generated using the same matrix composition and extraction conditions as the environmental samples and were used to calculate copy numbers and estimate assay-specific LOD. Positive qPCR results for ISAV and PRV1 were then repeated to estimate the copy number in 5 µL with the Luna Probe One-Step RT-qPCR 4X Mix with UDG (New England Biolabs, USA) using the following protocol: 5 µL of 4× Luna Probe One-Step RT-qPCR 4X Mix with UDG, 2 µL of the viral assay (0.4 µL of 10 µM of assay probe, 0.8 µL of 10 µM of each forward and reverse primers), 5 µL extracted viral RNA, and 8 µL of nuclease-free water. Thermocycling was performed on a CFX384 Bio-Rad and CFX-manager (software version 3.1, Bio-Rad, USA) under the following conditions: 30 s at 25°C, 10 min at 55°C, 1 min at 95°C, 45 cycles of 10 s at 95°C and 30 s at 60°C. The copy number/µL was calculated using a previously designed sequence-verified gene fragment, or gBlocks™ (Integrated DNA Technologies, BVBA, Leuven, Belgium) 18 . Briefly, the concentration from ng/µL was converted to copy number/µL by using the formula provided by IDT guidelines. Copy numbers for ISAV and PRV1 quantification were calculated using the standard Integrated DNA Technologies (IDT) formula: copies/µL = (DNA concentration [ng/µL]) × (Molecular weight [fmol/ng) × (1 × 10 − 15 ) × (6.022 × 10²³). The synthetic gBlock used in this study had a concentration of 8.408 ng/µL and a molecular weight of 2.36 fmol/ng, corresponding to an estimated 1.19 × 10¹⁰ copy number/µL in the undiluted gBlock. A stock solution of 5 × 10⁷ copy number/µL was prepared by combining 8.4 µL of the undiluted gBlock with 1991.6 µL of TE buffer. Serial 1:4 dilutions were then generated to obtain a standard curve ranging from 5 × 10⁷ to 1 × 10⁻² copy number/µL. Expected copy numbers for each dilution step (1 µL and 5 µL input) are provided in Supplementary Table S2 . All standards were prepared using the same extraction matrix and buffer conditions as the environmental samples to ensure comparable amplification efficiencies. After optimisation, the Stock − 4 dilution (1.56 x 10 4 copy number/µL) was used as a standard template to estimate the RNA copy numbers/µL for the viral pathogens (ISAV and PRV1) based on the amount captured by the filters (Table S2 ). To aid interpretation, qPCR copy numbers initially expressed per 5 µL of extract were converted into copies per litre of water (Table S2 ). Based on the standard curves generated under the same matrix and extraction conditions, the estimated limits of detection were ~ 100 copies per reaction for ISAV and ~ 10 copies per reaction for PRV1, corresponding to ~ 4×10³ copies/L and ~ 4×10² copies/L, respectively. Several detections, particularly those with Cq > 35, fell close to these thresholds; therefore, estimates at low copy number should be interpreted cautiously due to increased stochastic variation and reduced quantification precision. Despite this uncertainty, ISAV and PRV1 detections near the LOD were consistent across biological replicates, supporting their biological relevance in the context of environmental shedding. 4.3 DNA extraction and sequencing For microbial community analysis, 4.5 mL of preserved pooled DNA/RNA in DNA/RNA Shield™ was extracted from each sampling point with the ZymoBIOMICS™ DNA/RNA Miniprep Kit (Zymo Research, USA, provided by Nordic Biosite AS, Sweden), following the manufacturer's specifications. Briefly, 750 µL of liquid sample were transferred into a ZR BashingBead Lysis Tube (0.1 & 0.5 mm) and homogenised with FastPrep®-24 5G bead beating grinder and lysis system (MP Biomedicals, USA) using the following program: 2 cycles for 60 s (300 s of pause), speed of 6.0 m/s – Lys Matrix A, 1 mg. Samples were then centrifuged at 15,000 x g for 1 min and mixed with an equal volume (1:1) of DNA/RNA Lysis Buffer to the supernatant and processed through the DNA and RNA purification kit specifications. To elute the nucleic acids, 80 µL of nuclease-free water was added directly to the column matrix, incubated for 5 min, and then centrifuged at 15,000 x g for 2 min. The final step with the Zymo-Spin™ III-HRC Filter was also performed to remove inhibitors, salts, enzymes, or toxic compounds from the solution containing nucleic acids for sequencing purposes. Sample integrity and fragment size were assessed using the Genomic DNA (gDNA) ScreenTape assay on the TapeStation 4150 System (Agilent Technologies, USA). DNA Integrity Numbers (DIN) from water samples were greater than 6, with an average concentration of 10 ng/µL and a fragment size of approximately 10,000 base pairs. For sequencing, the Ligation Sequencing Kit V14 – PCR Barcoding protocol (SQK-LSK114 with EXP-PBC001, Oxford Nanopore Technologies, UK) was applied according to the manufacturer’s instructions, with some modifications: 45 µL of each sample were used as input material, 18 cycles were performed during the PCR barcoding step, and 50 fmol were loaded onto the flow cell. Sequencing was carried out using a MinION Flow Cell (R10.4.1) on the MinION Mk1B device, with the MinKNOW™ software version 25.09.16, excluding basecalling and saving pod5 files. Each sequencing run included two positive controls: 100 ng of the ZymoBIOMICS® Microbial Community DNA Standard (Zymo Research, USA; supplied by Nordic Biosite AS, Sweden), and 45 µL of a sample from a previously published pilot experiment spiked with known concentrations of selected pathogens 18 . Negative controls were also included for each sampling point (sterile Milli-Q water filtered on site), along with an extraction control containing only DNA/RNA Shield™. 4.4 Bioinformatic analysis Oxford Nanopore ligation sequencing data were processed using the NEPAL pipeline (Norwegian Veterinary Institute) 47 , a reproducible workflow built on Nextflow for processing raw nanopore data. The pipeline was configured to evaluate three processing strategies: simplex, duplex, and NanoFilt-filtered reads, enabling comparative analysis of read quality and downstream performance. Basecalling was performed using Dorado v0.2.1 (Oxford Nanopore Technologies), integrated into the NEPAL pipeline. The super accuracy model [email protected] was used to ensure the highest possible basecalling accuracy. NanoFilt was applied to duplex reads following the filtering parameters: minimum quality score: 9; minimum read length: 300 bp (to exclude very short fragments); and maximum read length: 2,147,483,647 bp (a technical upper limit to avoid excluding long reads). Outputs from each workflow were compared in terms of read quality, length distribution, GC content, and assembly metrics. Statistical analysis among the three processing strategies was performed using the Kruskal–Wallis non-parametric test 48 and displayed on the plots. When the Kruskal–Wallis test indicated significance (p ≤ 0.05), Wilcoxon rank-sum tests 34 were performed for pairwise comparisons, and p-values were adjusted using the Benjamini–Hochberg (BH) method 35 to control the false discovery rate. After NanoFilt filtering, taxonomic classification was performed using the nf-core 49 taxprofiler pipeline with Kraken2 50 as the classifier. The pipeline was configured to use the Kraken2 PlusPFP database ( https://benlangmead.github.io/aws-indexes/k2 ) and the long-read mode (--long) for optimal performance on nanopore data. To avoid host contamination, sequences from the Atlantic salmon genome ( S. salar ; reference: Ssal_v3.1; GCA_905237065.2) were excluded from the database. The pipeline generated taxonomic profiles at multiple ranks (domain, phylum, order, class, family, genus, and species) and summary reports for downstream comparative analysis. Statistical analyses and visualisation were performed in R (v4.4.0) 51 using RStudio (v2024.04.0 + 735) 52 . After an initial taxonomic composition assessment, reads classified as H. sapiens and all sequences assigned to Eukaryota or Archaea were excluded to focus on prokaryotic and viral fractions relevant to fish health analyses. In the taxonomic workflow, human-associated taxa (e.g., H. sapiens , Homo , Hominidae) were explicitly removed during data processing. In the diversity workflow, the entire Eukaryota domain was filtered out in R, which also eliminated human-derived and other nonbacterial/nonviral reads. As a result, only non-host and non-human microbial taxa were retained for all taxonomic, alpha-diversity, beta-diversity, and differential abundance analyses. Alpha diversity indices—species richness, Shannon diversity, Simpson diversity, and Pielou’s evenness—were calculated from relative abundances. Group comparisons (e.g., hatchery vs seawater sites or water sources in each environment) used Wilcoxon rank-sum 34 or Kruskal–Wallis 48 tests with FDR correction (Benjamini–Hochberg 35 ). Beta diversity was based on Bray–Curtis dissimilarity 33 and visualised with PCoA. Group separation was assessed using PERMANOVA ( adonis2 53 ) and multivariate dispersion tests ( betadisper 54 ), as implemented in the vegan R package 55 . Confidence ellipses (95%) were drawn for visual separation among environments. Differential abundance analyses were assessed with centred log-ratio (CLR) transformation and Dirichlet-Monte-Carlo model using the ALDEx2 package 56 , a negative binomial model to raw counts was used to estimate log₂ fold changes between groups with the DESeq2 package 57 , and non-parametric comparison of log-ratio-transformed relative abundances were assessed with Wilcoxon rank-sum 34 test with FDR correction (Benjamini–Hochberg 35 ) (FDR ≤ 0.10 and |effect size| ≥ 0.5). We applied an FDR threshold of ≤ 0.10, consistent with recommendations for flexible false discovery control in large-scale omics analyses, where rigid cut-offs (e.g., FDR ≤ 0.05) may be overly conservative and reduce sensitivity in high-dimensional datasets 58 . We report effect sizes alongside p-values, following recommendations that highlight the importance of effect magnitude in interpreting microbiome analyses, particularly in studies where statistical power may be limited 36 . A Custom Kraken2 database was developed to compare the taxonomic classification of fish pathogens with the standard Kraken2 PlusPFP database. The Standard PlusPFP database includes bacterial, archaeal, viral, fungal, and protozoan genomes distributed with Kraken2. The Custom database was built with reference genomes and complete assemblies of known Atlantic salmon pathogens, including R. salmoninarum , Salmon gill poxvirus, Phocoenobacter skyensis , Phocoenobacter atlanticus subsp. atlanticus , Y. ruckeri , M. viscosa , Aeromonas salmonicida subsp. salmonicida , Piscirickettsia salmonis , T. maritimum , Tenacibaculum dicentrarchi , T. piscium , Tenacibaculum finnmarkense , and F. noatunensis . Genomes were retrieved from NCBI RefSeq/GenBank, and they were manually curated to ensure completeness and taxonomic accuracy. The new entries were added to the Kraken2 library, and the database index was rebuilt. Comparative visualisations (species-level relative abundance barplots and Venn overlaps) were generated in R to identify taxa uniquely detected by either database. Declarations Competing interests The author(s) declare no competing interests. Funding Declaration The study was funded by the Research Council of Norway (RCN) through the project “SusOffAqua - Unleashing the sustainable value creation potential of offshore ocean aquaculture” (project number 328724) and by the FHF - Fiskeri- og havbruksnæringens forskningsfinansiering through the project “Biosikkerhetstiltak mot ILA i settefisk (ILA-SAFE)” (project number 901674). This work was co-funded by the European Union's Horizon Europe Project 101136346 EUPAHW. Views and opinions expressed are, however, those of the author(s) only and do not necessarily reflect those of the European Union or the European Research Executive Agency. Neither the European Union nor the granting authority can be held responsible for them. Author Contribution Study conceptualisation (O.B., D.S., H.S., and T.V.), experimental work (O.B., S.N.M., M.M.A., and D.S.), data analysis and visualisation (O.B. and S.G.), manuscript writing first draft (O.B., H.S. and T.V.), manuscript editing (O.B., D.S., S.G., M.M.A., S.N.M., H.S., and T.V.). All authors reviewed, edited, and approved the final manuscript. Acknowledgement We thank the staff at the hatchery and open seawater cage for the excellent collaboration. The computations were performed on resources provided by Sigma2 - the National Infrastructure for High-Performance Computing and Data Storage in Norway (project nn10070k). Data Availability The datasets generated during and/or analysed during the current study are available at the BioProject PRJNA1379277. All other data generated during this study are included in this published article and its Supplementary Information files. References FAO. International markets for fisheries and aquaculture products – Fourth issue 2024, with January–June 2024 statistics. (2024). Moldal, T., Wiik-Nielsen, J., Oliveira, V. H. S., Svendsen, J. C. & Sommerset, I. Norwegian Fish Health Report 2024. Norwegian Veterinary Institute Report series #1b/2025 . (Norwegian Veterinary Institute, 2025). Bernhardt, L. V. et al. Early detection of salmonid alphavirus in seawater from marine farm sites of Atlantic salmon Salmo salar . Diseases of Aquatic Organisms 146 , 41-52 (2021). https://doi.org:10.3354/dao03618 Bernhardt, L. V., Myrmel, M., Lillehaug, A., Qviller, L. & Weli, S. C. Filtration, concentration and detection of salmonid alphavirus in seawater during a post-smolt salmon ( Salmo salar ) cohabitant challenge. Diseases of Aquatic Organisms 144 , 61-73 (2021). https://doi.org:10.3354/dao03572 Weli, S. C., Bernhardt, L. V., Qviller, L., Myrmel, M. & Lillehaug, A. Development and evaluation of a method for concentration and detection of salmonid alphavirus from seawater. J Virol Methods 287 (2021). https://doi.org:10.1016/j.jviromet.2020.113990 Bohara, K., Yadav, A. K. & Joshi, P. Detection of Fish Pathogens in Freshwater Aquaculture Using eDNA Methods. Diversity 14 (2022). https://doi.org:10.3390/d14121015 Arbon, P. et al. Passive eDNA sampling to monitor infectious hypodermal and haematopoietic necrosis virus (IHHNV) in pond-reared giant black tiger shrimp ( Penaeus monodon ) aquaculture systems. Aquaculture , 742709 (2025). https://doi.org:10.1016/j.aquaculture.2025.742709 Krishna, D. et al. Environmental DNA/RNA for non-invasive early detection and monitoring of pathogen dynamics in Atlantic salmon ( Salmo salar ) recirculating aquaculture systems (RAS). Aquaculture 611 (2025). https://doi.org:10.1016/j.aquaculture.2025.743060 Pires, N. M. M., Dong, T., Yang, Z. & da Silva, L. F. B. A. Recent methods and biosensors for foodborne pathogen detection in fish: Progress and future prospects to sustainable aquaculture systems. Critical Reviews in Food Science and Nutrition 61 , 1852-1876 (2021). https://doi.org:10.1080/10408398.2020.1767032 Grunow, B. & Strauch, S. M. Status assessment and opportunities for improving fish welfare in animal experimental research according to the 3R-Guidelines. Rev Fish Biol Fisher 33 , 1075-1093 (2023). https://doi.org:10.1007/s11160-023-09781-8 Gundi, V. A. K. B. et al. Microbial pathogens in aquaculture: a review of emerging threats. Academia Biology 3 (2025). https://doi.org:10.20935/AcadBiol7814 Rieder, J., Kapopoulou, A., Bank, C. & Adrian-Kalchhauser, I. Metagenomics and metabarcoding experimental choices and their impact on microbial community characterization in freshwater recirculating aquaculture systems. Environmental Microbiome 18 , 8 (2023). https://doi.org:10.1186/s40793-023-00459-z Macedo, D. B. et al. New perspectives on metagenomic analysis for pathogen monitoring in sustainable freshwater aquaculture production: a systematic review. Frontiers in Freshwater Science 2 , 1459233 (2024). https://doi.org:10.3389/ffwsc.2024.1459233 Valenzuela-Miranda, D. et al. Nanopore-based metagenomic approaches for detection of bacterial pathogens in recirculating aquaculture systems. Fishes-Basel 10 , 496 (2025). https://doi.org:10.3390/fishes10100496 Acharya, K. et al. Metagenomic water quality monitoring with a portable laboratory. Water Research 184 , 116112 (2020). https://doi.org:10.1016/j.watres.2020.116112 Djikeng, A., Kuzmickas, R., Anderson, N. G. & Spiro, D. J. Metagenomic analysis of RNA viruses in a fresh water lake. PLOS One 4 , e7264 (2009). https://doi.org:10.1371/journal.pone.0007264 Liu, Z. et al. Comparative metagenomic analysis of microbial community compositions and functions in cage aquaculture and its nearby non-aquaculture environments. Front Microbiol 15 , 1398005 (2024). https://doi.org:10.3389/fmicb.2024.1398005 Benedicenti, O. et al. A refinement to eRNA and eDNA-based detection methods for reliable and cost-efficient screening of pathogens in Atlantic salmon aquaculture. PLOS One 19 , e0312337 (2024). https://doi.org:10.1371/journal.pone.0312337 Saggu, S. K., Kumar, M. & Kumar, S. Metagenomics and its impact on environmental and therapeutic microbiology. Archives of Microbiology 208 , 24 (2026). https://doi.org:10.1007/s00203-025-04588-6 Minich, J. J. et al. Microbial ecology of Atlantic salmon ( Salmo salar ) hatcheries: impacts of the built environment on fish mucosal microbiota. Appl Environ Microb 86 , e00411-00420 (2020). https://doi.org:10.1128/AEM.00411-20 Fiedler, A. W., Drågen, M. K. R., Lorentsen, E. D., Vadstein, O. & Bakke, I. The stability and composition of the gut and skin microbiota of Atlantic salmon throughout the yolk sac stage. Front Microbiol 14 , 1177972 (2023). https://doi.org:10.3389/fmicb.2023.1177972 Wuchter, C. et al. Archaeal nitrification in the ocean. Proceedings of the National Academy of Sciences 103 , 12317-12322 (2006). https://doi.org:10.1073/pnas.0600756103 Zhou, K. et al. Genomic and transcriptomic insights into complex virus–prokaryote interactions in marine biofilms. The ISME Journal 17 , 2303-2312 (2023). https://doi.org:10.1038/s41396-023-01546-2 Ha, A. D., Moniruzzaman, M. & Aylward, F. O. Assessing the biogeography of marine giant viruses in four oceanic transects. ISME Communications 3 , 43 (2023). https://doi.org:10.1038/s43705-023-00252-6 Avcı, B., Krüger, K., Fuchs, B. M., Teeling, H. & Amann, R. I. Polysaccharide niche partitioning of distinct Polaribacter clades during North Sea spring algal blooms. The ISME Journal 14 , 1369-1383 (2020). https://doi.org:10.1038/s41396-020-0601-y Bolaños, L. M. et al. Influence of short and long term processes on SAR11 communities in open ocean and coastal systems. ISME Communications 2 , 116 (2022). https://doi.org:10.1038/s43705-022-00198-1 Weynberg, K. D., Allen, M. J. & Wilson, W. H. Marine prasinoviruses and their tiny plankton hosts: a review. Viruses 9 , 43 (2017). https://doi.org:10.3390/v9030043 Kashulin, A., Seredkina, N. & Sørum, H. Cold‐water vibriosis. The current status of knowledge. Journal of Fish Diseases 40 , 119-126 (2017). https://doi.org:10.1111/jfd.12465 Richards, G. P. et al. Mechanisms for Pseudoalteromonas piscicida -induced killing of vibrios and other bacterial pathogens. Appl Environ Microb 83 , e00175-00117 (2017). https://doi.org:10.1128/AEM.00175-17 Deming, J. W., Somers, L. K., Straube, W. L., Swartz, D. G. & Macdonell, M. T. Isolation of an obligately barophilic bacterium and description of a new genus, Colwellia gen. nov. Syst Appl Microbiol 10 , 152-160 (1988). https://doi.org:10.1016/S0723-2020(88)80030-4 Bigalke, A., Meyer, N., Papanikolopoulou, L. A., Wiltshire, K. H. & Pohnert, G. The algicidal bacterium Kordia algicida shapes a natural plankton community. Appl Environ Microb 85 , e02779-02718 (2019). https://doi.org:10.1128/AEM.02779-18 Nicholls, S. M., Quick, J. C., Tang, S. & Loman, N. J. Ultra-deep, long-read nanopore sequencing of mock microbial community standards. GigaScience 8 , giz043 (2019). https://doi.org:10.1093/gigascience/giz043 Bray, J. R. & Curtis, J. T. An ordination of the upland forest communities of southern Wisconsin. Ecological Monographs 27 , 326-349 (1957). https://doi.org:10.2307/1942268 Wilcoxon, F. Individual comparisons by ranking methods. Biometrics Bulletin 1 , 80-83 (1945). https://doi.org:10.2307/3001968 Benjamini, Y. & Hochberg, Y. Controlling the False Discovery Rate: A Practical and Powerful Approach to Multiple Testing. Journal of the Royal Statistical Society. Series B (Methodological) 57 , 289-300 (1995). https://doi.org:10.1111/j.2517-6161.1995.tb02031.x Ferdous, T. et al. The rise to power of the microbiome: power and sample size calculation for microbiome studies. Mucosal Immunology 15 , 1060-1070 (2022). https://doi.org:10.1038/s41385-022-00548-1 Coutinho, F. H. et al. Marine viruses discovered via metagenomics shed light on viral strategies throughout the oceans. Nature Communications 8 , 15955 (2017). https://doi.org:10.1038/ncomms15955 Mayers, K. M. J. et al. Grazing on Marine Viruses and Its Biogeochemical Implications. mBio 14 , e01921-01921 (2023). https://doi.org:10.1128/mbio.01921-21 Calayag, A. M. et al. Arctic Ocean virus communities and their seasonality, bipolarity, and prokaryotic associations. Nature Communications 16 , 6427 (2025). https://doi.org:10.1038/s41467-025-61568-6 Snow, M. et al. Development, application and validation of a Taqman real-time RT-PCR assay for the detection of infectious salmon anaemia virus (ISAV) in Atlantic salmon ( Salmo salar ). Developments in Biologicals 126 , 133-145; discussion 325-136 (2006). Palacios, G. et al. Heart and Skeletal Muscle Inflammation of Farmed Salmon Is Associated with Infection with a Novel Reovirus. PLOS One 5 (2010). https://doi.org:10.1371/journal.pone.0011487 Løvoll, M. et al. A novel totivirus and piscine reovirus (PRV) in Atlantic salmon ( Salmo salar ) with cardiomyopathy syndrome (CMS). Virol J 7 , 309 (2010). https://doi.org:10.1186/1743-422X-7-309 Hodneland, K. & Endresen, C. Sensitive and specific detection of Salmonid alphavirus using real-time PCR (TaqMan®). J Virol Methods 131 , 184-192 (2006). https://doi.org:10.1016/j.jviromet.2005.08.012 Ørpetveit, I. et al. Detection of infectious pancreatic necrosis virus in subclinically infected Atlantic salmon by virus isolation in cell culture or real-time reverse transcription polymerase chain reaction: influence of sample preservation and storage. Journal of Veterinary Diagnostic Investigation 22 , 886-895 (2010). https://doi.org:10.1177/10406387100220060 Riborg, A. et al. qPCR screening for Yersinia ruckeri clonal complex 1 against a background of putatively avirulent strains in Norwegian aquaculture. Journal of Fish Diseases 45 , 1211-1224 (2022). https://doi.org:10.1111/jfd.13656 Lazado, C. C., Breiland, M. W., Furtado, F., Burgerhout, E. & Strand, D. The circulating plasma metabolome of Neoparamoeba perurans -infected Atlantic salmon ( Salmo salar ). Microbial Pathogenesis 166 , 105553 (2022). https://doi.org:10.1016/j.micpath.2022.105553 NorwegianVeterinaryInstitute/Nepal: Release of the update Nepal pipeline v. v1.0.0 (Zenodo, 2024). Kruskal, W. H. & Wallis, W. A. Use of ranks in one-criterion variance analysis. Journal of the American Statistical Association 47 , 583-621 (1952). https://doi.org:10.1080/01621459.1952.10483441 Ewels, P. A. et al. The nf-core framework for community-curated bioinformatics pipelines. Nature Biotechnology 38 , 276-278 (2020). https://doi.org:10.1038/s41587-020-0439-x Wood, D. E. & Salzberg, S. L. Kraken: ultrafast metagenomic sequence classification using exact alignments. Genome Biol 15 , R46 (2014). https://doi.org:10.1186/gb-2014-15-3-r46 Team, R. C. R: A language and environment for statistical computing. R Foundation for Statistical Computing, Vienna, Austria. https://www.R-project.org/ (2016). Team, R. S. RStudio: integrated development for R. https://posit.co/download/rstudio-desktop/ (2024). Anderson, M. J. A new method for non‐parametric multivariate analysis of variance. Austral Ecology 26 , 32-46 (2001). https://doi.org:10.1111/j.1442-9993.2001.01070.pp.x Anderson, M. J. Distance-based tests for homogeneity of multivariate dispersions. Biometrics 62 , 245-253 (2006). https://doi.org:10.1111/j.1541-0420.2005.00440.x vegan: Community Ecology Package. R package version 2.8-0 (2025). Fernandes, A. D. et al. Unifying the analysis of high-throughput sequencing datasets: characterizing RNA-seq, 16S rRNA gene sequencing and selective growth experiments by compositional data analysis. Microbiome 2 , 15 (2014). https://doi.org:10.1186/2049-2618-2-15 Love, M. I., Huber, W. & Anders, S. Moderated estimation of fold change and dispersion for RNA-seq data with DESeq2. Genome Biol 15 , 550 (2014). https://doi.org:10.1186/s13059-014-0550-8 Millstein, J. et al. fdrci: FDR confidence interval selection and adjustment for large-scale hypothesis testing. Bioinformatics Advances 2 , vbac047 (2022). https://doi.org:10.1093/bioadv/vbac047 Additional Declarations No competing interests reported. Supplementary Files Supplementaries.docx TableS2.xlsx TableS3.xlsx TableS4.xlsx TableS5.xlsx TableS6.xlsx TableS7.xlsx Cite Share Download PDF Status: Published Journal Publication published 14 Apr, 2026 Read the published version in Scientific Reports → Version 1 posted Editorial decision: Revision requested 18 Mar, 2026 Reviews received at journal 08 Mar, 2026 Reviews received at journal 08 Mar, 2026 Reviewers agreed at journal 16 Feb, 2026 Reviews received at journal 15 Feb, 2026 Reviewers agreed at journal 15 Feb, 2026 Reviewers agreed at journal 13 Feb, 2026 Reviewers invited by journal 13 Feb, 2026 Submission checks completed at journal 12 Feb, 2026 First submitted to journal 12 Feb, 2026 You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-8373998","acceptedTermsAndConditions":true,"allowDirectSubmit":false,"archivedVersions":[],"articleType":"Article","associatedPublications":[],"authors":[{"id":592415761,"identity":"3d543097-6936-4a6e-be6a-b535f2e95198","order_by":0,"name":"Ottavia Benedicenti","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAABKElEQVRIie3RMWuDQBTA8ScPzHIla6TFfIUXhFgX+1UMt2awWyCFtAg3pXQ10A+Rj6AIZpG6Wjo0IWsGS6dQKT1TstSYucP9EfHEH+dTAJXq30aHA4AdztrDur5ufVrHv0TXgnotCZ4hcCSHO6D3ftenSbf/FH8y3wW78xivd75r2v1ATNmkurK7eYT+pEEMgXjJiIMzf+GDZ+KWIzTxxjJiTsgBw6w5eIpghIRAxXgobTRappp4XQhiVCDghThFcB/SDOh9Z39JMqvJ7eJbkjxpI3qvpETuwoYoiUeSwMe9JBE/SQzBh9clrZgzH1uGnGWwTEdBr0wt+WKcEtacpYvxtvCqqWl3skG5q9w+JatN6d2ZN5THmy1rfrFjLf86agUqlUqlOtcPrfpg7qZe7gkAAAAASUVORK5CYII=","orcid":"","institution":"Norwegian Veterinary Institute","correspondingAuthor":true,"prefix":"","firstName":"Ottavia","middleName":"","lastName":"Benedicenti","suffix":""},{"id":592415762,"identity":"6171a6a0-72bc-476a-8cd8-ce7c31ec3f8f","order_by":1,"name":"David A. Strand","email":"","orcid":"","institution":"Norwegian Veterinary Institute","correspondingAuthor":false,"prefix":"","firstName":"David","middleName":"A.","lastName":"Strand","suffix":""},{"id":592415763,"identity":"9c819b0e-dd8b-4fae-afd7-5f7868fded64","order_by":2,"name":"Saima Nasrin Mohammad","email":"","orcid":"","institution":"Norwegian Veterinary Institute","correspondingAuthor":false,"prefix":"","firstName":"Saima","middleName":"Nasrin","lastName":"Mohammad","suffix":""},{"id":592415764,"identity":"c152da9a-35a6-497f-b0c6-305a42f1fc19","order_by":3,"name":"Snorre Gulla","email":"","orcid":"","institution":"Norwegian Veterinary Institute","correspondingAuthor":false,"prefix":"","firstName":"Snorre","middleName":"","lastName":"Gulla","suffix":""},{"id":592415765,"identity":"9920ae3b-ddea-4a92-9af4-c1318f24ccbb","order_by":4,"name":"Marit Måsøy Amundsen","email":"","orcid":"","institution":"Norwegian Veterinary Institute","correspondingAuthor":false,"prefix":"","firstName":"Marit","middleName":"Måsøy","lastName":"Amundsen","suffix":""},{"id":592415766,"identity":"058aaf48-c747-4878-9788-a32f25610e71","order_by":5,"name":"Hilde Sindre","email":"","orcid":"","institution":"Norwegian Veterinary Institute","correspondingAuthor":false,"prefix":"","firstName":"Hilde","middleName":"","lastName":"Sindre","suffix":""},{"id":592415767,"identity":"e53c3723-b7d6-4705-8a72-a6063d535d7b","order_by":6,"name":"Trude Vrålstad","email":"","orcid":"","institution":"Norwegian Veterinary Institute","correspondingAuthor":false,"prefix":"","firstName":"Trude","middleName":"","lastName":"Vrålstad","suffix":""}],"badges":[],"createdAt":"2025-12-16 09:09:10","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-8373998/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-8373998/v1","draftVersion":[],"editorialEvents":[{"content":"https://doi.org/10.1038/s41598-026-48791-x","type":"published","date":"2026-04-14T15:57:21+00:00"}],"editorialNote":"","failedWorkflow":false,"files":[{"id":103335805,"identity":"b7e9f5b8-2af2-4dc5-9cc4-25c84cf7ab49","added_by":"auto","created_at":"2026-02-24 14:36:30","extension":"png","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":1147784,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eRelative abundance (%) of the top five genera per sample after nanofilt processing.\u003c/strong\u003e Samples are grouped by origin: hatchery, seawater sites, wellboat, and controls. Genera are colour-coded, and “other” represents taxa outside the top five for each sample.\u003c/p\u003e","description":"","filename":"Screenshot20260224at9.36.03AM.png","url":"https://assets-eu.researchsquare.com/files/rs-8373998/v1/e6bed83abb5c3e318871edc8.png"},{"id":103050421,"identity":"b3f930e6-d985-4848-a2fa-5ca613caca86","added_by":"auto","created_at":"2026-02-20 07:50:00","extension":"png","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":4972498,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eAlpha and beta diversity analyses and differentially abundant taxa at the genus level after excluding Homo sapiens, Archaea, and Eukaryota.\u003c/strong\u003e \u003cstrong\u003e(a)\u003c/strong\u003e Alpha diversity metrics (Pilou’s evenness, richness, Shannon diversity, and Simpson diversity) with a Wilcoxon significant test. \u003cstrong\u003e(b)\u003c/strong\u003e Beta diversity based on Bray–Curtis dissimilarity indicates clear separation between hatchery and seawater samples (PERMANOVA R² = 0.323, p = 0.001). \u003cstrong\u003e(c)\u003c/strong\u003e Differential abundance analysis of the top five genera using centred log-ratio (CLR) transformation highlights genera enriched in hatchery versus seawater sites. \u003cstrong\u003e(d)\u003c/strong\u003eVolcano plot of CLR effect sizes (Wilcoxon test) and FDR-adjusted p-values identifies taxa with significant differences (FDR ≤ 0.1 and |effect size| ≥ 0.5).\u003c/p\u003e","description":"","filename":"Fig.2.png","url":"https://assets-eu.researchsquare.com/files/rs-8373998/v1/e13f9cd200e9a10a2cac582c.png"},{"id":103050383,"identity":"e2f10d2d-2fe1-470f-a949-8fc8908cb0ea","added_by":"auto","created_at":"2026-02-20 07:49:44","extension":"png","order_by":3,"title":"Figure 3","display":"","copyAsset":false,"role":"figure","size":2199343,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eAlpha and beta diversity of hatchery water samples at the genus level after excluding Homo sapiens, Archaea, and Eukaryota.\u003c/strong\u003e \u003cstrong\u003e(a)\u003c/strong\u003eBoxplots show alpha diversity metrics (Pielou’s evenness, richness, Shannon, and Simpson) across three water groups: inlet, tank, and outlet water. No significant differences were observed among these groups for any metric. \u003cstrong\u003e(b)\u003c/strong\u003ePrincipal Coordinates Analysis (PCoA) based on Bray–Curtis dissimilarity illustrates beta diversity patterns among the same water groups. PERMANOVA results (R² = 0.138, p = 0.868) indicate no significant compositional differences between inlet, tank, and outlet water samples.\u003c/p\u003e","description":"","filename":"Fig.3.png","url":"https://assets-eu.researchsquare.com/files/rs-8373998/v1/c3a10954786576249f62e814.png"},{"id":103056439,"identity":"f08f962a-d27d-4972-8c2f-0a15290f162e","added_by":"auto","created_at":"2026-02-20 09:10:29","extension":"png","order_by":4,"title":"Figure 4","display":"","copyAsset":false,"role":"figure","size":2248907,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eAlpha and beta diversity of seawater samples at the genus level after excluding Homo sapiens, Archaea, and Eukaryota.\u003c/strong\u003e \u003cstrong\u003e(a)\u003c/strong\u003eBoxplots show alpha diversity metrics (Pielou’s evenness, richness, Shannon, and Simpson) across three sampling conditions: 1 km from the seawater cage at 1 m depth, 1 km from the cage at 20 m depth, and directly at the seawater cage. No significant differences were observed among these groups for any metric. \u003cstrong\u003e(b)\u003c/strong\u003ePrincipal Coordinates Analysis (PCoA) based on Bray–Curtis dissimilarity illustrates beta diversity patterns among the same water groups. PERMANOVA results (R² = 0.123, p = 0.463) indicate no significant compositional differences between sampling depths or proximity to the cage.\u003c/p\u003e","description":"","filename":"Fig.4.png","url":"https://assets-eu.researchsquare.com/files/rs-8373998/v1/5e951a3e5ea92740eecc9eb5.png"},{"id":103029153,"identity":"b27311e3-11b6-4c81-8232-7326da65d063","added_by":"auto","created_at":"2026-02-19 21:37:00","extension":"png","order_by":5,"title":"Figure 5","display":"","copyAsset":false,"role":"figure","size":3524829,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eAlpha and beta diversity and differentially abundant taxa in seawater cage samples grouped by PRV1 status (positive vs. negative) at the genus level after excluding Homo sapiens, Archaea, and Eukaryota.\u003c/strong\u003e \u003cstrong\u003e(a)\u003c/strong\u003eBoxplots show alpha diversity metrics (Pielou’s evenness, richness, Shannon, and Simpson) for PRV1-positive and PRV1-negative samples. \u003cstrong\u003e(b)\u003c/strong\u003e Principal Coordinates Analysis (PCoA) based on Bray–Curtis dissimilarity illustrates beta diversity patterns between PRV1 groups. PERMANOVA results (R² = 0.162, p = 0.14) indicate no significant compositional differences. \u003cstrong\u003e(c)\u003c/strong\u003eDifferential abundance analysis using CLR transformation, which highlights the top six genera with the largest effect sizes between PRV1-positive and PRV1-negative samples. \u003cstrong\u003e(d)\u003c/strong\u003e Volcano plot of CLR effect sizes (Wilcoxon test) and FDR-adjusted p-values shows no taxa meeting the significance threshold (FDR ≤ 0.1 and |effect size| ≥ 0.5).\u003c/p\u003e","description":"","filename":"Fig.5.png","url":"https://assets-eu.researchsquare.com/files/rs-8373998/v1/26bfaab7073491b6d6f7754c.png"},{"id":103029154,"identity":"db6a9d17-ca19-4a63-b4e8-2a37b5e5841e","added_by":"auto","created_at":"2026-02-19 21:37:00","extension":"png","order_by":6,"title":"Figure 6","display":"","copyAsset":false,"role":"figure","size":7967286,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eComparison of taxonomic classification using Kraken2 with standard versus custom databases at the species level.\u003c/strong\u003e \u003cstrong\u003e(a)\u003c/strong\u003e Relative abundance of species across samples from hatchery, seawater sites, and wellboat using the two database approaches. \u003cstrong\u003e(b)\u003c/strong\u003e Venn diagram showing species detected in hatchery and seawater site samples by the standard and custom databases. \u003cstrong\u003e(c)\u003c/strong\u003eVenn diagram showing species detected in wellboat samples by the standard and custom databases. Custom database improved detection of pathogenic species relevant to Atlantic salmon farms compared to the standard database.\u003c/p\u003e","description":"","filename":"Fig.6.png","url":"https://assets-eu.researchsquare.com/files/rs-8373998/v1/bf71b4501950740d5236bcb1.png"},{"id":103029155,"identity":"95aded1c-f883-4acf-86a2-92ccf00926bc","added_by":"auto","created_at":"2026-02-19 21:37:00","extension":"png","order_by":7,"title":"Figure 7","display":"","copyAsset":false,"role":"figure","size":10517597,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eOverview of sampling workflow.\u003c/strong\u003e Water samples were collected from two environments: hatchery (inlet, tank, and outlet water) and seawater sites (water inside the pen and locations 1 km/200m away at 1 m and 20 m depth). Samples were filtered and preserved in storage buffer before nucleic acid purification. Extracted nucleic acids were analysed by qPCR and sequenced using Oxford Nanopore Technologies (ONT) MinION platform. Created in BioRender. Benedicenti, O. (2026) https://BioRender.com/0djdjnv https://BioRender.com/x9uc0sc.\u003c/p\u003e","description":"","filename":"Fig.7.png","url":"https://assets-eu.researchsquare.com/files/rs-8373998/v1/ca8be0c13e2f69ccfe4ca804.png"},{"id":107352814,"identity":"41371443-3601-4fdd-8a9e-2cab29854327","added_by":"auto","created_at":"2026-04-20 16:16:11","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":30833821,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-8373998/v1/a02d9029-f59a-433f-82b0-4522f7854a1c.pdf"},{"id":103050520,"identity":"7b2e71a1-be4f-482d-91e7-87fcaca21809","added_by":"auto","created_at":"2026-02-20 07:50:22","extension":"docx","order_by":0,"title":"","display":"","copyAsset":false,"role":"supplement","size":903825,"visible":true,"origin":"","legend":"","description":"","filename":"Supplementaries.docx","url":"https://assets-eu.researchsquare.com/files/rs-8373998/v1/3524675b7347859ab47856aa.docx"},{"id":103029146,"identity":"9d5ecc38-440a-417f-aabd-d041045c5584","added_by":"auto","created_at":"2026-02-19 21:37:00","extension":"xlsx","order_by":1,"title":"","display":"","copyAsset":false,"role":"supplement","size":172172,"visible":true,"origin":"","legend":"","description":"","filename":"TableS2.xlsx","url":"https://assets-eu.researchsquare.com/files/rs-8373998/v1/8685ed482c3d3136459f7a07.xlsx"},{"id":103050503,"identity":"60a9b3ce-f6ed-42ed-8017-da3cb9ef7cf2","added_by":"auto","created_at":"2026-02-20 07:50:18","extension":"xlsx","order_by":2,"title":"","display":"","copyAsset":false,"role":"supplement","size":15525,"visible":true,"origin":"","legend":"","description":"","filename":"TableS3.xlsx","url":"https://assets-eu.researchsquare.com/files/rs-8373998/v1/fafc83af2d02bcf5b27190e6.xlsx"},{"id":103050466,"identity":"b8afa60d-692c-412c-b880-4f4b71d190e5","added_by":"auto","created_at":"2026-02-20 07:50:09","extension":"xlsx","order_by":3,"title":"","display":"","copyAsset":false,"role":"supplement","size":14981,"visible":true,"origin":"","legend":"","description":"","filename":"TableS4.xlsx","url":"https://assets-eu.researchsquare.com/files/rs-8373998/v1/6fcbe0ab3499725451fcbae5.xlsx"},{"id":103029150,"identity":"85419bf4-bb7d-4407-80fc-528371a13300","added_by":"auto","created_at":"2026-02-19 21:37:00","extension":"xlsx","order_by":4,"title":"","display":"","copyAsset":false,"role":"supplement","size":13330,"visible":true,"origin":"","legend":"","description":"","filename":"TableS5.xlsx","url":"https://assets-eu.researchsquare.com/files/rs-8373998/v1/a7a1ead6f6bf7acd42a25fdd.xlsx"},{"id":103050461,"identity":"1e63ca9f-55bd-47e6-8bca-60130edc79d2","added_by":"auto","created_at":"2026-02-20 07:50:08","extension":"xlsx","order_by":5,"title":"","display":"","copyAsset":false,"role":"supplement","size":18534,"visible":true,"origin":"","legend":"","description":"","filename":"TableS6.xlsx","url":"https://assets-eu.researchsquare.com/files/rs-8373998/v1/672a8d3e917ac0b0b4e2934f.xlsx"},{"id":103056581,"identity":"25c06abe-f1db-4e7c-aebd-caca0907fd06","added_by":"auto","created_at":"2026-02-20 09:16:52","extension":"xlsx","order_by":6,"title":"","display":"","copyAsset":false,"role":"supplement","size":12803,"visible":true,"origin":"","legend":"","description":"","filename":"TableS7.xlsx","url":"https://assets-eu.researchsquare.com/files/rs-8373998/v1/c1aef19a099d52e2971f9db5.xlsx"}],"financialInterests":"No competing interests reported.","formattedTitle":"Integrated approaches for pathogen monitoring and shotgun metagenomic analysis in Atlantic salmon farming","fulltext":[{"header":"Introduction","content":"\u003cp\u003eThe production of Atlantic salmon, \u003cem\u003eSalmo salar\u003c/em\u003e (Linnaeus), was around 2.02\u0026nbsp;million tonnes during January\u0026ndash;September 2024, a slight 1% decline compared to 2023, and the main producer, Norway, accounted for ca. 1.5\u0026nbsp;million tonnes in the same period (\u0026ndash;1.3% year-on-year)\u003csup\u003e1\u003c/sup\u003e. However, infectious diseases continue to pose major challenges for the salmon farming industry, affecting fish health, welfare, and productivity\u003csup\u003e\u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2\u003c/span\u003e\u003c/sup\u003e. Therefore, specific and sensitive tools for pathogen detection are crucial for the development of efficient preventive/ mitigative strategies and to limit the spread of pathogens in fish farms.\u003c/p\u003e \u003cp\u003eIn recent years, environmental DNA/RNA (eDNA/eRNA) sampling has become an increasingly adopted approach for pathogen surveillance in aquaculture\u003csup\u003e\u003cspan additionalcitationids=\"CR4 CR5 CR6 CR7 CR8\" citationid=\"CR3\" class=\"CitationRef\"\u003e3\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR9\" class=\"CitationRef\"\u003e9\u003c/span\u003e\u003c/sup\u003e. Moreover, a good screening method could not only be used to detect pathogens shed from infected fish but also to find pathogens circulating within the farm environment. This could allow for early implementation of preventive measures in response to elevated pathogen levels and in advance of clinical symptoms emerging. For this reason, monitoring fish pathogens in ambient farm water (tanks or cages) offers a less invasive and more scalable alternative to direct fish sampling and may potentially provide earlier insights into pathogen presence during infection. This approach also aligns with a 3R (Replacement, Reduction, Refinement) strategy\u003csup\u003e\u003cspan citationid=\"CR10\" class=\"CitationRef\"\u003e10\u003c/span\u003e\u003c/sup\u003e and supports the One Health framework by promoting integrated biosecurity monitoring that benefits aquatic animal welfare, environmental sustainability, and human health\u003csup\u003e\u003cspan citationid=\"CR11\" class=\"CitationRef\"\u003e11\u003c/span\u003e\u003c/sup\u003e.\u003c/p\u003e \u003cp\u003eThere are recent studies that have explored metagenomics and metabarcoding in freshwater aquaculture systems, particularly focusing on microbial communities in Recirculating Aquaculture Systems (RAS), as they rely heavily on microbial communities (e.g., biofilters, water, and biofilms) for nutrient cycling, water quality and fish health\u003csup\u003e\u003cspan additionalcitationids=\"CR13\" citationid=\"CR12\" class=\"CitationRef\"\u003e12\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e14\u003c/span\u003e\u003c/sup\u003e. A recent study aimed to compare how different sequencing approaches affect the characterisation of microbial communities in RAS\u003csup\u003e\u003cspan citationid=\"CR12\" class=\"CitationRef\"\u003e12\u003c/span\u003e\u003c/sup\u003e. Here, while 16S rRNA gene sequencing was reliable for identifying community structure and spatiotemporal patterns, long-read sequencing was less effective for quantitative pattern analysis but useful for identifying functional roles and pathogens, and shotgun metagenomics provided broader insights, including detection of fungi, viruses, and bacteriophages, enabling exploration of inter-domain interactions\u003csup\u003e\u003cspan citationid=\"CR12\" class=\"CitationRef\"\u003e12\u003c/span\u003e\u003c/sup\u003e. Oxford Nanopore Technology was also used to analyse microbial communities and detect bacterial pathogens in an Atlantic salmon commercial freshwater RAS performing long-read 16S rRNA gene sequencing\u003csup\u003e\u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e14\u003c/span\u003e\u003c/sup\u003e. This technology enables real-time, high-resolution monitoring of microbial communities in water systems\u003csup\u003e\u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e14\u003c/span\u003e,\u003cspan citationid=\"CR15\" class=\"CitationRef\"\u003e15\u003c/span\u003e\u003c/sup\u003e. However, the choice of sequencing methods and reference databases might constitute a bottleneck in some studies. For instance, the diversity of RNA viruses in Lake Needwood (USA), explored using metagenomic sequencing, did not match any known viruses, indicating a large pool of novel viral diversity\u003csup\u003e\u003cspan citationid=\"CR16\" class=\"CitationRef\"\u003e16\u003c/span\u003e\u003c/sup\u003e. Another study reviewed and emphasised the potential role of metagenomics, coupled with bioinformatics tools and databases, as a non-invasive, rapid tool for environmental monitoring and disease prevention\u003csup\u003e\u003cspan citationid=\"CR13\" class=\"CitationRef\"\u003e13\u003c/span\u003e\u003c/sup\u003e. Although metagenomics is increasingly used in recirculating aquaculture systems (RAS), its application in tanks and cage-based systems remains limited, despite evidence showing its effectiveness in detecting pathogens and monitoring environmental impacts such as eutrophication\u003csup\u003e\u003cspan citationid=\"CR17\" class=\"CitationRef\"\u003e17\u003c/span\u003e\u003c/sup\u003e. Because each sequencing method has inherent strengths and weaknesses, microbial community studies must clearly document methodological choices (e.g., filtration, extraction protocols, sequencing depth), as these significantly affect inter-study comparability. In this context, a combined qPCR\u0026ndash;shotgun metagenomics approach may support early mitigation strategies in aquaculture by improving pathogen detection and providing ecological context relevant to fish health. For this reason, this work explored an innovative, non-invasive approach to monitor the presence of pathogens in aquaculture environments using a previously established method for filtering, storing, and extracting eDNA/eRNA from water samples\u003csup\u003e\u003cspan citationid=\"CR18\" class=\"CitationRef\"\u003e18\u003c/span\u003e\u003c/sup\u003e. In addition to targeted detection, shotgun metagenomics analysis was employed to characterise broader microbial community dynamics, providing insights into potential environmental shifts associated with pathogen presence or other indicators of disturbance (e.g., mucus shedding, stress, and pathogen occurrence).\u003c/p\u003e \u003cp\u003eIn this study, we used a previously published procedure for detecting various targeted pathogens\u003csup\u003e\u003cspan citationid=\"CR18\" class=\"CitationRef\"\u003e18\u003c/span\u003e\u003c/sup\u003e in both hatchery and seawater sites, where we followed the same Atlantic salmon population. Pathogen surveillance during stressful operational events, such as wellboat delousing, revealed in our study a marked increase in pathogen concentration post-treatment in the water. Moreover, we explored the feasibility of on-site microbial screening testing portable sequencing technologies, such as the Oxford Nanopore MinION, which could support rapid field assessments. In fact, preliminary microbiome analyses, based on metagenomic profiling, offered insights into microbial community dynamics. These included signs of decreased richness and shifts in the virome, which may reflect environmental changes potentially associated with the presence of a salmon pathogen in the water. Moreover, custom pathogen databases were also used to validate outputs from publicly available resources, highlighting the importance of curated reference data in improving the accuracy and relevance of metagenomic pathogen surveillance in aquaculture. The overarching aim of this study was therefore to evaluate the feasibility and performance of a practical, field-deployable workflow for pathogen surveillance in aquaculture, combining onsite filtration with targeted qPCR detection and complementary metagenomic profiling. While the central focus is methodological, we also assessed whether shotgun metagenomics provided additional ecological context \u0026mdash; such as microbial community shifts \u0026mdash; relevant to pathogen occurrence. This dual focus allowed us to determine both the practical usability of the workflow under real farm conditions and its potential to support broader health-monitoring frameworks.\u003c/p\u003e"},{"header":"Results","content":"\u003cdiv id=\"Sec3\" class=\"Section2\"\u003e \u003ch2\u003e2.1 qPCR / RT-qPCR assays\u003c/h2\u003e \u003cp\u003eThe same Atlantic salmon population was monitored from the hatchery stage (freshwater and smoltification process) through transfer to the seawater site. qPCR screening across all sampling points detected two targeted pathogens: infectious salmon anaemia virus (ISAV) in hatchery inlet-, tank-, and outlet waters, and piscine orthoreovirus genotype 1 (PRV1) in seawater cage samples from August 2023 onwards (Table\u0026nbsp;\u003cspan refid=\"Tab1\" class=\"InternalRef\"\u003e1\u003c/span\u003e \u0026amp; Table \u003cspan refid=\"MOESM2\" class=\"InternalRef\"\u003eS2\u003c/span\u003e). No pathogens were detected in offshore samples (1 km). PRV1 was consistently detected only inside the pens, except during the final sampling event, when low-level PRV1 signals were also observed 200 m from the cage. Additional sampling was performed before and after delousing treatments on wellboats. PRV1 concentrations increased following treatment, likely reflecting enhanced viral shedding associated with fish stress during handling procedures (Table\u0026nbsp;\u003cspan refid=\"Tab1\" class=\"InternalRef\"\u003e1\u003c/span\u003e \u0026amp; Table \u003cspan refid=\"MOESM2\" class=\"InternalRef\"\u003eS2\u003c/span\u003e).\u003c/p\u003e \u003cp\u003e \u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab1\" border=\"1\"\u003e \u003ccaption language=\"En\"\u003e \u003cdiv class=\"CaptionNumber\"\u003eTable 1\u003c/div\u003e \u003cdiv class=\"CaptionContent\"\u003e \u003cp\u003e\u003cb\u003eCq values and RT-qPCR assays estimated copy numbers per L for ISAV (hatchery) and PRV1 (seawater sites and wellboat).\u003c/b\u003e Results are shown as mean\u0026thinsp;\u0026plusmn;\u0026thinsp;standard deviation from three pooled biological replicates and two technical replicates.\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e \u003ccolgroup cols=\"4\"\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\"\u003e \u003cp\u003eSamples\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c2\"\u003e \u003cp\u003eGroup\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c3\"\u003e \u003cp\u003eCq Mean\u0026thinsp;\u0026plusmn;\u0026thinsp;St. Dev.\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c4\"\u003e \u003cp\u003eCopy numbers/ L\u0026thinsp;\u0026plusmn;\u0026thinsp;St. Dev.\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eOutlet water (Feb/23)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eHatchery\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e35.39\u0026thinsp;\u0026plusmn;\u0026thinsp;0.37\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e3952.00\u0026thinsp;\u0026plusmn;\u0026thinsp;1012.58\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eTank water (Feb/23)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eHatchery\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e35.42\u0026thinsp;\u0026plusmn;\u0026thinsp;0.09\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e3824.20\u0026thinsp;\u0026plusmn;\u0026thinsp;254.28\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eInlet water (Feb/23)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eHatchery\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e34.68\u0026thinsp;\u0026plusmn;\u0026thinsp;0.23\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e6414.00\u0026thinsp;\u0026plusmn;\u0026thinsp;1009.75\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eControl (Feb/23)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eHatchery\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eN/A\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eN/A\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eOutlet water (May/23)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eHatchery\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e34.30\u0026thinsp;\u0026plusmn;\u0026thinsp;0.08\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e8370.00\u0026thinsp;\u0026plusmn;\u0026thinsp;444.06\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eTank water (May/23)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eHatchery\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e34.25\u0026thinsp;\u0026plusmn;\u0026thinsp;0.30\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e8700.00\u0026thinsp;\u0026plusmn;\u0026thinsp;1753.62\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eInlet water (May/23)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eHatchery\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e34.38\u0026thinsp;\u0026plusmn;\u0026thinsp;0.12\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e7884.00\u0026thinsp;\u0026plusmn;\u0026thinsp;661.85\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eControl (May/23)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eHatchery\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eN/A\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eN/A\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e1km from cage \u0026minus;\u0026thinsp;1m depth (Apr/23)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eSeawater sites\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eN/A\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eN/A\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e1km from cage \u0026minus;\u0026thinsp;20m depth (Apr/23)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eSeawater sites\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eN/A\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eN/A\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eSeawater cage (Apr/23)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eSeawater sites\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eN/A\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eN/A\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eControl (Apr/23)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eSeawater sites\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eN/A\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eN/A\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e1km from cage \u0026minus;\u0026thinsp;1m depth (May/23)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eSeawater sites\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eN/A\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eN/A\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e1km from cage \u0026minus;\u0026thinsp;20m depth (May/23)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eSeawater sites\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eN/A\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eN/A\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eSeawater cage (May/23)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eSeawater sites\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eN/A\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eN/A\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eControl (May/23)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eSeawater sites\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eN/A\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eN/A\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eSeawater cage (Jun/23)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eSeawater sites\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eN/A\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eN/A\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eSeawater cage (Jul/23)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eSeawater sites\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eN/A\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eN/A\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eSeawater cage (Aug/23)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eSeawater sites\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e38.00\u0026thinsp;\u0026plusmn;\u0026thinsp;0.66\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e209.94\u0026thinsp;\u0026plusmn;\u0026thinsp;91.78\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eControl (Aug/23)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eSeawater sites\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eN/A\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eN/A\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eSeawater cage (Sep/23)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eSeawater sites\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e31.68\u0026thinsp;\u0026plusmn;\u0026thinsp;0.06\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e14826.00\u0026thinsp;\u0026plusmn;\u0026thinsp;557.20\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e1km from cage \u0026minus;\u0026thinsp;1m depth (Oct/23)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eSeawater sites\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eN/A\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eN/A\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e1km from cage \u0026minus;\u0026thinsp;20m depth (Oct/23)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eSeawater sites\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eN/A\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eN/A\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eSeawater cage (Oct/23)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eSeawater sites\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e34.12\u0026thinsp;\u0026plusmn;\u0026thinsp;0.08\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e2800.80\u0026thinsp;\u0026plusmn;\u0026thinsp;162.92\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eControl (Oct/23)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eSeawater sites\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eN/A\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eN/A\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eWellboat before treatment (Nov/23)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eWellboat\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e36.81\u0026thinsp;\u0026plusmn;\u0026thinsp;0.77\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e481.78\u0026thinsp;\u0026plusmn;\u0026thinsp;242.99\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eWellboat after treatment (Nov/23)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eWellboat\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e29.93\u0026thinsp;\u0026plusmn;\u0026thinsp;0.07\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e48760.00\u0026thinsp;\u0026plusmn;\u0026thinsp;2319.31\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eSeawater cage (Dec/23)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eSeawater sites\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e32.15\u0026thinsp;\u0026plusmn;\u0026thinsp;0.02\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e10770.00\u0026thinsp;\u0026plusmn;\u0026thinsp;132.94\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e1km from cage \u0026minus;\u0026thinsp;1m depth (Feb/24)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eSeawater sites\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eN/A\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eN/A\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e1km from cage \u0026minus;\u0026thinsp;20m depth (Feb/24)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eSeawater sites\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eN/A\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eN/A\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eSeawater cage (Feb/24)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eSeawater sites\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e39.20\u0026thinsp;\u0026plusmn;\u0026thinsp;1.46\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e111.30\u0026thinsp;\u0026plusmn;\u0026thinsp;95.69\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eControl (Feb/24)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eSeawater sites\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eN/A\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eN/A\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e200m from cage \u0026minus;\u0026thinsp;1m depth (Feb/24)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eSeawater sites\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e40.09\u0026thinsp;\u0026plusmn;\u0026thinsp;0.01\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e47.80\u0026thinsp;\u0026plusmn;\u0026thinsp;0.45\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e200m from cage \u0026minus;\u0026thinsp;20m depth (Feb/24)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eSeawater sites\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eN/A\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eN/A\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003c/tbody\u003e \u003c/colgroup\u003e \u003c/table\u003e\u003c/div\u003e \u003c/p\u003e \u003cp\u003eAlthough some detections occurred at high Cq values (\u0026gt;\u0026thinsp;35), these signals fell within the expected behaviour of low-concentration environmental samples approaching the assay\u0026rsquo;s limits of detection (LOD). Importantly, low-positive results were reproducible across biological (n\u0026thinsp;=\u0026thinsp;3) and technical (n\u0026thinsp;=\u0026thinsp;2) replicates and supported by consistent standard curve\u0026ndash;based quantification (Table \u003cspan refid=\"MOESM2\" class=\"InternalRef\"\u003eS2\u003c/span\u003e), indicating that these detections reflect true low-level environmental shedding rather than analytical noise.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec4\" class=\"Section2\"\u003e \u003ch2\u003e2.2 Sequencing pipeline output and quality assessment\u003c/h2\u003e \u003cp\u003eFrom the output of the Oxford Nanopore raw ligation sequencing data processed using the NEPAL pipeline, we evaluated three processing strategies\u0026mdash;simplex, duplex, and nanofilt-filtered reads\u0026mdash;to assess differences in read quality and downstream performance. Five key metrics were analysed for each run (runs 1\u0026ndash;4) using boxplots (Fig. \u003cspan refid=\"MOESM1\" class=\"InternalRef\"\u003eS1\u003c/span\u003e \u0026amp; Table \u003cspan refid=\"MOESM3\" class=\"InternalRef\"\u003eS3\u003c/span\u003e): 1) average read length, which varied between runs, with duplex reads generally shorter due to trimming during consensus generation; 2) average quality scores (range: Q15\u0026ndash;Q21), which differed significantly among workflows, with duplex reads exhibited the highest average quality, while nanofilt results depended on the filtering threshold; 3) GC content (%), which remained relatively stable across workflows and runs (35\u0026ndash;50%, as expected for environmental samples); 4) N50 (bp) values used as an indicator of assembly performance, with values exceeding 2,000 bp; and 5) Q30 (%), where duplex reads showed lower proportions of high-confidence bases compared to nanofilt-filtered reads. Based on these results, we proceeded with nanofilt-filtered reads for taxonomic profiling using the taxprofiler pipeline with the Kraken2 database. Although filtering may remove some informative reads, Kraken2 benefits from longer reads and higher Q30 values; therefore, nanofilt-filtered reads were considered more robust for taxonomic classification, offering improved sensitivity and specificity. The taxprofiler output was assessed using standard quality metrics (Fig. \u003cspan refid=\"MOESM2\" class=\"InternalRef\"\u003eS2\u003c/span\u003e \u0026amp; Table \u003cspan refid=\"MOESM4\" class=\"InternalRef\"\u003eS4\u003c/span\u003e). As expected, control samples exhibited a lower number of reads compared to experimental samples (Fig. \u003cspan refid=\"MOESM2\" class=\"InternalRef\"\u003eS2\u003c/span\u003e \u0026amp; Table \u003cspan refid=\"MOESM4\" class=\"InternalRef\"\u003eS4\u003c/span\u003e). Most processed reads had a median length between 1,749 and 2,749 bp, which is appropriate for long-read sequencing, especially for environmental or low-input samples (Fig. \u003cspan refid=\"MOESM2\" class=\"InternalRef\"\u003eS2\u003c/span\u003e \u0026amp; Table \u003cspan refid=\"MOESM4\" class=\"InternalRef\"\u003eS4\u003c/span\u003e). GC content was consistent across all samples (\u0026asymp;\u0026thinsp;40\u0026ndash;50%), supporting the overall uniformity of library preparation (Fig. \u003cspan refid=\"MOESM2\" class=\"InternalRef\"\u003eS2\u003c/span\u003e \u0026amp; Table \u003cspan refid=\"MOESM4\" class=\"InternalRef\"\u003eS4\u003c/span\u003e). However, \u003cem\u003eHomo sapiens\u003c/em\u003e sequences were detected in several samples, suggesting contamination during sampling rather than during library prep, despite the use of sterile materials (e.g., gloves, sterile filter cups, etc.) (Fig. \u003cspan refid=\"MOESM2\" class=\"InternalRef\"\u003eS2\u003c/span\u003e \u0026amp; Table \u003cspan refid=\"MOESM4\" class=\"InternalRef\"\u003eS4\u003c/span\u003e).\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec5\" class=\"Section2\"\u003e \u003ch2\u003e2.3 Taxonomic classification\u003c/h2\u003e \u003cp\u003eWe first analysed the overall relative abundance (%) among the top six classified phyla, excluding the \u003cem\u003eS. salar\u003c/em\u003e genome (taxprofiler pipeline) and \u003cem\u003eH. sapiens\u003c/em\u003e contamination. Samples were grouped by environment (hatchery, seawater sites, and wellboat) and controls (Fig. \u003cspan refid=\"MOESM3\" class=\"InternalRef\"\u003eS3\u003c/span\u003e). Hatchery samples exhibited the lowest proportion of unclassified reads (12.82\u0026thinsp;\u0026plusmn;\u0026thinsp;6.37%) and minimal variability in \u0026ldquo;other\u0026rdquo; taxa (1.88\u0026thinsp;\u0026plusmn;\u0026thinsp;0.07%) (Fig. \u003cspan refid=\"MOESM3\" class=\"InternalRef\"\u003eS3\u003c/span\u003e \u0026amp; Table \u003cspan refid=\"MOESM5\" class=\"InternalRef\"\u003eS5\u003c/span\u003e). In contrast, seawater sites showed higher unclassified percentages (23.99\u0026thinsp;\u0026plusmn;\u0026thinsp;4.71%) and greater variability in \u0026ldquo;other\u0026rdquo; taxa (5.03\u0026thinsp;\u0026plusmn;\u0026thinsp;1.25%) (Fig. \u003cspan refid=\"MOESM3\" class=\"InternalRef\"\u003eS3\u003c/span\u003e \u0026amp; Table \u003cspan refid=\"MOESM5\" class=\"InternalRef\"\u003eS5\u003c/span\u003e). The wellboat group displayed the greatest variability in taxonomic classification, driven by the second sample collected after treatment, which had an exceptionally high proportion of unclassified reads (43.5%) (Fig. \u003cspan refid=\"MOESM3\" class=\"InternalRef\"\u003eS3\u003c/span\u003e \u0026amp; Table \u003cspan refid=\"MOESM5\" class=\"InternalRef\"\u003eS5\u003c/span\u003e). Transient increases in the proportion of unclassified reads are common in environmental metagenomics, particularly when community composition shifts toward taxa that are poorly represented in reference databases or that require \u003cem\u003ede novo\u003c/em\u003e assembly for accurate annotation\u003csup\u003e\u003cspan citationid=\"CR19\" class=\"CitationRef\"\u003e19\u003c/span\u003e\u003c/sup\u003e. This suggests that treatment may have altered the microbial community or introduced sequences that were more difficult to classify. The Zymo microbial standard positive control was almost fully classified (99.99%), with negligible unclassified or \u0026ldquo;other\u0026rdquo; taxa (0.01%) (Fig. \u003cspan refid=\"MOESM3\" class=\"InternalRef\"\u003eS3\u003c/span\u003e \u0026amp; Table \u003cspan refid=\"MOESM5\" class=\"InternalRef\"\u003eS5\u003c/span\u003e), as expected for a well-characterised mock community. Similarly, the positive control from the spiked pilot experiment\u003csup\u003e\u003cspan citationid=\"CR18\" class=\"CitationRef\"\u003e18\u003c/span\u003e\u003c/sup\u003e was also nearly fully classified (97.56%), with only a small fraction unclassified (1.53%) (Fig. \u003cspan refid=\"MOESM3\" class=\"InternalRef\"\u003eS3\u003c/span\u003e \u0026amp; Table \u003cspan refid=\"MOESM5\" class=\"InternalRef\"\u003eS5\u003c/span\u003e). These results confirm that the pipeline performed well on known or control samples.\u003c/p\u003e \u003cp\u003eA general taxonomic composition profile showed a shift from \u003cem\u003eActinomycetota\u003c/em\u003e (phylum of Gram-positive bacteria with high GC content) and \u003cem\u003eBacteroidota\u003c/em\u003e (Gram-negative bacteria) dominance in the hatchery, to more diverse communities in seawater sites reflecting environmental changes (Fig. \u003cspan refid=\"MOESM3\" class=\"InternalRef\"\u003eS3\u003c/span\u003e). Hatchery water is nutrient-rich and controlled, favouring the enrichment of heterotrophic degraders typical of recirculating systems\u003csup\u003e\u003cspan citationid=\"CR20\" class=\"CitationRef\"\u003e20\u003c/span\u003e,\u003cspan citationid=\"CR21\" class=\"CitationRef\"\u003e21\u003c/span\u003e\u003c/sup\u003e, while seawater sites showed: \u003cem\u003eChlorophyta\u003c/em\u003e peaks during spring/algal bloom periods, indicating seasonal eutrophication and/or phytoplankton proliferation; \u003cem\u003eNitrososphaerota\u003c/em\u003e presence, which suggest active nitrification, likely linked to nitrogen cycling in the marine environment\u003csup\u003e\u003cspan citationid=\"CR22\" class=\"CitationRef\"\u003e22\u003c/span\u003e\u003c/sup\u003e; and \u003cem\u003eUroviricota\u003c/em\u003e and \u003cem\u003eNucleocytoviricota\u003c/em\u003e, which could be correlated to viral regulation of microbial and algal populations, influencing bloom dynamics and nutrient turnover\u003csup\u003e\u003cspan citationid=\"CR23\" class=\"CitationRef\"\u003e23\u003c/span\u003e,\u003cspan citationid=\"CR24\" class=\"CitationRef\"\u003e24\u003c/span\u003e\u003c/sup\u003e (Fig. \u003cspan refid=\"MOESM3\" class=\"InternalRef\"\u003eS3\u003c/span\u003e). For subsequent comparisons across environments, Archaea and Eukaryota were excluded from the analyses because their high abundance in marine sites, often driven by seasonal factors such as algal blooms, could mask bacterial and viral taxa. It was thus considered that focusing on bacteria, viruses, and bacteriophages would provide a more informative perspective on microbial dynamics throughout the salmon production cycle, from hatchery to seawater sites and ultimately slaughter. Therefore, we showed the relative abundance (%) of the top six genera per sample after nanofilt processing, including \u0026ldquo;others\u0026rdquo; (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003e \u0026amp; Table \u003cspan refid=\"MOESM6\" class=\"InternalRef\"\u003eS6\u003c/span\u003e). Hatchery samples showed the highest percentage for \u0026ldquo;other\u0026rdquo; (ca. 22.08%), but relatively low unclassified reads, while seawater sites had higher unclassified reads (ca. 23.99%), reflecting marine complexity with taxonomic gaps (Table \u003cspan refid=\"MOESM7\" class=\"InternalRef\"\u003eS7\u003c/span\u003e). The extreme unclassified proportion (43.50%) was driven by the wellboat post-treatment sample, while controls behaved as expected with low unclassified reads except for minor reagent contamination (negative control) (Table \u003cspan refid=\"MOESM7\" class=\"InternalRef\"\u003eS7\u003c/span\u003e). Hatchery water microbiota is dominated by \u003cem\u003eMycolicibacterium\u003c/em\u003e (\u003cem\u003eActinomycetota\u003c/em\u003e), \u003cem\u003eMycobacterium\u003c/em\u003e (\u003cem\u003eActinomycetota\u003c/em\u003e) and \u003cem\u003eFlavobacterium\u003c/em\u003e (\u003cem\u003eBacteroidota\u003c/em\u003e) (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003e \u0026amp; Table \u003cspan refid=\"MOESM6\" class=\"InternalRef\"\u003eS6\u003c/span\u003e). In contrast, seawater sites exhibited seasonal shifts that strongly influenced microbial composition: spring blooms (April\u0026ndash;May) favoured polysaccharide degraders (\u003cem\u003ePolaribacter\u003c/em\u003e\u003csup\u003e\u003cspan citationid=\"CR25\" class=\"CitationRef\"\u003e25\u003c/span\u003e\u003c/sup\u003e), along with oligotrophic specialists such as \u0026ldquo;\u003cem\u003eCandidatus\u003c/em\u003e Pelagibacter\u0026rdquo; (SAR11\u003csup\u003e26\u003c/sup\u003e), summer (July\u0026ndash;August) was characterised by viral\u0026ndash;algal interactions (\u003cem\u003ePrasinovirus\u003c/em\u003e\u003csup\u003e\u003cspan citationid=\"CR27\" class=\"CitationRef\"\u003e27\u003c/span\u003e\u003c/sup\u003e), autumn (September\u0026ndash;October) introduced potential opportunistic pathogens (\u003cem\u003eAliivibrio\u003c/em\u003e\u003csup\u003e\u003cspan citationid=\"CR28\" class=\"CitationRef\"\u003e28\u003c/span\u003e\u003c/sup\u003e, \u003cem\u003ePseudalteromonas\u003c/em\u003e\u003csup\u003e\u003cspan citationid=\"CR29\" class=\"CitationRef\"\u003e29\u003c/span\u003e\u003c/sup\u003e), and winter (December\u0026ndash;February) was dominated by \u0026ldquo;\u003cem\u003eCandidatus\u003c/em\u003e Pelagibacter\u0026rdquo; (SAR11), together with cold-adapted genera like \u003cem\u003eColwellia\u003c/em\u003e\u003csup\u003e\u003cspan citationid=\"CR30\" class=\"CitationRef\"\u003e30\u003c/span\u003e\u003c/sup\u003e and bloom-termination-associated genera such as \u003cem\u003eKordia\u003c/em\u003e\u003csup\u003e\u003cspan citationid=\"CR31\" class=\"CitationRef\"\u003e31\u003c/span\u003e\u003c/sup\u003e. The negative control showed a high proportion of \u0026ldquo;other\u0026rdquo; genera (59.94%), likely reflecting low-level contamination commonly reported in extraction kits and reagents. Therefore, potential contaminant taxa should be considered when interpreting low-biomass samples, particularly in the taxonomic composition of wellboat samples after treatment. The pilot experiment\u003csup\u003e\u003cspan citationid=\"CR18\" class=\"CitationRef\"\u003e18\u003c/span\u003e\u003c/sup\u003e was dominated by \u003cem\u003eYersinia\u003c/em\u003e (99.99%), confirming successful sequencing and pipeline accuracy (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003e \u0026amp; Table \u003cspan refid=\"MOESM6\" class=\"InternalRef\"\u003eS6\u003c/span\u003e). Moreover, the Zymo microbial standard positive control\u003csup\u003e\u003cspan citationid=\"CR32\" class=\"CitationRef\"\u003e32\u003c/span\u003e\u003c/sup\u003e was characterised by \u003cem\u003eStaphylococcus\u003c/em\u003e, \u003cem\u003eBacillus\u003c/em\u003e, \u003cem\u003eListeria\u003c/em\u003e, and \u003cem\u003eEnterococcus\u003c/em\u003e in balanced proportions (12\u0026ndash;18%), further validating pipeline accuracy (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003e \u0026amp; Table \u003cspan refid=\"MOESM6\" class=\"InternalRef\"\u003eS6\u003c/span\u003e).\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec6\" class=\"Section2\"\u003e \u003ch2\u003e2.3 Alpha and beta diversities\u003c/h2\u003e \u003cp\u003eAlpha and beta diversity analyses between hatchery and seawater sites were not significantly different at the phylum or species level; therefore, subsequent analyses focused on the genus level, as this was considered the lowest phylogenetic rank providing reliable information based on the selected standard Kraken2 database. Alpha diversity showed significant differences in Pielou\u0026rsquo;s evenness, Shannon, and Simpson indices, while richness did not differ significantly between environments (Wilcoxon test; Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003ea). Beta diversity based on Bray\u0026ndash;Curtis dissimilarity\u003csup\u003e\u003cspan citationid=\"CR33\" class=\"CitationRef\"\u003e33\u003c/span\u003e\u003c/sup\u003e revealed a clear separation between hatchery and seawater samples (PERMANOVA, R\u0026sup2; = 0.323, p\u0026thinsp;=\u0026thinsp;0.001; Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003eb). Differential abundance analysis of the six most abundant genera using centred log-ratio (CLR) transformation highlighted taxa enriched in hatchery versus seawater sites: hatchery showed higher relative abundance in \u003cem\u003eAurantimicrobium\u003c/em\u003e and \u003cem\u003eMycolicibacterium\u003c/em\u003e, while marine sites in \u0026ldquo;\u003cem\u003eCandidatus\u003c/em\u003e Pseudothioglobus\u0026rdquo;, \u003cem\u003eLyrvirus\u003c/em\u003e, \u003cem\u003ePelagivirus\u003c/em\u003e, and \u003cem\u003eSiovirus\u003c/em\u003e (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003ec). Volcano plots of CLR effect sizes (Wilcoxon test\u003csup\u003e\u003cspan citationid=\"CR34\" class=\"CitationRef\"\u003e34\u003c/span\u003e\u003c/sup\u003e) and false discovery rate (FDR)-adjusted p-values\u003csup\u003e\u003cspan citationid=\"CR35\" class=\"CitationRef\"\u003e35\u003c/span\u003e\u003c/sup\u003e indicated a consistent set of genera significantly different between environments (FDR\u0026thinsp;\u0026le;\u0026thinsp;0.1 and |effect size| \u0026ge; 0.5; Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003ed). Effect sizes (CLR) are reported alongside FDR-adjusted p-values in order to highlight not only statistically significant differences, but also the magnitude of compositional shifts between environments. The use of an effect size threshold (|effect size| \u0026ge; 0.5) ensures that only taxa showing biologically meaningful changes are retained for interpretation.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003eWe then analysed alpha and beta diversity within each environment, but no significant differences were observed among water sources within the same environment for any metric (Figs.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003e \u0026amp; \u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003e). Since marine site samples were initially negative for PRV1 and later tested positive, we investigated whether PRV1 status (positive vs negative) was associated with changes in alpha and beta diversity or differentially abundant taxa in seawater cage samples at the genus level. Alpha diversity did not show significant differences, although a trend toward decreased richness was observed when PRV1 was detected in water samples (Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003ea). Principal Coordinates Analysis (PCoA) based on Bray\u0026ndash;Curtis dissimilarity\u003csup\u003e\u003cspan citationid=\"CR33\" class=\"CitationRef\"\u003e33\u003c/span\u003e\u003c/sup\u003e illustrated beta diversity patterns between PRV1 groups, but PERMANOVA results (R\u0026sup2; = 0.162, p\u0026thinsp;=\u0026thinsp;0.14) indicated no significant differences (Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003eb). Differential abundance analysis of the top six genera suggested a potential increase in bacteriophage-associated taxa in PRV1-positive samples (Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003ec), although the volcano plot showed no taxa meeting the significance threshold (FDR\u0026thinsp;\u0026le;\u0026thinsp;0.1 and |effect size| \u0026ge; 0.5) (Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003ed). The lack of taxa meeting the significance threshold likely reflects limited effect magnitudes and reduced statistical power, rather than the absence of true biological differences. This is in line with recommendations for microbiome studies, which emphasise the need to interpret effect sizes together with p-values due to the high variability and constrained sample sizes typical of environmental sequencing datasets\u003csup\u003e\u003cspan citationid=\"CR36\" class=\"CitationRef\"\u003e36\u003c/span\u003e\u003c/sup\u003e.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec7\" class=\"Section2\"\u003e \u003ch2\u003e2.4 Databases\u0026rsquo; comparison\u003c/h2\u003e \u003cp\u003eWe compared the taxonomic classification using Kraken2 with standard vs custom pathogen reference databases at the species level. The custom pathogen reference database reflects the presence of strain and species-specific Atlantic salmon pathogenic taxa, highlighting the importance of using databases tailored to address the questions at hand. This comparison showed the total absence of \u003cem\u003eFrancisella noatunensis\u003c/em\u003e, \u003cem\u003eMoritella viscosa\u003c/em\u003e and \u003cem\u003eTenacibaculum piscium\u003c/em\u003e in hatchery and seawater site samples by the standard database, while \u003cem\u003eRenibacterium salmoninarum\u003c/em\u003e and \u003cem\u003eTenacibaculum maritimum\u003c/em\u003e were identified only from the standard database (Figs.\u0026nbsp;\u003cspan refid=\"Fig6\" class=\"InternalRef\"\u003e6\u003c/span\u003ea, \u003cspan refid=\"Fig6\" class=\"InternalRef\"\u003e6\u003c/span\u003eb \u0026amp; \u003cspan refid=\"Fig6\" class=\"InternalRef\"\u003e6\u003c/span\u003ec). In the wellboat water samples, the standard databases failed to detect most of the pathogenic species, especially after the treatment (Figs.\u0026nbsp;\u003cspan refid=\"Fig6\" class=\"InternalRef\"\u003e6\u003c/span\u003ea \u0026amp; \u003cspan refid=\"Fig6\" class=\"InternalRef\"\u003e6\u003c/span\u003ec). Therefore, the custom reference database improved the detection of pathogenic species in Atlantic salmon, compared to the standard database.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003c/div\u003e"},{"header":"Discussion","content":"\u003cp\u003eInfectious diseases represent a major challenge for the Atlantic salmon industry, and pathogen surveillance is a key preventive measure to limit pathogen spread and enable fish farm personnel to act before clinical symptoms appear in the population. Early detection is particularly important when pathogens are shed into the farm environment at high concentrations and persist over time. This study primarily aimed to assess the practicality and performance of a combined qPCR–shotgun metagenomics workflow for routine pathogen surveillance in aquaculture, using farm-deployable filtration and portable sequencing technologies. In addition to demonstrating the feasibility of implementing this approach under real production conditions, we also evaluated whether metagenomic data could provide complementary ecological context, including microbial community patterns potentially associated with pathogen presence. This methodological emphasis guided the design and interpretation of the study.\u003c/p\u003e \u003cp\u003eWe used a previously established method for filtering, storing, and extracting eDNA/eRNA from water samples\u003csup\u003e\u003cspan class=\"CitationRef\"\u003e18\u003c/span\u003e\u003c/sup\u003e and tested it under real farm conditions, with the support of trained fish farm personnel who evaluated the practicality of the filtration procedure for surveillance purposes. This approach was applied while monitoring the same Atlantic salmon population from the hatchery stage through seawater transfer until slaughter. We used specific targeted qPCR assays, and we were able to detect ISAV in all water sources (outlet-, tank- and inlet waters) at similar concentrations during the smoltification process, while PRV1 was detected in water inside the pens from August 2023. PRV1 is highly prevalent in Norwegian aquaculture: the Norwegian Fish Health Report 2024 documents that the virus was detected d at 336 sites, which is a significant increase compared to numbers from 2023 (267), confirming that PRV1 circulates widely in both freshwater and seawater production environments\u003csup\u003e\u003cspan class=\"CitationRef\"\u003e2\u003c/span\u003e\u003c/sup\u003e. Surveillance data also show recurrent detections in wild and escaped salmon. Thus, the low-level PRV1 signals identified in our water samples are arguably consistent with what might be expected in terms of background viral prevalence in Norwegian production systems. According to the farm’s fish health records, no notifiable diseases were reported throughout the full production cycle. Routine diagnostics identified low‑level ISAV (ISAV-HRP0) presence in the hatchery and PRV1 at the seawater site, but neither was associated with increased morbidity or mortality. No additional pathogens or infectious events were reported during the entire production cycle. Furthermore, pathogen surveillance during wellboat delousing revealed a marked increase in PRV1 concentration in water samples post-treatment.\u003c/p\u003e \u003cp\u003eEnvironmental pathogen surveillance inherently involves low target concentrations, which frequently produce high Cq values near the LOD. These signals must be interpreted with caution; however, when reproducible across biological replicates and consistent with expected shedding patterns, they remain informative for early detection purposes. None of the targeted pathogens were detected outside the pens, highlighting the importance of sampling location for accurate pathogen surveillance in open-water cages. Hydrodynamic transport between the sea pens and surrounding waters is expected in open water aquaculture systems. Although current data for the site were not available, the low pathogen loads detected suggest that any water exchange did not obscure the differentiation between pen-adjacent and offshore samples. Future studies incorporating site-specific current measurements could help refine the spatial interpretation of waterborne pathogen signals.\u003c/p\u003e \u003cp\u003eWe also analysed the same water samples for microbial community composition in the laboratory to assess the feasibility of using portable sequencing technologies, such as the Oxford Nanopore MinION, to evaluate whether indicators of balanced or disrupted health in fish farms could support targeted pathogen-level screening. The hatchery exhibited a predominance of phyla associated with nutrient-rich environments\u003csup\u003e\u003cspan class=\"CitationRef\"\u003e20\u003c/span\u003e,\u003cspan class=\"CitationRef\"\u003e21\u003c/span\u003e\u003c/sup\u003e, whereas seawater showed an expected seasonality in microbial community composition\u003csup\u003e\u003cspan class=\"CitationRef\"\u003e22\u003c/span\u003e,\u003cspan class=\"CitationRef\"\u003e23\u003c/span\u003e\u003c/sup\u003e. Across taxonomic levels, seawater cage sites consistently displayed higher diversity and evenness, particularly at the order, family, and genus levels, compared to the hatchery environment. Richness remained similar between environments at the genus level, suggesting that observed differences were driven by community structure rather than species count. However, alpha and beta diversity showed no significant differences among water sources within the same environment. PRV1-positive samples exhibited trends such as decreased microbial richness and alterations in the virome. Marine environments harbour a substantially higher abundance and diversity of bacteriophages, and their presence varies according to season, temperature and photic shifts, and according to targeted hosts, such as the “\u003cem\u003eCandidatus\u003c/em\u003e Pelagibacter” (SAR11) in cold water habitats\u003csup\u003e\u003cspan class=\"CitationRef\"\u003e37\u003c/span\u003e\u003c/sup\u003e. Bacteriophages play a pivotal role in marine microbial food webs by regulating bacterial mortality, recycling nutrients through the viral shunt, and influencing biogeochemical cycles, particularly carbon and nitrogen fluxes\u003csup\u003e\u003cspan class=\"CitationRef\"\u003e38\u003c/span\u003e\u003c/sup\u003e. In our study, the differential abundance of \u003cem\u003eLlyrvirus\u003c/em\u003e, \u003cem\u003ePelagivirus\u003c/em\u003e and \u003cem\u003eSiovirus\u003c/em\u003e was significantly higher in seawater sites than in the hatchery, while PRV1-positive samples showed an increased differential abundance of \u003cem\u003eIgirivirus\u003c/em\u003e, \u003cem\u003ePowvirus\u003c/em\u003e, and \u003cem\u003eSednavirus\u003c/em\u003e. All these viruses are marine phages belonging to the class \u003cem\u003eCaudoviricetes\u003c/em\u003e (tailed dsDNA phages)\u003csup\u003e\u003cspan class=\"CitationRef\"\u003e39\u003c/span\u003e\u003c/sup\u003e. These observations suggest possible early indications of microbiome and virome shifts, which might be associated with pathogen presence. However, given the low pathogen loads (high Cq values), the stable health status of the fish population, and the absence of clinical disease outbreaks, these patterns should be interpreted as preliminary and potentially influenced by seasonal environmental variability rather than as definitive signs of dysbiosis. The key limitation of our study was that the selected fish groups exhibited good health status, which significantly constrained our ability to establish meaningful correlations between pathogen presence and indicators of health or dysbiosis based on microbiome composition. This uncertainty highlights the need for future studies conducted under more variable or compromised health conditions, where stronger pathogen signals or disease events may allow clearer interpretation of community-level ecological changes. Such studies will be essential to determine whether the subtle trends observed here represent biologically relevant responses or fall within the expected range of natural variability.\u003c/p\u003e \u003cp\u003eWhile low-level human DNA contamination was detected in some samples, this signal was taxonomically restricted and completely removed through explicit filtering of all human-associated reads. Importantly, no evidence of cross-contamination with any of the microorganisms analysed in this study was observed: all negative controls remained free of target pathogens, and microbial community profiles showed coherent ecological patterns rather than signatures of random contamination. Therefore, contamination is unlikely to have influenced the prokaryotic or viral fractions retained for downstream analysis.\u003c/p\u003e \u003cp\u003eAlthough shotgun metagenomics can also be used for antimicrobial resistance (AMR) and virulence gene profiling, this was not included in our analyses because sequencing depth and filtering parameters were optimised for taxonomic classification rather than functional annotation. In addition, the presence of human-associated contamination in some samples could confound AMR interpretation. Future studies with higher sequencing depth and stricter assembly criteria will be required to robustly integrate AMR screening into this workflow. Moreover, although \u003cem\u003eStreptomyces\u003c/em\u003e was detected in some samples, no antibiotic treatments were applied during the production cycle. \u003cem\u003eStreptomyces\u003c/em\u003e is commonly found in aquatic and soil-derived environmental microbiota, and its presence likely reflects natural background communities rather than the use of antimicrobial agents. Furthermore, no high-quality metagenome‑assembled genomes (MAGs) were reconstructed in this study. The sequencing depth achievable with portable field‑deployable Nanopore runs, together with the high complexity and viral load of marine water samples, was insufficient for robust genome binning. Higher‑coverage sequencing and longer runs will be required to recover MAGs in future applications. Despite all limitations, we successfully established a protocol to extract sufficient DNA from water samples for shotgun metagenomic sequencing, where pipeline quality metrics and taxonomic classification of positive controls confirmed the robustness of the procedure. Moreover, our study emphasised two critical aspects: (i) the persistent issue of human contamination in field samples and its associated microbiota, despite precautions to maintain sterile conditions, and (ii) the importance of reference database selection for achieving comprehensive and accurate taxonomic profiling. Most published studies rely on 16S rRNA gene sequencing\u003csup\u003e\u003cspan class=\"CitationRef\"\u003e12\u003c/span\u003e–\u003cspan class=\"CitationRef\"\u003e14\u003c/span\u003e\u003c/sup\u003e, which provides only a bacterial community snapshot and does not capture viral components. In contrast, our shotgun metagenomics approach revealed shifts in the virome, highlighting the ecological relevance of viruses and bacteriophages in farm water, where they interact closely with fish and their mucosal surfaces. Furthermore, we demonstrate that this workflow can be integrated into targeted DNA-based pathogen screening by using a curated reference database containing high-quality genomes of Atlantic salmon pathogens. These curated entries were added to the Kraken2 library to improve classification accuracy. However, sequence matches alone should not be interpreted as definitive evidence of pathogen presence, as short regions of high similarity may reflect closely related environmental taxa rather than the target pathogen. This was evident when comparing taxonomic classifications using Kraken2 with the standard PlusPFP database vs our custom reference database. This comparison has important implications for practical pathogen surveillance in aquaculture. Standard reference databases such as PlusPFP are highly comprehensive but may occasionally produce false positives for high-consequence pathogens — for example, \u003cem\u003eR. salmoninarum\u003c/em\u003e — due to short conserved genomic regions shared with related environmental taxa rather than true pathogen presence. In contrast, curated pathogen-specific databases substantially reduce this risk by restricting the search space to well-annotated, biologically relevant genomes. Our custom database was assembled using stringent criteria, including the selection of complete or near-complete assemblies from NCBI, the exclusion of poorly annotated or fragmented genomes, and the retention only of taxa with confirmed pathogenic relevance for Atlantic salmon. This targeted approach enhances classification accuracy, reduces taxonomic ambiguity, and provides a more reliable foundation for DNA-based pathogen surveillance in open water farming systems.\u003c/p\u003e \u003cp\u003eTaken together, our findings suggest that this integrated qPCR–metagenomics workflow is likely to be most informative under conditions where pathogen activity or fish stress is elevated — for example, during known or suspected outbreaks, after events such as delousing or grading, or when monitoring multiple cohorts or farms simultaneously. Additional validation across diverse production systems, health statuses, and environmental conditions will be essential to determine the sensitivity, robustness, and operational value of this approach in routine farm-level decision making.\u003c/p\u003e \u003cp\u003eAs a conclusion, our study establishes a baseline for future applications of continuous pathogen monitoring using water samples throughout the production cycle via qPCR, combined with microbiome profiling and adaptive sampling pipelines supported by custom pathogen reference databases. While shotgun metagenomics can reveal ecological patterns relevant for understanding disease risk, its application in low-biomass water matrices requires careful contamination control and conservative interpretation. In contrast, targeted qPCR provides robust and sensitive pathogen detection even under these conditions and remains the most appropriate tool for routine surveillance.\u003c/p\u003e "},{"header":"Methods","content":"\u003ch2\u003e4.1 Sampling workflow\u003c/h2\u003e\u003cp\u003eWater samples were collected from two distinct environments associated with the same Atlantic salmon population (from ca. 250 g to 4–6 kg): a hatchery (including inlet-, tank-, and outlet waters) and the marine grow-out site. In the marine environment, routine monitoring samples were collected from within the sea pen and from two offshore reference locations situated 1 km away from the pen, at depths of 1 m and 20 m. Additionally, during the final sampling, two further offshore samples were collected at 200 m from the pen, also at 1 m and 20 m depth, in order to increase the spatial resolution around the farm during this time point (Fig.\u0026nbsp;\u003cspan class=\"InternalRef\"\u003e7\u003c/span\u003e). In the hatchery, tank‑associated water was collected from a fixed sampling point located immediately adjacent to the tanks, corresponding to the outflow water that had passed through the tank system. At the marine site, water was collected from the dead‑fish collection point, which draws water directly from inside the sea pen and is routinely used by the farm for monitoring purposes, or offshore from a service boat. These sampling points provided water representative of the environment immediately surrounding the fish. Environmental variables (e.g., temperature, salinity, dissolved oxygen) were not available from the farm and could therefore not be incorporated. Given the nature of the sampling procedure and the focus of the study on pathogen detection rather than environmental drivers, this limitation does not influence the interpretation of the results. Additional samples were taken on the wellboat immediately before and after delousing. For each sampling event and location, three biological replicates of 500 mL were collected and filtered on site, together with a 500 mL sterile water control (autoclaved and 0.22 µm-filtered laboratory-grade water). This control was included to detect any contamination introduced during field filtration or laboratory processing (filtered using the same equipment, filters, and workflow as the samples). The choice of 500 mL per sample was based on previous method development work, where this volume provided the best balance between sensitivity and filtration efficiency\u003csup\u003e\u003cspan class=\"CitationRef\"\u003e18\u003c/span\u003e\u003c/sup\u003e. Larger volumes increased the risk of membrane clogging and did not improve detection probability, whereas 500 mL was sufficient to detect even low amounts of viral and bacterial pathogens in field conditions. All replicates were processed using the same filtration setup, on-site and based on the results from a previous work\u003csup\u003e\u003cspan class=\"CitationRef\"\u003e18\u003c/span\u003e\u003c/sup\u003e: a combination of the Mixed Cellulose Ester membrane filter (MF-Millipore® Membrane Filters, 0.45 µm pore size, 47 mm diameter, hydrophilic, Millipore, USA, provided by Merck Life Science AS, Norway) and Glass Fibre Filter (2.0 µm pore size, hydrophilic glass fibre with a binder resin, and 47 mm diameter, Millipore, USA, provided by Merck Life Science AS, Norway) were used in place of the original Nalgene™ filters within the Single Use Analytical Filter Funnels (Thermo Fisher Scientific, USA) (Fig.\u0026nbsp;\u003cspan class=\"InternalRef\"\u003e7\u003c/span\u003e). 500 mL of sampled water and 500 mL of sterile water (control) were filtered with a flow rate setting of 3.8 to 4.0 L/min with the EZ-Stream vacuum pump (Millipore, USA) (Fig.\u0026nbsp;\u003cspan class=\"InternalRef\"\u003e7\u003c/span\u003e). All three biological replicates were analysed by targeted qPCR and (RT)qPCR in duplicate reactions, whereas shotgun metagenomics was performed on one pooled extract per sampling location, generated by combining the three biological replicates to ensure sufficient DNA yield for sequencing. These technical duplicates should not be interpreted as independent biological replicates, which were instead represented by the three separately collected and processed 500 mL water samples per location. This approach provided complete qPCR coverage across all environments and time points, while delivering representative metagenomic profiles for hatchery samples, all seawater sampling events, and both wellboat samples.\u003c/p\u003e\u003cp\u003e \u003c/p\u003e\u003ch2\u003e4.2 eDNA/eRNA extraction and qPCR / RT-qPCR assays\u003c/h2\u003e\u003cp\u003eAutomated DNA and viral RNA extractions were performed on a MagNA Pure 96 instrument (Roche) with the MagNA Pure 96 DNA and Viral NA Large Volume Kit (Roche, France), using Pathogen Universal LV protocol with a sample input volume of 1000 µL and an elution volume of 100 µL, per sample, respectively. The pathogen panel screened at the hatchery and seawater sites was defined by the collaboration agreement with the farm. Only pathogens included in the farm’s authorised diagnostic panel could be tested, and the authors were not permitted to add additional targets. As a result, the hatchery screening did not include some freshwater-specific bacterial pathogens. This study followed the farm’s established surveillance procedures, in line with its methodological focus on evaluating a field-deployable qPCR–metagenomics workflow rather than conducting a full epidemiological assessment. In the hatchery, screening included ISAV, PRV1, piscine myocarditis virus (PMCV), infectious pancreatic necrosis virus (IPNV), and \u003cem\u003eYersinia ruckeri\u003c/em\u003e, while at marine sites the panel comprised ISAV, PRV1, PMCV, IPNV, salmonid alphavirus (SAV), and \u003cem\u003eParamoeba perurans\u003c/em\u003e.\u003c/p\u003e\u003cp\u003eRT-qPCR was carried out for ISAV\u003csup\u003e\u003cspan class=\"CitationRef\"\u003e40\u003c/span\u003e\u003c/sup\u003e, PRV1\u003csup\u003e41\u003c/sup\u003e, PMCV\u003csup\u003e\u003cspan class=\"CitationRef\"\u003e42\u003c/span\u003e\u003c/sup\u003e and SAV\u003csup\u003e\u003cspan class=\"CitationRef\"\u003e43\u003c/span\u003e\u003c/sup\u003e with the Brilliant III Ultra-Fast QRT-PCR Master Mix (Agilent Technologies, USA) using the following protocol: 10 µL of 2× QRT-PCR master mix, 0.2 µL of 100 mM DTT, 1 µL of RT/RNase block, 1 µL of the 20× viral assay (6 µM of assay probe, 10 µM of each forward and reverse primers), and 5 µL extracted viral RNA. Thermocycling was performed on a CFX384 Bio-Rad and CFX-manager (software version 3.1, Bio-Rad, USA) under the following conditions: 10 min at 50°C, 3 min at 95°C, 45 cycles of 5 s at 95°C and 10 s at 60°C. Moreover, IPNV\u003csup\u003e\u003cspan class=\"CitationRef\"\u003e44\u003c/span\u003e\u003c/sup\u003e RT-qPCR was performed with the QIAGEN OneStep RT-PCR Kit (Qiagen, USA) using the following protocol: 4 µL of 5× QIAGEN OneStep RT-PCR Buffer, 0.8 µL of dNTP Mix (containing 10 mM of each dNTP), 1 µL of MgCl\u003csub\u003e2\u003c/sub\u003e (25 mM), 1 µL of the 20× viral assay (6 µM of assay probe, 10 µM of each forward and reverse primers), 0.1 µL of RNaseOUT™ Recombinant Ribonuclease Inhibitor (40 U/µL, Invitrogen™, USA), 0.8 µL of QIAGEN OneStep RT-PCR Enzyme Mix, 7.3 µL of nuclease-free water, and 5 µL extracted viral RNA. Thermocycling was performed on a CFX384 Bio-Rad and CFX-manager (software version 3.1, Bio-Rad, USA) under the following conditions: 30 min at 50°C, 15 min at 95°C, 45 cycles of 30 s at 94°C and 60 s at 60°C.\u003c/p\u003e\u003cp\u003eqPCR was carried out for \u003cem\u003eY. ruckeri\u003c/em\u003e\u003csup\u003e\u003cspan class=\"CitationRef\"\u003e45\u003c/span\u003e\u003c/sup\u003e with TaqMan™ Fast Advanced Master Mix (Applied Biosystems™, USA) using the following protocol: 10 µL of 2× TaqMan Fast Advanced Master Mix, 5 µL of the bacterial assay (10 µM of assay probe and each forward and reverse primers), and 5 µL extracted DNA. Thermocycling was performed on a CFX384 Bio-Rad and CFX-manager (software version 3.1, Bio-Rad, USA) under the following conditions: 2 min at 50°C, 20 s at 95°C, 50 cycles of 3 s each at 95°C, and 20 s at 62°C. Specific qPCR assay for \u003cem\u003eP. perurans\u003c/em\u003e was carried out with 25 µL reactions consisting of 12.5 µL TaqPath™ qPCR Master Mix, CG (Applied Biosystems™, USA), 500 nM of each primer and 250 nM of probe, nuclease-free water and 5 µL DNA sample\u003csup\u003e\u003cspan class=\"CitationRef\"\u003e46\u003c/span\u003e\u003c/sup\u003e. The following qPCR cycling conditions were used: an initial denaturation at 95°C for 20 s, followed by 50 cycles of denaturation at 95°C for 3 s and annealing at 60°C for 30 s\u003csup\u003e46\u003c/sup\u003e.\u003c/p\u003e\u003cp\u003eAll qPCR assays were analysed in duplicate, and a no-template control (H\u003csub\u003e2\u003c/sub\u003eO) and water sample control from the MagNA Pure 96 extraction protocol were included on each plate as negative controls. Standard curves of serial dilution of the relevant target sequences (CFX-manager software version 3.1, Bio-Rad, USA) were used for the calculation of the concentration used for the experiments. Primers and probes are shown in the supplementary (Supplementary Table \u003cspan class=\"InternalRef\"\u003eS1\u003c/span\u003e). Full standard curves, including amplification efficiencies, R² values, slope, and y-intercept, are provided in the Supplementary Material (Table \u003cspan class=\"InternalRef\"\u003eS2\u003c/span\u003e). These curves were generated using the same matrix composition and extraction conditions as the environmental samples and were used to calculate copy numbers and estimate assay-specific LOD.\u003c/p\u003e\u003cp\u003ePositive qPCR results for ISAV and PRV1 were then repeated to estimate the copy number in 5 µL with the Luna Probe One-Step RT-qPCR 4X Mix with UDG (New England Biolabs, USA) using the following protocol: 5 µL of 4× Luna Probe One-Step RT-qPCR 4X Mix with UDG, 2 µL of the viral assay (0.4 µL of 10 µM of assay probe, 0.8 µL of 10 µM of each forward and reverse primers), 5 µL extracted viral RNA, and 8 µL of nuclease-free water. Thermocycling was performed on a CFX384 Bio-Rad and CFX-manager (software version 3.1, Bio-Rad, USA) under the following conditions: 30 s at 25°C, 10 min at 55°C, 1 min at 95°C, 45 cycles of 10 s at 95°C and 30 s at 60°C. The copy number/µL was calculated using a previously designed sequence-verified gene fragment, or gBlocks™ (Integrated DNA Technologies, BVBA, Leuven, Belgium)\u003csup\u003e\u003cspan class=\"CitationRef\"\u003e18\u003c/span\u003e\u003c/sup\u003e. Briefly, the concentration from ng/µL was converted to copy number/µL by using the formula provided by IDT guidelines. Copy numbers for ISAV and PRV1 quantification were calculated using the standard Integrated DNA Technologies (IDT) formula: copies/µL = (DNA concentration [ng/µL]) × (Molecular weight [fmol/ng) × (1 × 10\u003csup\u003e− 15\u003c/sup\u003e) × (6.022 × 10²³). The synthetic gBlock used in this study had a concentration of 8.408 ng/µL and a molecular weight of 2.36 fmol/ng, corresponding to an estimated 1.19 × 10¹⁰ copy number/µL in the undiluted gBlock. A stock solution of 5 × 10⁷ copy number/µL was prepared by combining 8.4 µL of the undiluted gBlock with 1991.6 µL of TE buffer. Serial 1:4 dilutions were then generated to obtain a standard curve ranging from 5 × 10⁷ to 1 × 10⁻² copy number/µL. Expected copy numbers for each dilution step (1 µL and 5 µL input) are provided in Supplementary Table \u003cspan class=\"InternalRef\"\u003eS2\u003c/span\u003e. All standards were prepared using the same extraction matrix and buffer conditions as the environmental samples to ensure comparable amplification efficiencies. After optimisation, the Stock\u003csup\u003e− 4\u003c/sup\u003e dilution (1.56 x 10\u003csup\u003e4\u003c/sup\u003e copy number/µL) was used as a standard template to estimate the RNA copy numbers/µL for the viral pathogens (ISAV and PRV1) based on the amount captured by the filters (Table \u003cspan class=\"InternalRef\"\u003eS2\u003c/span\u003e). To aid interpretation, qPCR copy numbers initially expressed per 5 µL of extract were converted into copies per litre of water (Table \u003cspan class=\"InternalRef\"\u003eS2\u003c/span\u003e). Based on the standard curves generated under the same matrix and extraction conditions, the estimated limits of detection were ~ 100 copies per reaction for ISAV and ~ 10 copies per reaction for PRV1, corresponding to ~ 4×10³ copies/L and ~ 4×10² copies/L, respectively. Several detections, particularly those with Cq \u0026gt; 35, fell close to these thresholds; therefore, estimates at low copy number should be interpreted cautiously due to increased stochastic variation and reduced quantification precision. Despite this uncertainty, ISAV and PRV1 detections near the LOD were consistent across biological replicates, supporting their biological relevance in the context of environmental shedding.\u003c/p\u003e\u003ch2\u003e4.3 DNA extraction and sequencing\u003c/h2\u003e\u003cp\u003eFor microbial community analysis, 4.5 mL of preserved pooled DNA/RNA in DNA/RNA Shield™ was extracted from each sampling point with the ZymoBIOMICS™ DNA/RNA Miniprep Kit (Zymo Research, USA, provided by Nordic Biosite AS, Sweden), following the manufacturer's specifications. Briefly, 750 µL of liquid sample were transferred into a ZR BashingBead Lysis Tube (0.1 \u0026amp; 0.5 mm) and homogenised with FastPrep®-24 5G bead beating grinder and lysis system (MP Biomedicals, USA) using the following program: 2 cycles for 60 s (300 s of pause), speed of 6.0 m/s – Lys Matrix A, 1 mg. Samples were then centrifuged at 15,000 x g for 1 min and mixed with an equal volume (1:1) of DNA/RNA Lysis Buffer to the supernatant and processed through the DNA and RNA purification kit specifications. To elute the nucleic acids, 80 µL of nuclease-free water was added directly to the column matrix, incubated for 5 min, and then centrifuged at 15,000 x g for 2 min. The final step with the Zymo-Spin™ III-HRC Filter was also performed to remove inhibitors, salts, enzymes, or toxic compounds from the solution containing nucleic acids for sequencing purposes. Sample integrity and fragment size were assessed using the Genomic DNA (gDNA) ScreenTape assay on the TapeStation 4150 System (Agilent Technologies, USA). DNA Integrity Numbers (DIN) from water samples were greater than 6, with an average concentration of 10 ng/µL and a fragment size of approximately 10,000 base pairs. For sequencing, the Ligation Sequencing Kit V14 – PCR Barcoding protocol (SQK-LSK114 with EXP-PBC001, Oxford Nanopore Technologies, UK) was applied according to the manufacturer’s instructions, with some modifications: 45 µL of each sample were used as input material, 18 cycles were performed during the PCR barcoding step, and 50 fmol were loaded onto the flow cell. Sequencing was carried out using a MinION Flow Cell (R10.4.1) on the MinION Mk1B device, with the MinKNOW™ software version 25.09.16, excluding basecalling and saving pod5 files. Each sequencing run included two positive controls: 100 ng of the ZymoBIOMICS® Microbial Community DNA Standard (Zymo Research, USA; supplied by Nordic Biosite AS, Sweden), and 45 µL of a sample from a previously published pilot experiment spiked with known concentrations of selected pathogens\u003csup\u003e\u003cspan class=\"CitationRef\"\u003e18\u003c/span\u003e\u003c/sup\u003e. Negative controls were also included for each sampling point (sterile Milli-Q water filtered on site), along with an extraction control containing only DNA/RNA Shield™.\u003c/p\u003e\u003ch2\u003e4.4 Bioinformatic analysis\u003c/h2\u003e\u003cp\u003eOxford Nanopore ligation sequencing data were processed using the NEPAL pipeline (Norwegian Veterinary Institute)\u003csup\u003e\u003cspan class=\"CitationRef\"\u003e47\u003c/span\u003e\u003c/sup\u003e, a reproducible workflow built on Nextflow for processing raw nanopore data. The pipeline was configured to evaluate three processing strategies: simplex, duplex, and NanoFilt-filtered reads, enabling comparative analysis of read quality and downstream performance. Basecalling was performed using Dorado v0.2.1 (Oxford Nanopore Technologies), integrated into the NEPAL pipeline. The super accuracy model [email protected] was used to ensure the highest possible basecalling accuracy. NanoFilt was applied to duplex reads following the filtering parameters: minimum quality score: 9; minimum read length: 300 bp (to exclude very short fragments); and maximum read length: 2,147,483,647 bp (a technical upper limit to avoid excluding long reads). Outputs from each workflow were compared in terms of read quality, length distribution, GC content, and assembly metrics. Statistical analysis among the three processing strategies was performed using the Kruskal–Wallis non-parametric test\u003csup\u003e\u003cspan class=\"CitationRef\"\u003e48\u003c/span\u003e\u003c/sup\u003e and displayed on the plots. When the Kruskal–Wallis test indicated significance (p ≤ 0.05), Wilcoxon rank-sum tests\u003csup\u003e\u003cspan class=\"CitationRef\"\u003e34\u003c/span\u003e\u003c/sup\u003e were performed for pairwise comparisons, and p-values were adjusted using the Benjamini–Hochberg (BH) method\u003csup\u003e\u003cspan class=\"CitationRef\"\u003e35\u003c/span\u003e\u003c/sup\u003e to control the false discovery rate. After NanoFilt filtering, taxonomic classification was performed using the nf-core\u003csup\u003e\u003cspan class=\"CitationRef\"\u003e49\u003c/span\u003e\u003c/sup\u003e taxprofiler pipeline with Kraken2\u003csup\u003e50\u003c/sup\u003e as the classifier. The pipeline was configured to use the Kraken2 PlusPFP database (\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://benlangmead.github.io/aws-indexes/k2\u003c/span\u003e\u003cspan class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e) and the long-read mode (--long) for optimal performance on nanopore data. To avoid host contamination, sequences from the Atlantic salmon genome (\u003cem\u003eS. salar\u003c/em\u003e; reference: Ssal_v3.1; GCA_905237065.2) were excluded from the database. The pipeline generated taxonomic profiles at multiple ranks (domain, phylum, order, class, family, genus, and species) and summary reports for downstream comparative analysis. Statistical analyses and visualisation were performed in R (v4.4.0)\u003csup\u003e51\u003c/sup\u003e using RStudio (v2024.04.0 + 735)\u003csup\u003e52\u003c/sup\u003e. After an initial taxonomic composition assessment, reads classified as \u003cem\u003eH. sapiens\u003c/em\u003e and all sequences assigned to Eukaryota or Archaea were excluded to focus on prokaryotic and viral fractions relevant to fish health analyses. In the taxonomic workflow, human-associated taxa (e.g., \u003cem\u003eH. sapiens\u003c/em\u003e, \u003cem\u003eHomo\u003c/em\u003e, Hominidae) were explicitly removed during data processing. In the diversity workflow, the entire Eukaryota domain was filtered out in R, which also eliminated human-derived and other nonbacterial/nonviral reads. As a result, only non-host and non-human microbial taxa were retained for all taxonomic, alpha-diversity, beta-diversity, and differential abundance analyses. Alpha diversity indices—species richness, Shannon diversity, Simpson diversity, and Pielou’s evenness—were calculated from relative abundances. Group comparisons (e.g., hatchery vs seawater sites or water sources in each environment) used Wilcoxon rank-sum\u003csup\u003e\u003cspan class=\"CitationRef\"\u003e34\u003c/span\u003e\u003c/sup\u003e or Kruskal–Wallis\u003csup\u003e\u003cspan class=\"CitationRef\"\u003e48\u003c/span\u003e\u003c/sup\u003e tests with FDR correction (Benjamini–Hochberg\u003csup\u003e\u003cspan class=\"CitationRef\"\u003e35\u003c/span\u003e\u003c/sup\u003e). Beta diversity was based on Bray–Curtis dissimilarity\u003csup\u003e\u003cspan class=\"CitationRef\"\u003e33\u003c/span\u003e\u003c/sup\u003e and visualised with PCoA. Group separation was assessed using PERMANOVA (\u003cem\u003eadonis2\u003c/em\u003e\u003csup\u003e\u003cem\u003e53\u003c/em\u003e\u003c/sup\u003e) and multivariate dispersion tests (\u003cem\u003ebetadisper\u003c/em\u003e\u003csup\u003e\u003cem\u003e\u003cspan class=\"CitationRef\"\u003e54\u003c/span\u003e\u003c/em\u003e\u003c/sup\u003e), as implemented in the vegan R package\u003csup\u003e\u003cspan class=\"CitationRef\"\u003e55\u003c/span\u003e\u003c/sup\u003e. Confidence ellipses (95%) were drawn for visual separation among environments. Differential abundance analyses were assessed with centred log-ratio (CLR) transformation and Dirichlet-Monte-Carlo model using the \u003cem\u003eALDEx2\u003c/em\u003e package\u003csup\u003e\u003cspan class=\"CitationRef\"\u003e56\u003c/span\u003e\u003c/sup\u003e, a negative binomial model to raw counts was used to estimate log₂ fold changes between groups with the \u003cem\u003eDESeq2\u003c/em\u003e package\u003csup\u003e\u003cspan class=\"CitationRef\"\u003e57\u003c/span\u003e\u003c/sup\u003e, and non-parametric comparison of log-ratio-transformed relative abundances were assessed with Wilcoxon rank-sum\u003csup\u003e\u003cspan class=\"CitationRef\"\u003e34\u003c/span\u003e\u003c/sup\u003e test with FDR correction (Benjamini–Hochberg\u003csup\u003e\u003cspan class=\"CitationRef\"\u003e35\u003c/span\u003e\u003c/sup\u003e) (FDR ≤ 0.10 and |effect size| ≥ 0.5). We applied an FDR threshold of ≤ 0.10, consistent with recommendations for flexible false discovery control in large-scale omics analyses, where rigid cut-offs (e.g., FDR ≤ 0.05) may be overly conservative and reduce sensitivity in high-dimensional datasets\u003csup\u003e\u003cspan class=\"CitationRef\"\u003e58\u003c/span\u003e\u003c/sup\u003e. We report effect sizes alongside p-values, following recommendations that highlight the importance of effect magnitude in interpreting microbiome analyses, particularly in studies where statistical power may be limited\u003csup\u003e\u003cspan class=\"CitationRef\"\u003e36\u003c/span\u003e\u003c/sup\u003e. A Custom Kraken2 database was developed to compare the taxonomic classification of fish pathogens with the standard Kraken2 PlusPFP database. The Standard PlusPFP database includes bacterial, archaeal, viral, fungal, and protozoan genomes distributed with Kraken2. The Custom database was built with reference genomes and complete assemblies of known Atlantic salmon pathogens, including \u003cem\u003eR. salmoninarum\u003c/em\u003e, Salmon gill poxvirus, \u003cem\u003ePhocoenobacter skyensis\u003c/em\u003e, \u003cem\u003ePhocoenobacter atlanticus\u003c/em\u003e subsp. \u003cem\u003eatlanticus\u003c/em\u003e, \u003cem\u003eY. ruckeri\u003c/em\u003e, \u003cem\u003eM. viscosa\u003c/em\u003e, \u003cem\u003eAeromonas salmonicida\u003c/em\u003e subsp. \u003cem\u003esalmonicida\u003c/em\u003e, \u003cem\u003ePiscirickettsia salmonis\u003c/em\u003e, \u003cem\u003eT. maritimum\u003c/em\u003e, \u003cem\u003eTenacibaculum dicentrarchi\u003c/em\u003e, \u003cem\u003eT. piscium\u003c/em\u003e, \u003cem\u003eTenacibaculum finnmarkense\u003c/em\u003e, and \u003cem\u003eF. noatunensis\u003c/em\u003e. Genomes were retrieved from NCBI RefSeq/GenBank, and they were manually curated to ensure completeness and taxonomic accuracy. The new entries were added to the Kraken2 library, and the database index was rebuilt. Comparative visualisations (species-level relative abundance barplots and Venn overlaps) were generated in R to identify taxa uniquely detected by either database.\u003c/p\u003e"},{"header":"Declarations","content":"\u003ch2\u003eCompeting interests\u003c/h2\u003e\n\u003cp\u003eThe author(s) declare no competing interests.\u003c/p\u003e\n\u003ch2\u003eFunding Declaration\u003c/h2\u003e\n\u003cp\u003eThe study was funded by the Research Council of Norway (RCN) through the project \u0026ldquo;SusOffAqua - Unleashing the sustainable value creation potential of offshore ocean aquaculture\u0026rdquo; (project number 328724) and by the FHF - Fiskeri- og havbruksn\u0026aelig;ringens forskningsfinansiering through the project \u0026ldquo;Biosikkerhetstiltak mot ILA i settefisk (ILA-SAFE)\u0026rdquo; (project number 901674). This work was co-funded by the European Union\u0026apos;s Horizon Europe Project 101136346 EUPAHW. Views and opinions expressed are, however, those of the author(s) only and do not necessarily reflect those of the European Union or the European Research Executive Agency. Neither the European Union nor the granting authority can be held responsible for them.\u003c/p\u003e\n\u003ch2\u003eAuthor Contribution\u003c/h2\u003e\n\u003cp\u003eStudy conceptualisation (O.B., D.S., H.S., and T.V.), experimental work (O.B., S.N.M., M.M.A., and D.S.), data analysis and visualisation (O.B. and S.G.), manuscript writing first draft (O.B., H.S. and T.V.), manuscript editing (O.B., D.S., S.G., M.M.A., S.N.M., H.S., and T.V.). All authors reviewed, edited, and approved the final manuscript.\u003c/p\u003e\n\u003ch2\u003eAcknowledgement\u003c/h2\u003e\n\u003cp\u003eWe thank the staff at the hatchery and open seawater cage for the excellent collaboration. The computations were performed on resources provided by Sigma2 - the National Infrastructure for High-Performance Computing and Data Storage in Norway (project nn10070k).\u003c/p\u003e\n\u003ch2\u003eData Availability\u003c/h2\u003e\n\u003cp\u003eThe datasets generated during and/or analysed during the current study are available at the BioProject PRJNA1379277. All other data generated during this study are included in this published article and its Supplementary Information files.\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\n\u003cli\u003eFAO. \u003cem\u003eInternational markets for fisheries and aquaculture products \u0026ndash; Fourth issue 2024, with January\u0026ndash;June 2024 statistics.\u003c/em\u003e \u0026lt;https://openknowledge.fao.org/handle/20.500.14283/cd3517e\u0026gt; (2024).\u003c/li\u003e\n\u003cli\u003eMoldal, T., Wiik-Nielsen, J., Oliveira, V. H. S., Svendsen, J. C. \u0026amp; Sommerset, I. \u003cem\u003eNorwegian Fish Health Report 2024. Norwegian Veterinary Institute Report series #1b/2025\u003c/em\u003e. (Norwegian Veterinary Institute, 2025).\u003c/li\u003e\n\u003cli\u003eBernhardt, L. V.\u003cem\u003e et al.\u003c/em\u003e Early detection of salmonid alphavirus in seawater from marine farm sites of Atlantic salmon \u003cem\u003eSalmo salar\u003c/em\u003e. \u003cem\u003eDiseases of Aquatic Organisms\u003c/em\u003e \u003cstrong\u003e146\u003c/strong\u003e, 41-52 (2021). https://doi.org:10.3354/dao03618\u003c/li\u003e\n\u003cli\u003eBernhardt, L. V., Myrmel, M., Lillehaug, A., Qviller, L. \u0026amp; Weli, S. C. Filtration, concentration and detection of salmonid alphavirus in seawater during a post-smolt salmon (\u003cem\u003eSalmo salar\u003c/em\u003e) cohabitant challenge. \u003cem\u003eDiseases of Aquatic Organisms\u003c/em\u003e \u003cstrong\u003e144\u003c/strong\u003e, 61-73 (2021). https://doi.org:10.3354/dao03572\u003c/li\u003e\n\u003cli\u003eWeli, S. C., Bernhardt, L. V., Qviller, L., Myrmel, M. \u0026amp; Lillehaug, A. Development and evaluation of a method for concentration and detection of salmonid alphavirus from seawater. \u003cem\u003eJ Virol Methods\u003c/em\u003e \u003cstrong\u003e287\u003c/strong\u003e (2021). https://doi.org:10.1016/j.jviromet.2020.113990\u003c/li\u003e\n\u003cli\u003eBohara, K., Yadav, A. K. \u0026amp; Joshi, P. Detection of Fish Pathogens in Freshwater Aquaculture Using eDNA Methods. \u003cem\u003eDiversity\u003c/em\u003e \u003cstrong\u003e14\u003c/strong\u003e (2022). https://doi.org:10.3390/d14121015\u003c/li\u003e\n\u003cli\u003eArbon, P.\u003cem\u003e et al.\u003c/em\u003e Passive eDNA sampling to monitor infectious hypodermal and haematopoietic necrosis virus (IHHNV) in pond-reared giant black tiger shrimp (\u003cem\u003ePenaeus monodon\u003c/em\u003e) aquaculture systems. \u003cem\u003eAquaculture\u003c/em\u003e, 742709 (2025). https://doi.org:10.1016/j.aquaculture.2025.742709\u003c/li\u003e\n\u003cli\u003eKrishna, D.\u003cem\u003e et al.\u003c/em\u003e Environmental DNA/RNA for non-invasive early detection and monitoring of pathogen dynamics in Atlantic salmon (\u003cem\u003eSalmo salar\u003c/em\u003e) recirculating aquaculture systems (RAS). \u003cem\u003eAquaculture\u003c/em\u003e \u003cstrong\u003e611\u003c/strong\u003e (2025). https://doi.org:10.1016/j.aquaculture.2025.743060\u003c/li\u003e\n\u003cli\u003ePires, N. M. M., Dong, T., Yang, Z. \u0026amp; da Silva, L. F. B. A. Recent methods and biosensors for foodborne pathogen detection in fish: Progress and future prospects to sustainable aquaculture systems. \u003cem\u003eCritical Reviews in Food Science and Nutrition\u003c/em\u003e \u003cstrong\u003e61\u003c/strong\u003e, 1852-1876 (2021). https://doi.org:10.1080/10408398.2020.1767032\u003c/li\u003e\n\u003cli\u003eGrunow, B. \u0026amp; Strauch, S. M. Status assessment and opportunities for improving fish welfare in animal experimental research according to the 3R-Guidelines. \u003cem\u003eRev Fish Biol Fisher\u003c/em\u003e \u003cstrong\u003e33\u003c/strong\u003e, 1075-1093 (2023). https://doi.org:10.1007/s11160-023-09781-8\u003c/li\u003e\n\u003cli\u003eGundi, V. A. K. B.\u003cem\u003e et al.\u003c/em\u003e Microbial pathogens in aquaculture: a review of emerging threats. \u003cem\u003eAcademia Biology\u003c/em\u003e \u003cstrong\u003e3\u003c/strong\u003e (2025). https://doi.org:10.20935/AcadBiol7814\u003c/li\u003e\n\u003cli\u003eRieder, J., Kapopoulou, A., Bank, C. \u0026amp; Adrian-Kalchhauser, I. Metagenomics and metabarcoding experimental choices and their impact on microbial community characterization in freshwater recirculating aquaculture systems. \u003cem\u003eEnvironmental Microbiome\u003c/em\u003e \u003cstrong\u003e18\u003c/strong\u003e, 8 (2023). https://doi.org:10.1186/s40793-023-00459-z\u003c/li\u003e\n\u003cli\u003eMacedo, D. B.\u003cem\u003e et al.\u003c/em\u003e New perspectives on metagenomic analysis for pathogen monitoring in sustainable freshwater aquaculture production: a systematic review. \u003cem\u003eFrontiers in Freshwater Science\u003c/em\u003e \u003cstrong\u003e2\u003c/strong\u003e, 1459233 (2024). https://doi.org:10.3389/ffwsc.2024.1459233\u003c/li\u003e\n\u003cli\u003eValenzuela-Miranda, D.\u003cem\u003e et al.\u003c/em\u003e Nanopore-based metagenomic approaches for detection of bacterial pathogens in recirculating aquaculture systems. \u003cem\u003eFishes-Basel\u003c/em\u003e \u003cstrong\u003e10\u003c/strong\u003e, 496 (2025). https://doi.org:10.3390/fishes10100496\u003c/li\u003e\n\u003cli\u003eAcharya, K.\u003cem\u003e et al.\u003c/em\u003e Metagenomic water quality monitoring with a portable laboratory. \u003cem\u003eWater Research\u003c/em\u003e \u003cstrong\u003e184\u003c/strong\u003e, 116112 (2020). https://doi.org:10.1016/j.watres.2020.116112\u003c/li\u003e\n\u003cli\u003eDjikeng, A., Kuzmickas, R., Anderson, N. G. \u0026amp; Spiro, D. J. Metagenomic analysis of RNA viruses in a fresh water lake. \u003cem\u003ePLOS One\u003c/em\u003e \u003cstrong\u003e4\u003c/strong\u003e, e7264 (2009). https://doi.org:10.1371/journal.pone.0007264\u003c/li\u003e\n\u003cli\u003eLiu, Z.\u003cem\u003e et al.\u003c/em\u003e Comparative metagenomic analysis of microbial community compositions and functions in cage aquaculture and its nearby non-aquaculture environments. \u003cem\u003eFront Microbiol\u003c/em\u003e \u003cstrong\u003e15\u003c/strong\u003e, 1398005 (2024). https://doi.org:10.3389/fmicb.2024.1398005\u003c/li\u003e\n\u003cli\u003eBenedicenti, O.\u003cem\u003e et al.\u003c/em\u003e A refinement to eRNA and eDNA-based detection methods for reliable and cost-efficient screening of pathogens in Atlantic salmon aquaculture. \u003cem\u003ePLOS One\u003c/em\u003e \u003cstrong\u003e19\u003c/strong\u003e, e0312337 (2024). https://doi.org:10.1371/journal.pone.0312337\u003c/li\u003e\n\u003cli\u003eSaggu, S. K., Kumar, M. \u0026amp; Kumar, S. Metagenomics and its impact on environmental and therapeutic microbiology. \u003cem\u003eArchives of Microbiology\u003c/em\u003e \u003cstrong\u003e208\u003c/strong\u003e, 24 (2026). https://doi.org:10.1007/s00203-025-04588-6\u003c/li\u003e\n\u003cli\u003eMinich, J. J.\u003cem\u003e et al.\u003c/em\u003e Microbial ecology of Atlantic salmon (\u003cem\u003eSalmo salar\u003c/em\u003e) hatcheries: impacts of the built environment on fish mucosal microbiota. \u003cem\u003eAppl Environ Microb\u003c/em\u003e \u003cstrong\u003e86\u003c/strong\u003e, e00411-00420 (2020). https://doi.org:10.1128/AEM.00411-20\u003c/li\u003e\n\u003cli\u003eFiedler, A. W., Dr\u0026aring;gen, M. K. R., Lorentsen, E. D., Vadstein, O. \u0026amp; Bakke, I. The stability and composition of the gut and skin microbiota of Atlantic salmon throughout the yolk sac stage. \u003cem\u003eFront Microbiol\u003c/em\u003e \u003cstrong\u003e14\u003c/strong\u003e, 1177972 (2023). https://doi.org:10.3389/fmicb.2023.1177972\u003c/li\u003e\n\u003cli\u003eWuchter, C.\u003cem\u003e et al.\u003c/em\u003e Archaeal nitrification in the ocean. \u003cem\u003eProceedings of the National Academy of Sciences\u003c/em\u003e \u003cstrong\u003e103\u003c/strong\u003e, 12317-12322 (2006). https://doi.org:10.1073/pnas.0600756103\u003c/li\u003e\n\u003cli\u003eZhou, K.\u003cem\u003e et al.\u003c/em\u003e Genomic and transcriptomic insights into complex virus\u0026ndash;prokaryote interactions in marine biofilms. \u003cem\u003eThe ISME Journal\u003c/em\u003e \u003cstrong\u003e17\u003c/strong\u003e, 2303-2312 (2023). https://doi.org:10.1038/s41396-023-01546-2\u003c/li\u003e\n\u003cli\u003eHa, A. D., Moniruzzaman, M. \u0026amp; Aylward, F. O. Assessing the biogeography of marine giant viruses in four oceanic transects. \u003cem\u003eISME Communications\u003c/em\u003e \u003cstrong\u003e3\u003c/strong\u003e, 43 (2023). https://doi.org:10.1038/s43705-023-00252-6\u003c/li\u003e\n\u003cli\u003eAvcı, B., Kr\u0026uuml;ger, K., Fuchs, B. M., Teeling, H. \u0026amp; Amann, R. I. Polysaccharide niche partitioning of distinct Polaribacter clades during North Sea spring algal blooms. \u003cem\u003eThe ISME Journal\u003c/em\u003e \u003cstrong\u003e14\u003c/strong\u003e, 1369-1383 (2020). https://doi.org:10.1038/s41396-020-0601-y\u003c/li\u003e\n\u003cli\u003eBola\u0026ntilde;os, L. M.\u003cem\u003e et al.\u003c/em\u003e Influence of short and long term processes on SAR11 communities in open ocean and coastal systems. \u003cem\u003eISME Communications\u003c/em\u003e \u003cstrong\u003e2\u003c/strong\u003e, 116 (2022). https://doi.org:10.1038/s43705-022-00198-1\u003c/li\u003e\n\u003cli\u003eWeynberg, K. D., Allen, M. J. \u0026amp; Wilson, W. H. Marine prasinoviruses and their tiny plankton hosts: a review. \u003cem\u003eViruses\u003c/em\u003e \u003cstrong\u003e9\u003c/strong\u003e, 43 (2017). https://doi.org:10.3390/v9030043\u003c/li\u003e\n\u003cli\u003eKashulin, A., Seredkina, N. \u0026amp; S\u0026oslash;rum, H. Cold‐water vibriosis. The current status of knowledge. \u003cem\u003eJournal of Fish Diseases\u003c/em\u003e \u003cstrong\u003e40\u003c/strong\u003e, 119-126 (2017). https://doi.org:10.1111/jfd.12465\u003c/li\u003e\n\u003cli\u003eRichards, G. P.\u003cem\u003e et al.\u003c/em\u003e Mechanisms for \u003cem\u003ePseudoalteromonas piscicida\u003c/em\u003e-induced killing of vibrios and other bacterial pathogens. \u003cem\u003eAppl Environ Microb\u003c/em\u003e \u003cstrong\u003e83\u003c/strong\u003e, e00175-00117 (2017). https://doi.org:10.1128/AEM.00175-17\u003c/li\u003e\n\u003cli\u003eDeming, J. W., Somers, L. K., Straube, W. L., Swartz, D. G. \u0026amp; Macdonell, M. T. Isolation of an obligately barophilic bacterium and description of a new genus, \u003cem\u003eColwellia\u003c/em\u003e gen. nov. \u003cem\u003eSyst Appl Microbiol\u003c/em\u003e \u003cstrong\u003e10\u003c/strong\u003e, 152-160 (1988). https://doi.org:10.1016/S0723-2020(88)80030-4\u003c/li\u003e\n\u003cli\u003eBigalke, A., Meyer, N., Papanikolopoulou, L. A., Wiltshire, K. H. \u0026amp; Pohnert, G. The algicidal bacterium \u003cem\u003eKordia algicida\u003c/em\u003e shapes a natural plankton community. \u003cem\u003eAppl Environ Microb\u003c/em\u003e \u003cstrong\u003e85\u003c/strong\u003e, e02779-02718 (2019). https://doi.org:10.1128/AEM.02779-18\u003c/li\u003e\n\u003cli\u003eNicholls, S. M., Quick, J. C., Tang, S. \u0026amp; Loman, N. J. Ultra-deep, long-read nanopore sequencing of mock microbial community standards. \u003cem\u003eGigaScience\u003c/em\u003e \u003cstrong\u003e8\u003c/strong\u003e, giz043 (2019). https://doi.org:10.1093/gigascience/giz043\u003c/li\u003e\n\u003cli\u003eBray, J. R. \u0026amp; Curtis, J. T. An ordination of the upland forest communities of southern Wisconsin. \u003cem\u003eEcological Monographs\u003c/em\u003e \u003cstrong\u003e27\u003c/strong\u003e, 326-349 (1957). https://doi.org:10.2307/1942268\u003c/li\u003e\n\u003cli\u003eWilcoxon, F. Individual comparisons by ranking methods. \u003cem\u003eBiometrics Bulletin\u003c/em\u003e \u003cstrong\u003e1\u003c/strong\u003e, 80-83 (1945). https://doi.org:10.2307/3001968\u003c/li\u003e\n\u003cli\u003eBenjamini, Y. \u0026amp; Hochberg, Y. Controlling the False Discovery Rate: A Practical and Powerful Approach to Multiple Testing. \u003cem\u003eJournal of the Royal Statistical Society. Series B (Methodological)\u003c/em\u003e \u003cstrong\u003e57\u003c/strong\u003e, 289-300 (1995). https://doi.org:10.1111/j.2517-6161.1995.tb02031.x\u003c/li\u003e\n\u003cli\u003eFerdous, T.\u003cem\u003e et al.\u003c/em\u003e The rise to power of the microbiome: power and sample size calculation for microbiome studies. \u003cem\u003eMucosal Immunology\u003c/em\u003e \u003cstrong\u003e15\u003c/strong\u003e, 1060-1070 (2022). https://doi.org:10.1038/s41385-022-00548-1 \u003c/li\u003e\n\u003cli\u003eCoutinho, F. H.\u003cem\u003e et al.\u003c/em\u003e Marine viruses discovered via metagenomics shed light on viral strategies throughout the oceans. \u003cem\u003eNature Communications\u003c/em\u003e \u003cstrong\u003e8\u003c/strong\u003e, 15955 (2017). https://doi.org:10.1038/ncomms15955\u003c/li\u003e\n\u003cli\u003eMayers, K. M. J.\u003cem\u003e et al.\u003c/em\u003e Grazing on Marine Viruses and Its Biogeochemical Implications. \u003cem\u003emBio\u003c/em\u003e \u003cstrong\u003e14\u003c/strong\u003e, e01921-01921 (2023). https://doi.org:10.1128/mbio.01921-21\u003c/li\u003e\n\u003cli\u003eCalayag, A. M.\u003cem\u003e et al.\u003c/em\u003e Arctic Ocean virus communities and their seasonality, bipolarity, and prokaryotic associations. \u003cem\u003eNature Communications\u003c/em\u003e \u003cstrong\u003e16\u003c/strong\u003e, 6427 (2025). https://doi.org:10.1038/s41467-025-61568-6\u003c/li\u003e\n\u003cli\u003eSnow, M.\u003cem\u003e et al.\u003c/em\u003e Development, application and validation of a Taqman real-time RT-PCR assay for the detection of infectious salmon anaemia virus (ISAV) in Atlantic salmon (\u003cem\u003eSalmo salar\u003c/em\u003e). \u003cem\u003eDevelopments in Biologicals\u003c/em\u003e \u003cstrong\u003e126\u003c/strong\u003e, 133-145; discussion 325-136 (2006). \u003c/li\u003e\n\u003cli\u003ePalacios, G.\u003cem\u003e et al.\u003c/em\u003e Heart and Skeletal Muscle Inflammation of Farmed Salmon Is Associated with Infection with a Novel Reovirus. \u003cem\u003ePLOS One\u003c/em\u003e \u003cstrong\u003e5\u003c/strong\u003e (2010). https://doi.org:10.1371/journal.pone.0011487\u003c/li\u003e\n\u003cli\u003eL\u0026oslash;voll, M.\u003cem\u003e et al.\u003c/em\u003e A novel totivirus and piscine reovirus (PRV) in Atlantic salmon (\u003cem\u003eSalmo salar\u003c/em\u003e) with cardiomyopathy syndrome (CMS). \u003cem\u003eVirol J\u003c/em\u003e \u003cstrong\u003e7\u003c/strong\u003e, 309 (2010). https://doi.org:10.1186/1743-422X-7-309\u003c/li\u003e\n\u003cli\u003eHodneland, K. \u0026amp; Endresen, C. Sensitive and specific detection of \u003cem\u003eSalmonid alphavirus\u003c/em\u003e using real-time PCR (TaqMan\u0026reg;). \u003cem\u003eJ Virol Methods\u003c/em\u003e \u003cstrong\u003e131\u003c/strong\u003e, 184-192 (2006). https://doi.org:10.1016/j.jviromet.2005.08.012\u003c/li\u003e\n\u003cli\u003e\u0026Oslash;rpetveit, I.\u003cem\u003e et al.\u003c/em\u003e Detection of infectious pancreatic necrosis virus in subclinically infected Atlantic salmon by virus isolation in cell culture or real-time reverse transcription polymerase chain reaction: influence of sample preservation and storage. \u003cem\u003eJournal of Veterinary Diagnostic Investigation\u003c/em\u003e \u003cstrong\u003e22\u003c/strong\u003e, 886-895 (2010). https://doi.org:10.1177/10406387100220060\u003c/li\u003e\n\u003cli\u003eRiborg, A.\u003cem\u003e et al.\u003c/em\u003e qPCR screening for \u003cem\u003eYersinia ruckeri\u003c/em\u003e clonal complex 1 against a background of putatively avirulent strains in Norwegian aquaculture. \u003cem\u003eJournal of Fish Diseases\u003c/em\u003e \u003cstrong\u003e45\u003c/strong\u003e, 1211-1224 (2022). https://doi.org:10.1111/jfd.13656\u003c/li\u003e\n\u003cli\u003eLazado, C. C., Breiland, M. W., Furtado, F., Burgerhout, E. \u0026amp; Strand, D. The circulating plasma metabolome of \u003cem\u003eNeoparamoeba perurans\u003c/em\u003e-infected Atlantic salmon (\u003cem\u003eSalmo salar\u003c/em\u003e). \u003cem\u003eMicrobial Pathogenesis\u003c/em\u003e \u003cstrong\u003e166\u003c/strong\u003e, 105553 (2022). https://doi.org:10.1016/j.micpath.2022.105553\u003c/li\u003e\n\u003cli\u003eNorwegianVeterinaryInstitute/Nepal: Release of the update Nepal pipeline v. v1.0.0 (Zenodo, 2024).\u003c/li\u003e\n\u003cli\u003eKruskal, W. H. \u0026amp; Wallis, W. A. Use of ranks in one-criterion variance analysis. \u003cem\u003eJournal of the American Statistical Association\u003c/em\u003e \u003cstrong\u003e47\u003c/strong\u003e, 583-621 (1952). https://doi.org:10.1080/01621459.1952.10483441\u003c/li\u003e\n\u003cli\u003eEwels, P. A.\u003cem\u003e et al.\u003c/em\u003e The nf-core framework for community-curated bioinformatics pipelines. \u003cem\u003eNature Biotechnology\u003c/em\u003e \u003cstrong\u003e38\u003c/strong\u003e, 276-278 (2020). https://doi.org:10.1038/s41587-020-0439-x\u003c/li\u003e\n\u003cli\u003eWood, D. E. \u0026amp; Salzberg, S. L. Kraken: ultrafast metagenomic sequence classification using exact alignments. \u003cem\u003eGenome Biol\u003c/em\u003e \u003cstrong\u003e15\u003c/strong\u003e, R46 (2014). https://doi.org:10.1186/gb-2014-15-3-r46\u003c/li\u003e\n\u003cli\u003eTeam, R. C. R: A language and environment for statistical computing. R Foundation for Statistical Computing, Vienna, Austria. \u003cem\u003ehttps://www.R-project.org/\u003c/em\u003e (2016). \u003c/li\u003e\n\u003cli\u003eTeam, R. S. RStudio: integrated development for R. \u003cem\u003ehttps://posit.co/download/rstudio-desktop/\u003c/em\u003e (2024). \u003c/li\u003e\n\u003cli\u003eAnderson, M. J. A new method for non‐parametric multivariate analysis of variance. \u003cem\u003eAustral Ecology\u003c/em\u003e \u003cstrong\u003e26\u003c/strong\u003e, 32-46 (2001). https://doi.org:10.1111/j.1442-9993.2001.01070.pp.x\u003c/li\u003e\n\u003cli\u003eAnderson, M. J. Distance-based tests for homogeneity of multivariate dispersions. \u003cem\u003eBiometrics\u003c/em\u003e \u003cstrong\u003e62\u003c/strong\u003e, 245-253 (2006). https://doi.org:10.1111/j.1541-0420.2005.00440.x\u003c/li\u003e\n\u003cli\u003evegan: Community Ecology Package. R package version 2.8-0 (2025).\u003c/li\u003e\n\u003cli\u003eFernandes, A. D.\u003cem\u003e et al.\u003c/em\u003e Unifying the analysis of high-throughput sequencing datasets: characterizing RNA-seq, 16S rRNA gene sequencing and selective growth experiments by compositional data analysis. \u003cem\u003eMicrobiome\u003c/em\u003e \u003cstrong\u003e2\u003c/strong\u003e, 15 (2014). https://doi.org:10.1186/2049-2618-2-15\u003c/li\u003e\n\u003cli\u003eLove, M. I., Huber, W. \u0026amp; Anders, S. Moderated estimation of fold change and dispersion for RNA-seq data with DESeq2. \u003cem\u003eGenome Biol\u003c/em\u003e \u003cstrong\u003e15\u003c/strong\u003e, 550 (2014). https://doi.org:10.1186/s13059-014-0550-8\u003c/li\u003e\n\u003cli\u003eMillstein, J.\u003cem\u003e et al.\u003c/em\u003e fdrci: FDR confidence interval selection and adjustment for large-scale hypothesis testing. \u003cem\u003eBioinformatics Advances\u003c/em\u003e \u003cstrong\u003e2\u003c/strong\u003e, vbac047 (2022). https://doi.org:10.1093/bioadv/vbac047 \u003c/li\u003e\n\u003c/ol\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":true,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":false,"hideJournal":false,"highlight":"","institution":"","isAcceptedByJournal":true,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"scientific-reports","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"scirep","sideBox":"Learn more about [Scientific Reports](http://www.nature.com/srep/)","snPcode":"","submissionUrl":"","title":"Scientific Reports","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"stoa","reportingPortfolio":"Scientific Reports","inReviewEnabled":true,"inReviewRevisionsEnabled":true},"keywords":"Atlantic salmon pathogens, environmental DNA, non-invasive sampling, shotgun metagenomic sequencing","lastPublishedDoi":"10.21203/rs.3.rs-8373998/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-8373998/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003eSpecific tools for detecting waterborne pathogens are essential for limiting disease spread in aquaculture. We evaluated a field-deployable workflow combining filtration of eDNA/eRNA with targeted (RT-)qPCR and complementary shotgun metagenomics to monitor pathogens and microbial community dynamics across an Atlantic salmon production cycle, from hatchery to offshore cages. The primary aim was to assess workflow feasibility and performance under real farm conditions, while secondarily examining whether metagenomic profiles could contextualise microbial shifts associated with pathogen presence.\u003c/p\u003e \u003cp\u003eISAV was consistently detected in hatchery water at ~\u0026thinsp;4\u0026times;10\u0026sup3;\u0026ndash;9\u0026times;10\u0026sup3; copies per litre, whereas PRV1 was detected only inside sea pens from August onward (~\u0026thinsp;4\u0026times;10\u0026sup2;\u0026ndash;1.5\u0026times;10⁴ copies per litre) and increased by more than two orders of magnitude after wellboat delousing. Shotgun metagenomics yielded a median of ~\u0026thinsp;1.5\u0026times;10⁵ reads per sample (mean read length\u0026thinsp;~\u0026thinsp;2.5 kb; N50\u0026thinsp;\u0026gt;\u0026thinsp;2 kb), enabling broad taxonomic screening. PRV1-positive seawater samples showed modest decreases in richness and shifts in viral taxa, though patterns were subtle and should be interpreted cautiously given low pathogen loads.\u003c/p\u003e \u003cp\u003eThe workflow was practical for trained farm personnel, and this integrated approach offers a scalable system for routine pathogen surveillance and supports earlier, evidence-based biosecurity actions, providing broader microbial information than qPCR alone.\u003c/p\u003e","manuscriptTitle":"Integrated approaches for pathogen monitoring and shotgun metagenomic analysis in Atlantic salmon farming","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2026-02-19 21:36:55","doi":"10.21203/rs.3.rs-8373998/v1","editorialEvents":[{"type":"communityComments","content":0},{"type":"decision","content":"Revision requested","date":"2026-03-19T03:30:26+00:00","index":"","fulltext":""},{"type":"editorInvitedReview","content":"","date":"2026-03-08T14:08:56+00:00","index":"hide","fulltext":""},{"type":"editorInvitedReview","content":"","date":"2026-03-08T09:27:51+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"33178629617108030236117231846851177758","date":"2026-02-16T23:07:37+00:00","index":"hide","fulltext":""},{"type":"editorInvitedReview","content":"","date":"2026-02-15T20:30:55+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"102981814605918374582067647051821383651","date":"2026-02-15T11:45:00+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"166151953924345566130642542196267566524","date":"2026-02-13T19:36:10+00:00","index":"hide","fulltext":""},{"type":"reviewersInvited","content":"","date":"2026-02-13T10:32:11+00:00","index":"","fulltext":""},{"type":"checksComplete","content":"","date":"2026-02-12T09:50:21+00:00","index":"","fulltext":""},{"type":"submitted","content":"Scientific Reports","date":"2026-02-12T08:28:57+00:00","index":"","fulltext":""}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"scientific-reports","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"scirep","sideBox":"Learn more about [Scientific Reports](http://www.nature.com/srep/)","snPcode":"","submissionUrl":"","title":"Scientific Reports","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"stoa","reportingPortfolio":"Scientific Reports","inReviewEnabled":true,"inReviewRevisionsEnabled":true}}],"origin":"","ownerIdentity":"7a078458-8380-4a2c-9bda-96a83e41af4b","owner":[],"postedDate":"February 19th, 2026","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"published-in-journal","subjectAreas":[{"id":63035417,"name":"Biological sciences/Biological techniques"},{"id":63035418,"name":"Biological sciences/Biotechnology"},{"id":63035419,"name":"Earth and environmental sciences/Environmental sciences"},{"id":63035420,"name":"Biological sciences/Microbiology"}],"tags":[],"updatedAt":"2026-04-20T16:15:43+00:00","versionOfRecord":{"articleIdentity":"rs-8373998","link":"https://doi.org/10.1038/s41598-026-48791-x","journal":{"identity":"scientific-reports","isVorOnly":false,"title":"Scientific Reports"},"publishedOn":"2026-04-14 15:57:21","publishedOnDateReadable":"April 14th, 2026"},"versionCreatedAt":"2026-02-19 21:36:55","video":"","vorDoi":"10.1038/s41598-026-48791-x","vorDoiUrl":"https://doi.org/10.1038/s41598-026-48791-x","workflowStages":[]},"version":"v1","identity":"rs-8373998","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-8373998","identity":"rs-8373998","version":["v1"]},"buildId":"XKTyCvWXoU3ODBz1xrDgd","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}