Nanopore-Based Metagenomic Approaches for Detection of Bacterial Pathogens in Recirculating Aquaculture Systems

Abstract

The microbial community in a recirculating aquaculture system (RAS) is pivotal in fish health, contributing significantly to the productive performance during the growing-out phase. Classical and molecular methods using PCR for species-specific amplifications have traditionally been used for bacterial community surveillance. Unfortunately, these approaches mask the real bacterial diversity and abundance, population dynamics, and prevalence of pathogenic bacteria. In this study, we explored the use of Oxford Nanopore Technology to characterize the microbiota and functional metagenomics in a commercial freshwater RAS. Intestine samples from Atlantic salmon (Salmo salar (85 ± 5.7 g)) and water samples from the inlet/outlet water, settling tank, and biofilters were collected. The full-length 16S rRNA gene was sequenced to reconstruct the microbial community, and bioinformatic tools were applied to estimate the functional potential in the RAS and fish microbiota. The analysis showed that bacteria involved in denitrification processes were found in water samples, as well as metabolic pathways related to hydrogen sulfide metabolism. Observations suggested that fish classified as sick exhibited decreased microbial diversity compared with fish without clinical symptomatology (p < 0.05). Proteobacteria were predominant in ill fish, and pathogens of the genera Aeromonas, Aliivibrio, and Vibrio were detected in all intestinal samples. Notably, Aliivibrio wodanis was detected in fish showing abnormal clinical conditions. Healthy salmon showed higher contributions of pathways related to amino acid metabolism and short-chain fatty acid fermentation (p < 0.05), which may indicate more favorable fish conditions. These findings suggest the utility of nanopore sequencing methods in assessing the microbial community in RASs for salmon aquaculture.

1. Introduction

The aquaculture industry has been challenged to develop sustainable systems for environmental impact issues such as water availability, habitat destruction, parasite transmission, and others [1,2]. Recirculating aquatic systems (RASs) have emerged as a solution to overcome the limitations of conventional systems, both in freshwater and saltwater [1]. These systems can be used to implement techniques and parameters to increase production, reduce environmental impact, and increase adaptation to climate change [2,3,4], while minimizing the risk for opportunistic fish pathogens [5]. Salmon farming has become one of Chile’s main economic activities, making it the second-largest world producer of salmonids. However, several sanitary challenges hinder the sustainable development of the sector [6]. Here, the implementation of RASs can help to improve salmon farming conditions in Chile.

Microorganisms play a fundamental role in an RAS, both in maintaining the system’s proper functioning and in the health of the farmed fish [7,8,9]. Understanding the microbial communities that compose the microbiota in an RAS, as well as the factors that influence their structure, may help to obtain a better understanding of the performance and health of the cultivated organisms in these systems. Studying various system components, such as tank biofilms, biofilters, and the water column, has become essential in analyzing RAS microbiomes. Studies indicate that the phyla Proteobacteria, Bacteroidetes, and Actinobacteria dominate water samples in Atlantic salmon RASs, while the dominant genera vary [10,11,12]. These studies have also evaluated the risk of infectious diseases in RASs [13]. Given the densities in RASs, the appropriate monitoring of the sanitary conditions of the fish is critical, since diseases can spread rapidly during an outbreak [2]. Potentially pathogenic bacteria of the genera Vibrio, Erwinia, Coxiella, and Aeromonas, among others, have been found in RASs [1]. The study and modulation of the microbiome arise as an element to predict and improve the health status of farmed fish [10,14]. The evidence suggests that mucus environments, such as the intestines, are critical components of the immune response and, thus, a point of particular interest in studying microbial dysbiosis in aquaculture systems [15,16].

Long-read sequencing technologies applied to genetic marker profiling, such as bacterial 16S RNA, have improved the taxonomic resolution of bacterial communities up to the species level [17]. Together with the improvement of bioinformatic algorithms, sequencing the complete 16S gene allows for predictions of the functional potential of microbial communities [18,19], which has been previously used to study the effect of a fish farm in its ecological niche [20]. The evaluation of specific functional pathways related to the production of toxic inorganic compounds is particularly relevant in an RAS, such as monitoring the production of hydrogen sulfide and nitrogen removal, which are indispensable for proper RAS functioning [21,22]. Moreover, microbiota sequencing can also be applied to detect potential pathogenic agents that might impact fish health during the saltwater transfer phase.

Several different factors have been described that may influence the microbial communities of RASs, such as water recirculation rates, UV treatments, and the type of diet [23,24,25]. However, little is known about the microbiota of RASs and fish under commercial farming conditions, where pathogens might arise during the culture cycle. Thus, this study aimed to characterize the microbiota dynamics of a commercial RAS that evidenced an outbreak from an unknown etiological agent. Our results evidence that 16S sequencing is a valuable approach to assess the health status of both RAS components and fish within them.

2. Materials and Methods

2.1. Ethical Statement

This study did not involve any experimental procedures on live animals beyond standard husbandry and health-monitoring practices. Sampling was performed by trained personnel from the MOWI company, with the approval of their Animal Welfare Committee, ensuring minimal stress and discomfort to the fish. No additional ethical approval was required, as samples were collected during routine operations in a commercial facility.

2.2. Sampling

A mortality event from an unknown etiological agent was reported in May 2021 in a commercial RAS in Puerto Montt, Chile, after 35 days post-vaccination with an attenuated vaccine against Salmon Rickettsial Septicemia (SRS). Thus, to evaluate the health status of this RAS and the fish, complete intestine and pyloric caecum samples from Atlantic salmon (Salmo salar (85 ± 5.7 g)) with abdominal distention (BL) and without any clinical sign (UnBL) of abdominal distension were collected in 95% ethanol, 35 days post-vaccination. All fish were maintained under the following water parameters: an O₂ level of 11.93 mg/L, a salinity of 4.58 ppt, a temperature of 14.3 °C, a pH of 6.8, a TAN level of 0.48 mg/L, a 106 NO₂⁻ level of 0.69 mg/L, a NO₃⁻ level of 101.2 mg/L, and an alkalinity level of 92.73 mg/L. Additionally, water samples were collected from a freshwater input called the inlet water (In.water); the sector where solids were decanted from the culture tanks, called the settling tank (S.Tank); and two sectors in the biofilter of the recirculation system, one at the entrance to the inlet biofilter (In.biofilter) and one at the outlet biofilter (O.biofilter). The water samples were passed through 0.45 µm filters, and the filters were stored in 95% ethanol. All samples were transported on dry ice and analyzed as soon as they were received at the same time.

2.3. DNA Extraction and Full 16S rRNA Amplification

The fish intestine DNA without feces was extracted using a DNAeasy Blood & Tissue Kit (Qiagen, Germantown, MD, USA) according to the manufacturer’s instructions. The DNA in the water samples was extracted from the 0.45 µm filters using a Quiamp Fast DNA Stool (Qiagen, USA) Kit according to the manufacturer’s instructions. The DNA concentration was measured with a Nanodrop spectrophotometer. Three water samples per site and three intestine fish samples per condition were pooled in a final 100 ng/µL concentration. The full 16S rRNA gene was amplified using the primers 27 F 5′-AGAGTTTGATCCTGGCTCAG-3′ and 1492 R 5′-GGTTACCTTGTTACGACTT-3′ [26]. Taq DNA polymerase LongAmp (New England Biolabs, Ipswich, MA, USA), with a reaction volume of 15 µL, was used for amplification under the following conditions: 95 °C for 1 min, followed by 30 cycles at 95 °C for 20 sec, 56 °C for 30 sec, and 65 °C for 1 min, ending with a final extension at 65 °C for 5 min The effectiveness of the PCR was evaluated using 1% agarose gel electrophoresis.

2.4. Nanopore Library Synthesis and Sequencing

PCR products were purified in a 1:2 sample-to-bead ratio using Agencourt AMPure XP beads (Beckman Coulter, Brea, CA, USA) for Nanopore library synthesis. The samples were placed in the magnetic stand for washing with freshly prepared 70% ethanol. The washed product was separated from the beads by placing it in a magnetic stand after rinsing it in ultrapure sterile water. The purified amplicon was quantified in the Qubit 4 fluorometer (ThermoScientific, Waltham, MA, USA) and used as a template for library synthesis with the 16S Barcoding Kit (SQK-RAB204, Oxford Nanopore Technologies, Oxford, UK), according to the manufacturer’s instructions for a 1D sequencing strategy. Briefly, the purified 16S amplicon was mixed with barcodes. Then, a PCR was performed as in the previous step, with a final reaction volume of 50 µL, using the LongAmp Taq Polymerase (New England Biolabs). The PCR product was incubated for 5 min at room temperature in a HulaMixer (ThermoScientific, USA), and cleaning was carried out by washing with Agencourt AMPure XP beads, according to the manufacturer’s instructions. After washing the magnetic beads, the purified product was eluted in 10 µL of an elution buffer (10 mM Tris-HCl (pH 8.0) with 50 mM NaCl). The final concentration of the library was measured using a Qubit 4 fluorometer (Thermoscientific, USA), and the libraries were tested using High-Sensitivity D5000 ScreenTape (Agilent, Wilmington, DE, USA), according to the manufacturer’s instructions on a TapeStation Bioanalyzer 2100 (Agilent, USA). Following Oxford Nanopore Technologies’ methodology, the libraries were pooled in multiplex mode and put into the MK1 Spot-ON FLO-MIN107-R9 flowcell. As an internal control, the DNA of a microbial mock community (ZymoBiomics Microbial Community Standard) was extracted and sequenced following the same procedures, and the observed abundances of taxa were compared with their expected abundance. Sequencing efficiency was monitored using the software MinKNOW 2.0 (Oxford Nanopore Technologies). Raw sequencing reads were uploaded to SRA under the BioProject ID PRJNA1182272, and the barcode information for each sample is provided in Supplementary Table S1.

2.5. Bioinformatic Analysis

The generated fast5 files were base-called using Guppy (version 3.2.2, Oxford Nanopore Technologies, UK), and a filter step was applied to retain only sequences with a Q-score ≥ 7 (quality filter). The demultiplexing primers and barcode trimming were performed with Porechop [27]. A filter-by-sequence size of 1000–1800 bp and Q-score ≥ 8 was applied using Nanofilt (v2.8.0) [28]. Each amplicon sequencing variant (ASV) was determined in the Epi2Me platform (Oxford Nanopore Technologies, UK) using the 16S rRNA NCBI database. The lca = −1 (unsuccessful classification) results were discarded from the analysis. The resulting data were used to calculate alpha and beta diversity indices, calculated in the R software (v4.4.1) using the “vegan” R package (v 2.7-1) [29,30] and plotted using the “ggplot2” R package [31]. The taxonomic lineage was queried by the TaxonKit pipeline [32]. For variations of the intestinal microbiota exploration, the relative abundance per sample was calculated in R. For visualization, the data were plotted by phyla, genera, and species using the “ggplot2” R package. The core microbiota was identified with a prevalence cut-off of 80% and a lower relative abundance limit of 0.05%. The Log2 relative abundance by taxon across samples was plotted like a heatmap with the “Pheatmap” R package. Finally, potential fish pathogens were identified from a list of species reported by Austin and Austin [33] (Table S2).

2.6. Functional Prediction

The functional pathways for the entire microbiota were predicted from the 16S rRNA amplicon with the PICRUSt2 software (v2.5.2) [34]. These regions were extracted using the Hyperex pipeline with V3-V4 default regions. The pathway levels were built using the MetaCyc database [35]. For data visualization, the relative abundance per sample was calculated and plotted using the “ggplot” R package. We used the STAMP 2.1.3 program to evaluate the pathway contributions and comparison plots [36]. A two-sided G test corrected by Benjamini–Hochberg’s false discovery rate was applied. The differences in proportions (DPs) and confidence intervals (CIs) were calculated using the Newcombe–Wilson method. Multiple test corrections were performed using the Storey q-value [37]. Data with a minimum relative abundance of 0.5 were used for graph building to determine the proportion difference. Nitrogen and sulfur metabolism analyses of water samples were performed in the R package. A relative abundance plot was generated using the Inorganic Nutrient Metabolism MetaCYC superclass, including OTUs’ contribution by pathway. A map of the cellular function of nitrogen and sulfur metabolism and the KEGG pathway contribution of water samples [38] was built on the KEGG Mapper platform [39,40]. Finally, the KOs of virulence factors were evaluated from the database developed by Pattaroni et al. [41].

3. Results

This study aimed to characterize the microbiota dynamics of a commercial RAS that evidenced an outbreak from an unknown etiological agent. The initial clinical report indicated that the bloated fish (30%) had different degrees of peritonitis (Figure S1A), while the histological exam showed severe multifocal granulomatous polymorphonuclear inflammation of pyloric caeca with central necrosis (Figure S1B). Infection by Rhodococcus sp. was suspected based on local epidemiology and the staining affinity of the Gram stain (Figure S1C); however, the PCR results were negative for Rhodococcus sp. and Francisella sp. Septate hyphae suggested Exophiala salmonis in the gills, pancreas, and kidney (Figure S1D).

3.1. Microbiota Diversity Indexes

In the sequencing analysis, 1,118,214 reads passed the Guppy, demultiplex, and trimming filters; 89,743 passed the size filters and were used in taxonomic classification with Epi2me. A total of 95% of the reads were classified; 62% of these had lca = 0, and 38% had lca = 1. Those sequences with lca = 1 (if the top three classifications represent more than one genus) were only assigned to a family (Table S3). A total of 746 taxa were identified in water and intestine samples (Table S4). Regarding diversity analysis, the rarefaction curves indicated that all samples reached a plateau, and the lowest number of observed taxa was found in the intestines of fish with bloating, and the largest was in the settling tank (Figure 1A). The principal coordinates analysis using Bray–Curtis-transformed data placed the settling tank, inlet biofilter, and outlet biofilter pooled samples in a cluster. In contrast, the other pooled samples were independent (Figure 1B). However, the fish intestine samples with and without abdominal distension (BL, UnBL1, and UnBL2) were closer together than the other samples (Figure 1B). The settling tank (S.tank) and biofilter outlet (O.Biofilter) had the highest diversity, with Simpson’s index values of 0.98 and 0.99 and Shannon’s H’ values of 5.05 and 5.12 (Figure 1C,D). The intestines of fish with bloating (BL) had the lowest values of the diversity indices (with a Simpson’s value of 0.58 and a Shannon’s H’ value of 1.93) and Pielou’s evenness (Figure 1E).

3.2. Potential Pathogens in RAS Microbiota

Fourteen phyla were identified in the samples analyzed; we considered those with a relative abundance >0.5%. From 57% to 98% of all models were of the phylum Proteobacteria (Figure S2A), followed by Firmicutes (1.1–35.42%) and Bacteroidetes (1.4–25.5%). Cyanobacteria and Planctomycetes were only present in the intake, while Fusobacteria and Verrucomicrobia were only in the water samples. The intestinal microbiota of fish had the lowest richness of phyla; those with abdominal distension had only Proteobacteria and Bacteroidetes (98% and 2%, respectively).

The family Arcobacteraceae and the genera Clostridium and Aliivibrio were the most important in the intestinal microbiota of bloated fish (Figure S2B). A general analysis of the microbial core community showed that the phyla Proteobacteria and Bacteroidetes were representative in all samples (Figure 2). Some fish pathogen genera were identified as Shewanella, Flavobacterium, Acinetobacter, Pseudomonas, Vibrio, and Aeromonas (Figure 2). A more specific analysis showed that the water samples had fewer potentially pathogenic microorganisms than the fish intestine samples (Figure 3A); the relative abundance of potential pathogens in water samples was less than 1.7%, while all fish intestine samples had at least 15%. Bacteria of the genus Aeromonas were present in all samples; Aeromonas salmonicida and A. hydrophila were most abundant in one of the intestinal samples of fish without abdominal distension. In fish with abdominal distension, Aliivibrio wodanis replaced almost all the pathogens present in other samples; they represented 64.01% of the total relative abundance of microorganisms. However, this was only found in bloated fish and the biofilter outlet. Flavobacterium succinicans was present in all water samples of the interior circuit of the RAS. Salmonella enterica was found in the inlet but not in the other samples; Shewanella putrefaciens was detected in several samples of water and intestines. Species of the genus Vibrio, particularly V. anguillarum, were found in all intestinal samples (Figure 3A). The functional prediction of the pathogen-related microbiome evidenced 546 KOs associated with virulence factors. All samples had KO K03088, related to antiphagocytosis, sigma factor, and iron capture, especially the water samples. Another KO series related to iron capture (K0201, K02015, K02013, K06147, and K02014) was found in all samples (Figure 3B).

3.3. Functional Role of RAS Microbiota

Most species had a relative abundance below 10% (Figure S2C). There were some exceptions. In the inlet, 26.6% were Streptococcus thermophilus, while Sulfurimonas denitrificans were abundant in the settling tank and outlet biofilter. Clostridium chromiireducens had an abundance of 13.97% in one of the healthy fish groups (Figure S2C). The functional prediction analysis identified 398 pathways represented, which were reduced to 52 in level-two functional pathways and 6 in level one (Table S5; Figure S3). Seventeen of the functional pathways found in samples of fish intestinal microbiota had a significant difference in proportions of greater than 0.2%. Functional pathways related to biosynthesis significantly contributed, particularly the biosynthesis of fatty acids and lipids in bloated fish (Figure 4).

The second most significant difference in proportions in functional pathways of level sec was fermentation (generation of precursor metabolites and energy pathway); healthy fish made the most considerable contribution to the microbiota. Fish with abdominal distension contributed more significantly to the functional pathways related to cell replication, such as cell structure biosynthesis, than fish without distension. Although aromatic compound degradation contributed less to the metagenome, the difference between proportions significantly benefited healthy fish. By contrast, although amino acid biosynthesis was the third most-represented pathway, it showed a minor difference between proportions. Twenty-six pathways with a relative abundance of greater than 0.5% were identified in the water samples (Figure S4). However, fewer showed proportional differences more significant than 0.2%, and no difference was greater than 0.8%. Like intestine samples, the most significant contribution to the microbiota was associated with biosynthesis processes.

3.4. Sulfur and Nitrogen Metabolism

Sulfur metabolism was most represented by sulfur MetaCyc and favored in the intestinal microbiota of bloated fish (Figure 5A). Sulfur reduction and assimilation for cysteine biosynthesis were particularly preferred in the intestinal microbiota of bloated fish (Figure 5A). The KEGG map indicates that the KO related to sulfur metabolism was more complete in the bloated fish (Figure S5A). Bacteria of the genus Aliivibrio contributed most to H₂S production related to the sulfate reduction I (assimilatory) in the intestine samples of bloated fish (Figure 5B). Another production pathway not present in the KEGG map or MetaCyc was the trans-sulfuration pathway represented by KOs K01697, K01758, and K01011 (Table S6). Curiously, the difference in proportions indicated that the presence of these genes in intestinal microbiota samples was more favored in the healthy fish of group 2 than in the others.

4. Discussion

Recirculating aquaculture systems (RASs) represent a promising technology for controlling and improving parameters that are not feasible in flow-through systems, offering advantages in fish health management and environmental impact reduction [15,42]. They also allow the implementation of technological innovations for their improvement, a reduction in environmental impact, and adaptation to climate change [2,3,4]. In RASs, rapid pathogen detection is essential for disease control. Compared with short-read Illumina sequencing of partial 16S regions (e.g., V3–V4), nanopore full-length 16S rRNA sequencing (V1–V9) provides a higher taxonomic resolution, enabling more reliable species-level identification rather than limiting analyses to the genus level [43]. This higher resolution helps distinguish specific pathogenic species from near-relatives, an important advantage for surveillance, although the 16S gene’s limited variability means certain closely related taxa remain indistinguishable at the strain level [44]. While Illumina platforms deliver highly accurate reads but require centralized facilities and longer turnaround times, portable nanopore devices can produce actionable results within ~24 h [45]. Moreover, ONT’s low startup cost and flexible throughput make it well-suited for routine aquaculture monitoring, complementing Illumina’s role in high-throughput microbiome profiling [43].

In this study, we explored the microbiota composition of a freshwater RAS through a nanopore-based metagenomic approach, focusing on Atlantic salmon exhibiting clinical signs of abdominal bloating and granulomatous peritonitis. Traditional PCR methods failed to identify an etiological agent, prompting us to employ full-length 16S rRNA gene sequencing using Oxford Nanopore Technology to investigate both the fish and water microbiota. Our analysis suggested a low species diversity in the intestinal microbiome of bloated fish, which is often indicative of dysbiosis and has been associated with diseases in both humans and fish [46,47]. This finding aligns with Ofek et al. [46], who reported that healthy tilapia’s microbial alpha diversity was significantly lower in unhealthy fish. In our study, the bloated fish had a high abundance (>98%) of Proteobacteria, consistent with previous reports in fish microbiomes, including Atlantic salmon [48,49,50,51,52,53,54]. Interestingly, diseased hybrid tilapia also showed higher levels of Proteobacteria than healthy fish [46], suggesting a potential association between Proteobacteria dominance and disease states in fish. Notably, we detected a high abundance of Aliivibrio, particularly A. wodanis, in the intestines of bloated fish, whereas this pathogen was absent in healthy fish. A. wodanis has been associated with winter ulcer disease in Atlantic salmon, often isolated alongside Moritella viscosa [55]. However, the clinical symptoms observed in our study were inconsistent with typical winter ulcers [56], indicating a potential variation in pathogenicity or disease manifestation. Additionally, Aeromonas spp. were present in healthy fish, which is intriguing given that Aeromonas hydrophila and A. salmonicida are known pathogens with high adhesion velocities [57]. The presence of potentially pathogenic Vibrio species in both healthy and bloated fish suggests a complex microbial environment where opportunistic pathogens may coexist with commensals. Comparatively, the water samples revealed the presence of bacteria involved in nitrogen fixation, such as Pseudomonadaceae, Rhizobiaceae, and Rhodobacteraceae, consistent with previous studies [58]. However, specific nitrogen metabolism pathways were incomplete, suggesting that these processes occur primarily in biofilter sediments rather than the water column, as previously described [38]. The detection of A. wodanis in the outlet of the biofilter (Z200) indicates potential cross-contamination between water and fish microbiota. Biofilms within the RAS can serve as reservoirs for pathogens, which may detach and cause outbreaks [59,60]. This underscores the need for vigilant microbiota surveillance in salmon farming to preempt potential disease events.

Functionally, healthy fish microbiota showed enrichment in amino acid metabolism pathways and short-chain fatty acid fermentation. These metabolic functions are crucial for fish health, contributing to energy metabolism and overall well-being [61,62,63,64]. Amino acid metabolism pathways, including those for valine, leucine, and isoleucine, play significant roles in energy metabolism [62]. The increased representation of these pathways in healthy fish aligns with reports that such metabolic activities are associated with better fish health [63,64]. Another pathway with a significant proportional difference, although low in overall contribution, was aromatic compound degradation, favoring healthy fish, which has been linked to dysbiosis events in the skin and gills [65]. In contrast, bloated fish microbiota exhibited pathways related to sulfur metabolism, specifically, hydrogen sulfide (H₂S) production. While H₂S is a normal component of fish intestinal metabolism [66,67], excessive production can have cytoprotective effects on bacteria, potentially contributing to antibiotic resistance and complicating disease treatment [68,69]. A. wodanis, identified as a primary H₂S producer in bloated fish intestines, may benefit from these protective mechanisms. Hydrogen sulfide production has become an emerging problem in RASs [58], potentially due to the accumulation of solid waste in tanks, leading to chronic low-level exposure in fish [58,70]. Contrary to expectations, filters aimed at reducing turbidity may be insufficient to eliminate H₂S-producing bacteria [50].

Many virulence factors were found in the samples; in particular, KEGG Orthology (KO) term K03088 was linked to the sigma 70 family factor (rpoE gene), a response to bacterial stress that governs survival inside macrophages and the production of siderophores promoting bacterial growth in the host [71,72,73]. Iron acquisition systems, one of the main virulence factors of fish pathogens, were also significant [74]. We identified bacteria of importance in food safety within fish intestines, such as Clostridium spp., Aeromonas hydrophila, and Plesiomonas shigelloides. Although these species are part of the normal intestinal microbiota, monitoring is essential to prevent high concentrations that could cause gastrointestinal issues in consumers [75]. Our findings both align with and differ from previous reports, highlighting the complexity of microbiota interactions in RASs. While certain pathogens like A. wodanis are typically associated with specific diseases [55,56], their role may vary under different environmental conditions or in combination with other microbial communities. Although A. wodanis was detected in association with the bloating condition, the 16S rRNA approach used here cannot establish causality or confirm the presence and expression of virulence genes; such confirmation would require complementary shotgun metagenomics or transcriptomic analyses. However, the application of full-length 16S rRNA nanopore sequencing in RAS offers a practical framework for the early detection of potential pathogens, enabling producers and regulators to implement timely interventions before disease outbreaks occur. By providing rapid, high-resolution insights into microbiome composition and highlighting the presence of opportunistic or emerging pathogens, this approach can support proactive management strategies that ultimately contribute to reducing pathogen pressure, improving fish health, and enhancing the sustainability of aquaculture operations.

5. Conclusions

Our study demonstrates the potential of nanopore sequencing as a powerful tool for pathogen detection and microbial community assessment in recirculating aquaculture systems. By combining full-length 16S rRNA profiling with functional metagenomics, we revealed distinct microbial signatures associated with fish health status and identified Aliivibrio wodanis as a pathogen putatively linked to abnormal clinical conditions such as bloating. Importantly, this approach allowed us to detect clinically relevant bacterial genera, including Aeromonas, Aliivibrio, and Vibrio, across fish and water samples, thereby providing a more comprehensive view of pathogen prevalence in aquaculture environments. Enhanced monitoring of both fish and water microbiota through real-time sequencing technologies could strengthen disease prevention strategies and support sustainable health management in aquaculture. Future studies should validate these findings with larger cohorts and robust statistical analyses to further establish nanopore sequencing as a routine method for infection surveillance and pathogen detection in fish farming systems.