Recirculating aquaculture systems (RAS), often used in fish farming, rely on microorganisms to maintain healthy water quality, nutrient cycling, animal welfare, and disease control. However, many daily operations in fish farms (e.g., stocking) may negatively affect the microorganisms' community composition and create a favorable environment for opportunistic pathogens. Currently, understanding microorganisms' communities within RAS is scarce, which presents an obstacle for pro-active system management. To better understand microorganism communities' spatial and temporal structure within fish farms using a RAS, we collected samples of filtered water and biofilm swabs from two different Swiss fish farms and two different locations within each farm. Water was collected from within one tank and the biofilter, while biofilm swabs were collected from the same tank's wall where the water sample was collected. DNA was extracted using the Purelink Microbial DNA Purification kit, and then each sample was prepared with three different primer pairs, 341F/805R (V3V4 region), 515F/806R (V4), 27F/534R (V1-3), and ran on the MiSeq platform (v3 600 cycles). The pilot study aimed to understand 1) how the microbiota composition changes regarding spatial and temporal scales within and between farms, 2) the primer effect on detected taxa, and 3) the difference between commonly-used 16s pipelines.

No abstract

Recirculating aquaculture systems (RAS) heavily depend on microbial communities to maintain water quality. These communities therefore influence the growth, development, and welfare of farmed fish. With the increasing socio-economic role of fish farming e.g. regarding food security, an in-depth understanding of aquaculture microbial communities is also relevant from a management perspective. However, the data situation regarding the composition of microbial communities within RAS is patchy. Since this is partly ascribed to method choices, there clearly is a need for accurate, standardized, and user-friendly methods to study microbial communities in aquaculture systems. Here, we compare the performance of 16S amplicon sequencing, Pac-Bio long-read amplicon sequencing, and amplification-free shotgun metagenomics in the characterization of microbial communities in two commercial-size RAS fish farms. We show that, even though primer choice affects read quality, diversity, and assigned taxa, distinct primer pairs uncover similar spatio-temporal patterns between sample types, farms, and time points. We find that long-read amplicons underperform regarding quantitative resolution of spatio-temporal patterns, but allow for species-level identification of functional services and pathogens. Finally, shotgun metagenomics data identified fungi, viruses, and bacteriophages, opening avenues for an exploration of natural approaches regarding antipathogenic treatments. Overall, the datasets agreed on major prokaryotic players. In conclusion, different sequencing approaches yield overlapping and highly complementary results, with each contributing data no other approach could. Such a tiered approach therefore constitutes a practical and cost-effective strategy for obtaining the maximum amount of information on aquaculture microbial communities. These data could lead to better farm management practices and at the same time inform basic research on community evolution dynamics.