Introduction
Salmon aquaculture loses c.20% of annual production as a result of gill disease triggered by plankton organisms, including jellyfish and harmful algal blooms - as such, health challenges from the plankton threaten the sustainability of the European salmonid aquaculture industry. Current methods of plankton monitoring lack sensitivity and specificity, can be time consuming, and require taxonomic expertise. For these reasons, data generated fail to provide actionable and timely information to aquaculture managers. We have previously provided proof-of-principle that environmental DNA (eDNA) metabarcoding may be deployed to identify and quantify multiple planktonic threats of salmon farms. In the current study we present a seven-month longitudinal planktonic survey of salmon aquaculture on the west coast of Scotland (UK), covering both spring and autumn blooms when the periods of highest mortalities in the farms typically occur. This dataset allowed us to assess biodiversity and abundance from DNA-based data in comparison with plankton morphological identification, and to identify planktonic threats for salmon health.
Materials & Methods
Metabarcoding, plankton morphology and fish health were readily combined to identify key drivers of salmon gill health. Two sea cage sites (Site A (exposed), Site B (sheltered)) were monitored for c. 200 days (March - October 2021) by daily eDNA, phyto- and zooplankton samplings. Prior to both morphological and molecular analyses, a hit list of species of interest was compiled based on previously published literature and personal communications with the farms in order to record target planktonic taxa causing salmon gill disease. This list would allow us to inform our industrial partners of our preliminary findings in case of a bloom, and to perform quality control on our species database for metabarcoding analyses. Using GLMs and morphological species data, we have so far been able to investigate potential effects of species abundances on the incidence of AGD and PGD scores and mortality. We first tested these relationships assuming no temporal lag between species dynamics and fish mortalities, as well as assuming that mortality lagged 1, 2 and 3 and 7 days behind species dynamics. Preliminary eDNA data indicate that the method is sensitive to detect both phyto- and zooplankton species of interest and that some species might have been misclassified via the microscopy method. Next steps using the full eDNA dataset will aim to benchmark the molecular method against the microscopy data and, combined with abiotic variables (e.g., temperature, salinity, turbidity), to assay their ability to predict salmon mortality and gill disease (AGD, PGD).
Preliminary results
Up to 96 phytoplankton and 74 zooplankton taxa were recorded after morphological identification analyses, including most of the target species from our hit list. The dynamics of the hit species as well as of the most abundant plankton groups are shown in Fig 1, where abundance of taxa is represented in a logarithmic scale. Regarding phytoplankton, the two sites presented similar species dynamics; the HAB genus Pseudo-nitzschia was present on both sites and was abundant from April onwards whereas a March Skeletonema bloom was only observed in site B. Regarding zooplankton, total abundances, as well as bivalves and ophiura larvae were higher in site B. The most abundant hydromedusae species from our target list were Obelia sp. and Lizzia blondina. The latter showed higher abundances during late summer, same as for doliolids (only for Site A) and Oikopleura.dioica (more clearly observed in Site B).
Preliminary data (Table 1) shows the potential of eDNA metabarcoding in detecting planktonic organisms at the species level when morphological analyses is only reaching to genus level. Moreover, our eDNA analyses will detect benthic life stages from meroplanktonic target species, allowing for a thorough analysis of the potential threats for salmon health that are in the water off the pens. We expect that our upcoming sequencing data will include many more taxa than those recorded by microscopy, hence improving our initial hit list. Lagged statistical modelling will enable the detection of species that have a significant immediate, delayed and cumulative impacts on fish health
Outcome
Our project will augment our proof-of-principle that environmental DNA monitoring can be used to detect planktonic threats (algae, jellyfish, amoebozoans) to salmonid aquaculture and potentially translate this into a low-cost on-site quantitative test for key drivers of Complex Gill Disease (CGD). Rapid and early identification of the planktonic threats will thus enable salmon farmers to take effective, timely steps to mitigate their impact on finfish. We hope our technology will provide a step change in mitigating losses caused by planktonic threats, improving both the sustainability and productivity of the salmonid aquaculture industry.