eDNA from water column to characterise fish and invertebrate communities from 30 sites in the Belgian Part of the North Sea -COI

Study Extent

The construction of offshore wind farms may affect local soft-sediment fauna. Hence, an efficient monitoring technique is needed to monitor the potential effects on the marine ecosystem. Here, we assess whether eDNA metabarcoding is a suitable alternative to monitor fish and epibenthos biodiversity in these difficult to access marine habitats. Water sampling and trawl surveys were conducted in parallel in 12 coastal and 18 offshore sites, the latter located inside and outside two offshore wind farms in the Belgian part of the North Sea.

Sampling

Sampling: During two different field campaigns in September and November 2021, a total of 12 coastal and 18 offshore locations, situated inside and outside the OWFs C-power (transition zone) and Belwind (offshore zone), were samples. The coastal locations were sampled in triplicate during the September field campaign with the research vessel Simon Stevin using a Niskin carousel. The offshore locations and one coastal location (ft230) were sampled in November 2021 with the research vessel GeoOcean V. During this campaign five biological replicates were taken by successively lowering one Niskin bottle five times. One exception was the coastal site ft230, where only three biological replicates were taken. At each location, seawater was collected at 1 m above the seafloor using a 10 L Niskin bottle. From each 10 L Niskin bottle, a subsample of 2 L was collected in clean commercial plastic drinking water bottles, using a sterilized 200 µm mesh nylon prefilter to remove bigger pieces of debris. Between locations, the Niskin bottles were thoroughly rinsed with commercial source water. The water samples were either immediately filtered on board (GeoOcean V) or stored in the dark at -20 °C (Simon Stevin) until further processing. Each sample was filtered over a 0.45 µm Sterivex polyvinylidene fluoride (PVDF) filter until the filter was nearly clogged or until 1 L was filtered. The filters were stored at -20 °C until eDNA extraction.

Quality Control

Negative control samples were collected in the field and laboratory environments. Negative field controls were taken by collecting commercial source water from the Niskin bottles after they were carefully rinsed using commercial source water, also using the prefilter. Negative filter controls were included by filtering source water over a blanco 0.45 µm Sterivex filter. Negative extraction controls were included by applying the extraction protocol on blanco 0.45 µM Sterivex filters. Negative PCR controls were included by replace the extracted eDNA with 3 µl of UltraPure™ water.

Method steps

1. eDNA extraction: After overnight incubation with the lysisbuffer at 56°C, the eDNA was extracted using the DNeasy Blood and Tissue kit (Qiagen).
2. Library preparation: The genetic analysis was based on two molecular markers (12S for the fish species, and COI for the invertebrate species). A one-step amplification protocol was used in triplicate using fusion primers (Sigma Aldrich), which contained the template specific primer sequence and a unique barcode tag of 6 to 10 nucleotides. After amplification, the PCR replicates were pooled and purified using magnetic CleanNGS beads (CleanNA). After purification, the COI pools were quality checked on the BioAnalyzer. The eDNA libraries were sequenced with Illumina MiSeq, 2 x 300 bp.
3. Bioinformatic processing: The quality of the raw Illumina MiSeq sequencing reads was verified with FASTQC v0.11.9. The paired-end reads were then reorientated, demultiplexed and trimmed by using cutadapt. After demultiplexing, DADA2 was used for denoising, dereplication, merging, and removing of chimeric reads from the demultiplexed sequences. The taxonomic assignment of the resulting ASV sequences was performed against a custom made reference database using RDP classifier in DADA2 with a minimum bootstrapping support of 80. ASVs that remained unassigned at species level with RDP were successively run with BLASTn v2.12.0 against the custom made reference databases and the GenBank nucleotide database (from October 2022). After taxonomic assignment, the count table was cleaned by removing all the ASVs identified as contaminant by the prevalence method in Decontam using the field, filter, DNA extraction and PCR negative control samples.

Belgian part of the North Sea