Environmental DNA (eDNA) metabarcoding is emerging as a novel, objective tool for monitoring marine metazoan biodiversity. Zooplankton biodiversity in the vast and important open ocean is currently monitored through continuous plankton recorder (CPR) surveys, using ship-based bulk plankton sampling and morphological identification. We assessed whether eDNA metabarcoding (2 L filtered seawater) could capture similar Southern Ocean biodiversity as conventional CPR bulk sampling (~1500 L filtered seawater per CPR sample). We directly compared eDNA metabarcoding with (i) conventional morphological CPR sampling and (ii) bulk DNA metabarcoding of CPR collected plankton (two transects for each comparison, 40 and 44 paired samples respectively). A metazoan‐targeted cytochrome c oxidase I (COI) marker was used to characterize species-level diversity. In the 2 L eDNA samples this marker amplified large amounts of non‐metazoan picoplanktonic algae, but eDNA metabarcoding still detected up to 1.6 times more zooplankton species than morphologically analysed bulk CPR samples. COI metabarcoding of bulk DNA samples mostly avoided non-metazoan amplifications and recovered more zooplankton species than eDNA metabarcoding. However, eDNA metabarcoding detected roughly two thirds of metazoan species and identified similar taxa contributing to community differentiation across the subtropical front separating transects. We observed a diurnal pattern in eDNA data for copepods which perform diel vertical migrations, indicating a surprisingly short temporal eDNA signal. Compared to COI, a eukaryote-targeted 18S ribosomal RNA marker detected a higher proportion, but lower diversity, of metazoans in eDNA. With refinement and standardization of methodology, eDNA metabarcoding could become an efficient tool for monitoring open ocean biodiversity.

See materials and methods of Molecular Ecology paper "Capturing open ocean biodiversity: comparing environmental DNA metabarcoding to the continuous plankton recorder"

All raw fastq files were uploaded (gz compressed). They originate from two separate sequencing runs - file names starting with "V4..." are from the first run, file names starting with e.g. "003_V4..." are from the second run. Sample specific tags are in the respective COITags.csv files.

The pipeline to process the raw data is in the file "Pipeline_processing.R".

The html output of the R markdown code to analyse the data (includes code for all figures and supplemental material) is in the file Figures_and_Analyses_R_Markdown_2020-06-25.html, and all processed data input files are provided in .csv or .fasta format.