Comparison of DNA metabarcoding and microscopy in analysing planktonic protists from the European Arctic

In the era of climate change-related restructuring of planktonic protist communities, it is especially important to identify possible shifts in their taxonomic composition. While traditional microscopy-based morphological classification is time-consuming and requires experienced taxonomists, metabarcoding seems to substantially accelerate the determination of taxonomic composition. In this study, based on samples collected in summer 2019 from the West Spitsbergen Current, we analysed planktonic protists using both methods. Metabarcoding, based on high-throughput sequencing of the V4 region of the 18S rRNA gene, resulted in a much higher number of operational taxonomic units (OTUs) and sample diversity than microscopy, although the resolution of taxonomic identification ranged from species to phyla. Most morphology-based identification was performed at the species or genus level, additionally allowing us to include information about dominants and size fractions. The highest proportion of 45% shared taxa by both methods was recorded at the class level. The composition of dominant protists differed between the approaches, with most similarities being observed in Bacillariophyceae, for which two genera, Thalassiosira and Eucampia, were found to be the most abundant with both methods. For Dinophyceae, the most abundant representatives identified by microscopy were Gymnodinium spp., Prorocentrum minimum and Gonyaulax gracilis, while in the metabarcoding approach, most dinoflagellates were identified to the class level only. Given the different levels of accuracy of taxonomic determinations and possible biases in results connected to the chosen methodology, we advocate using an integrative taxonomic approach for the classification of planktonic protists based on the combination of microscopy and molecular methods.