Data from Ershova E, Wangensteen OS, Descoteaux R, Barth-Jensen C, Præbel K. 2021. Metabarcoding as a quantitative tool for estimating biodiversity and relative biomass of marine zooplankton
Description
We explored an enhanced quantitative approach by metabarcoding whole zooplankton communities using a highly degenerate primer set for the mitochondrial marker COI and compared the results to biomass estimates obtained using the traditional morphological approach of processing zooplankton samples. As expected, detected species richness using the metabarcoding approach was 3-4 times higher compared to morphological processing, with the highest differences found in the meroplankton fraction. About 75% of the species identified using microscopy were also recovered in the metabarcoding run. Within the taxa detected using both approaches, the relative numbers of sequence counts showed a strong quantitative relationship to their relative biomass, estimated from length-weight regressions, for a wide range of metazoan taxa. The highest correlations were found for crustaceans and the lowest for meroplanktonic larvae. Our results show that the reported approach of using a metabarcoding marker with improved taxonomic resolution, universal coverage for metazoans, reduced primer bias and availability of a comprehensive reference database, allow for rapid and relatively inexpensive processing of hundreds of samples at a higher taxonomic resolution than traditional zooplankton sorting. The described approach can therefore be widely applied for monitoring or ecological studies.
Files
Steps to reproduce
The first quality filtering steps of the bioinformatics pipeline were conducted using OBITools v1.01.22 (Boyer et al. 2016). Paired-end reads were aligned with illuminapairedend and reads with alignment score > 40 were kept. Demultiplexing and removal of primer sequences were done using ngsfilter. Reads with length between 299 and 320 and without ambiguous nucleotides were selected using obigrep and dereplicated using obiuniq. The uchime_denovo algorithm (Edgar et al 2011) implemented in vsearch v1.10.1 (Rognes et al., 2016) was then used to remove chimeric sequences. Step-by-step clustering was performed in SWARM 2.1.13 (Mahé et al., 2015) using a distance value of d=13 to cluster individual sequences into Molecular Operational Taxonomic Units (MOTUs). This distance value has previously been used to cluster similar datasets using the same COI fragment (e.g. Bakker et al., 2019; Antich et al., 2020; Atienza et al., 2020). After removing singletons (MOTUs with abundance of 1 read), taxonomic assignment of the representative sequence of remaining MOTUs was then performed using ecotag (Boyer et al., 2016) against DUFA-Leray v.2020-06-10, a custom reference database (publicly available from github.com/uit-metabarcoding/DUFA), which included Leray fragment sequences extracted from BOLD and Genbank, completed with in-house generated sequences.