Biases in bulk: DNA metabarcoding of marine communities and the methodology involved

Abstract

With the growing anthropogenic pressure on marine ecosystems, the need for efficient monitoring of biodiversity grows stronger. DNA metabarcoding of bulk samples is increasingly implemented in ecosystem assessments and is more cost-efficient and less time-consuming than monitoring based on morphology. However, before raw sequences are obtained from bulk samples, a profound number of methodological choices must be made. Here, we critically review the recent methods used for metabarcoding of marine bulk samples (including benthic, plankton and diet samples) and indicate how potential biases can be introduced throughout sampling, pre-processing, DNA extraction, marker and primer selection, PCR amplification and sequencing. From a total of 64 studies evaluated, our recommendations for best practices include to (a) consider DESS as a fixative instead of ethanol, (b) use the DNeasy PowerSoil kit for any samples containing traces of sediment, (c) not limit the marker selection to COI only, but preferably include multiple markers for higher taxonomic resolution, (d) avoid touchdown PCR profiles, (e) use a fixed annealing temperature for each primer pair when comparing across studies or institutes, (f) use a minimum of 3 PCR replicates and (g) include both negative and positive controls. Although the implementation of DNA metabarcoding still faces several technical complexities, we foresee wide-ranging advances in the near future, including improved bioinformatics for taxonomic assignment, sequencing of longer fragments, and the use of whole-genome information. Despite the bulk of biases involved in metabarcoding of bulk samples, it is clear that DNA metabarcoding provides a valuable tool in ecosystem assessments. Biases in bulk: DNA metabarcoding of marine communities and the methodology involved

Towards reproducible metabarcoding data: Lessons from an international cross‐laboratory experiment

“…Reporting unverified biosecurity risks from HTS data may lead to inadequate management responses and even potential legal entanglements for researchers (Darling et al, 2020). Despite recent calls for improved consistency in HTS‐based research outputs (Jeunen et al, 2019; van der Loos & Nijland, 2020), laboratories working in this field often follow different analytical workflows, develop and apply in‐house reference sequence databases, use customized bioinformatic pipelines, and often fail to report methodologically relevant metadata consistently (Goldberg et al, 2016; Nicholson et al, 2020). This lack of harmonization may impede the reproducibility of research, transferability and comparability of results, and ultimately, the credibility of HTS‐based biodiversity assessments (Zaiko et al, 2018).…”

Environmental DNA for biomonitoring

“…However, the authors consider both approaches as complementary and suggest that they should be combined for comprehensive assessment of the invertebrate community. The importance of bulk DNA metabarcoding as a tool for the assessment of marine ecosystems is also highlighted by van de Loos and Nijland (2021). The authors review various technical biases affecting bulk DNA metabarcoding workflow and discuss possible improvements that could help overcoming these biases in the future.…”

DNA metabarcoding marker choice skews perception of marine eukaryotic biodiversity

“…To surmount these limitations, we encourage employing a multimarker approach to metabarcoding across a variety of marine habitats and regions (see also van der Loos & Nijland, 2020). Such studies are essential to identify “dark” taxa that remain undescribed, even morphologically, and will help prioritize voucher‐based sampling to advance catalogues of marine biodiversity and hasten the completion of reference databases.…”

“…In studies focused on eukaryotic biodiversity (e.g., ARMS sampling), the mitochondrial cytochrome c oxidase subunit I (COI) and small subunit (SSU) 18S ribosomal RNA (rRNA) gene regions are two common targets that are considered “universal markers” that capture a broad range of eukaryotic taxa (Hebert et al, 2003; Leray & Knowlton, 2016; Tragin et al, 2018). Of these, 18S is the more commonly selected universal marker (59% and 28% of studies target the 18S and COI gene, respectively; van der Loos & Nijland, 2020). While both are considered universal markers that perform well across many eukaryotic taxa, they have known limitations that impact their ability to faithfully detect taxonomic diversity.…”