DNA Metabarcoding of Deep-Sea Sediment Communities Using COI: Community Assessment, Spatio-Temporal Patterns and Comparison with 18S rDNA

Abstract

Among the complex ecosystems and habitats that form the deep sea, submarine canyons and open slope systems are regarded as potential hotspots of biodiversity. We assessed the spatial and temporal patterns of biodiversity in sediment communities of a NW Mediterranean Canyon and its adjacent open slope (Blanes Canyon) with DNA metabarcoding. We sampled three layers of sediment and four different depths (900–1750 m) at two seasons, and used a fragment of the mitochondrial gene cytochrome c oxidase subunit I (COI) as a metabarcoding marker. The final dataset contained a total of 15,318 molecular operational taxonomic units (MOTUs). Metazoa, Stramenopiles and Archaeplastida were the dominant taxa and, within metazoans, Arthropoda, Nematoda and Cnidaria were the most diverse. There was a trend towards decreasing MOTU richness and diversity in the first few cm (1 to 5) of the sediment, with only 26.3% of the MOTUs shared across sediment layers. Our results show the presence of heterogeneous communities in the studied area, which was significantly different between zones, depths and seasons. We compared our results with the ones presented in a previous study, obtained using the v7 region of the 18S rRNA gene in the same samples. There were remarkable differences in the total number of MOTUs and in the most diverse taxa. COI recovered a higher number of MOTUs, but more remained unassigned taxonomically. However, the broad spatio-temporal patterns elucidated from both datasets coincided, with both markers retrieving the same ecological information. Our results showed that COI can be used to accurately characterize the studied communities and constitute a high-resolution method to detect ecological shifts. We also highlight that COI reference databases for deep-sea organisms have important gaps, and their completeness is essential in order to successfully apply metabarcoding techniques.

Bioinformatic pipelines combining denoising and clustering tools allow for more comprehensive prokaryotic and eukaryotic metabarcoding

“…The databases were downloaded from the DADA2 website (https://benjjneb.github.io/dada2/training.html) and from the FROGS website (http://genoweb.toulouse.inra.fr/frogs_databanks/assignation). Finally, to evaluate the effect of swarm clustering, ASV tables were clustered with swarm v2 (Mahé et al, 2015) in FROGS (http://frogs.toulouse.inra.fr/) at d ‐values (i.e., nucleotide differences) ranging from 1 to 13 ( d = 1, 3, 4, 5, 11 for 18S/16S, and d = 1, 5, 6, 7, 13 for COI), based on settings previously used in the literature (Andújar, Arribas, Gray, et al, 2018; Atienza et al, 2020; Clare et al, 2016; Cordier et al, 2019; Djurhuus et al, 2017; Laroche et al, 2018; Sawaya et al, 2019; Turon et al, 2020; Wood et al, 2019). Resulting OTUs were chimera‐filtered and taxonomically assigned via RDP and BLAST+ with the databases stated above, using standard FROGS procedures.…”

DNA Metabarcoding Methods for the Study of Marine Benthic Meiofauna: A Review

“…Compared to isolation of meiofauna, using sediment samples may be easier, it limits processing time, it avoids contamination, and it does not introduce any bias due to different methods of specimen isolation (Fonseca et al, 2018;Fais et al, 2020b). Our survey highlights that sediment samples are dominant in studies targeting environments hard to sample, such as the deep sea (e.g., Sinniger et al, 2016;Atienza et al, 2020;Kitahashi et al, 2020), and muddy and fine sediments difficult to elutriate (i.e., separating meiofauna from sediment by specific gravity using a liquid stream, Somerfield et al, 2005;viz., Lanzén et al, 2017;Faria et al, 2018;Salonen et al, 2019;Brandt et al, 2020). In addition, studies including a wider size range of benthic organisms such as micro-and macrofauna assessed through metabarcoding of sediment eDNA (e.g., Xie et al, 2017Xie et al, , 2018Fais et al, 2020a,b;Klunder et al, 2020a,b) might have artificially raised the number of studies in which DNA was directly extracted from sediment samples.…”

“…Notwithstanding, meiofaunal species are often underrepresented in molecular reference databases (Sinniger et al, 2016;Curry et al, 2018;Wangensteen et al, 2018), leading to high percentages of uncharacterized OTUs. For example, 70.9% of the eukaryotic clusters remained unassignable at the phylum level in a deep-sea study by Atienza et al (2020). However, it is important to consider that a high number of unassigned reads in COI datasets can be a result of non-target amplification of bacterial and other non-target species (Weigand and Macher, 2018).…”

“…Moreover, a consensus on an ideal metabarcoding approach for monitoring and assessing benthic meiofauna has not yet emerged. Each step in metabarcoding workflows is a potential source of bias in recovering community diversity and composition, as shown for example for the type of sample used for DNA extraction (i.e., whole sediment cores versus meiofauna isolated from sediment, Brannock and Halanych, 2015;Klunder et al, 2019;Pansu et al, 2021), sampling depth and design (Montagna et al, 2017;Nascimento et al, 2018;Fais et al, 2020b), target genes (Haenel et al, 2017;Cordier et al, 2019;Atienza et al, 2020;Laroche et al, 2020), primer pairs (Cowart et al, 2015;Fais et al, 2020b), and bioinformatic processing (Brannock and Halanych, 2015;Leasi et al, 2018;Antich et al, 2021). This implies that careful consideration and evaluation of the applied methods depending on the research question is crucial, and accurate reporting of adopted methodologies is important for better comparability between studies.…”

Grab what you can—an evaluation of spatial replication to decrease heterogeneity in sediment eDNA metabarcoding

“…Conversely, when extracting eDNA directly from sediment, it is more common to use molecular markers that target microorganisms, such as the eukaryotic 18S rRNA or the prokaryotic 16S rRNA genes (e.g., Guardiola et al, 2016 ; Keeley, Wood & Pochon, 2018 ; Lanzén et al, 2016 ; Lanzén et al, 2020 ; Sinniger et al, 2016 ; Stoeck et al, 2018 ). In several recent sediment extract studies using multiple markers, however, ribosomal markers have been supplemented with e.g., COI with promising results ( Atienza et al, 2020 ; Laroche et al, 2020 ; Macher et al, 2018 ; Mauffrey & Cordier, 2020 ). These studies suggest that sediment extracts may provide useful results despite a potential increase in spatial heterogeneity, which is significant as direct sediment extraction is a faster approach than separating organisms from the sediment, and thus especially suitable when processing a large number of samples for e.g., marine monitoring purposes.…”