Metabarcoding the marine environment: from single species to biogeographic patterns

Abstract

In this special issue, we assemble novel research and reviews that contribute to the growing literature on the utility of metabarcoding for the study of marine systems. Included are important contributions that highlight methodological advancements and that employ metabarcoding techniques across a diversity of taxa to answer questions ranging from the ecology of eDNA to food web dynamics and biogeographic drivers. Marine systems are some of the last “global frontiers” and home to abundant biodiversity, yet more than 80% of ocean habitats remain unexplored. With ongoing anthropogenic and climatic pressures leading to significant changes in ecosystem functioning, the continued resilience of marine systems is uncertain (Darling & Côté, 2018). Effective solutions for the management and conservation of these vast habitats are reliant on our understanding of their structure and interconnectedness, yet even the detection and cataloging of species can be logistically difficult. Here, we assemble novel research and reviews that contribute to the growing literature on the utility of metabarcoding for the study of marine systems. Included are important contributions that highlight methodological advancements and that employ metabarcoding techniques across a diversity of taxa to answer questions ranging from the ecology of eDNA to food web dynamics and biogeographic drivers. The articles published herein include studies conducted in eight countries with authors from across the globe (Figure 1), highlighting the importance of collaborative science. Over two-thirds of the studies include an industry or government partner emphasizing the applied nature of this technology and its potential to address an array of scientific and conservation-based questions. The growing use of metabarcoding in marine systems will allow for not only much needed biodiscovery and the augmenting of species inventories but importantly this work also provides baseline data to monitor and measure future changes in the ocean. Despite growing interest and nearly a decade of study, the application of metabarcoding for eDNA analyses is still laden with unanswered questions and highly variable methodological approaches. Differences in sampling and laboratory protocols, as well as methods of data analyses, challenge experimental replication and cross-study interpretation. In this special issue, Kumar et al. (2021) contribute one of only a few published studies that employ replicated experiments designed specifically to assess the impact of sampling effort, DNA extraction method, and filter pore size on DNA yield, PCR inhibition, and metabarcoding results (16S rDNA for fishes). Working in both inshore turbid- and offshore clear-water environments, their findings reinforce the importance of methodological validation when working in novel environments as the results for the two habitat types were strikingly different. The high turbidity samples taken from a biodiverse estuarine environment resulted in high DNA yield but also high levels of PCR inhibition that could only be alleviated with either extensive dilution or a commercial cleaning kit. Furthermore, offshore- or clear-water samples resulted in very low DNA yields demonstrating that when working in these habitats much larger volumes of water need to be filtered. A critical hurdle for eDNA research is our current lack of understanding of eDNA decay rates under different environmental conditions. This is particularly true for metabarcoding experiments, as much of what has been published has been from qPCR studies on single species. Holman et al. (2021) employed mesocosm experiments and COI metabarcoding of eDNA to examine decay rates in mixed communities of temperate marine metazoans. After stocked organisms were removed from the tanks, eDNA remained detectable for 24 hours with a substantial change in beta diversity between 24 and 48 hours. After 48 hours, most stocked species became undetectable. The authors compared log-linear and non-linear models for estimating eDNA decay functions in eight marine species and found the former proved the best fit with estimates for decay rate constants (k, which is equal to the fitted slope of the log-linear regression) varying across species. The calculation of eDNA decay rates is vital for modeling eDNA persistence and transport, and thus Holman et al. (2021) not only made an important contribution but also emphasize that such parameters should be applied with caution. The use of metabarcoding of eDNA to characterize marine communities and monitor populations is emerging as an important management tool. Traditional survey methods can be expensive and time-consuming, and the potential efficiency of eDNA techniques is set to change the way we conduct biodiversity research. Metabarcoding of eDNA may prove to be particularly useful for monitoring hyper-diverse systems such as coral reefs and estuarine habitats. However, prior to altering or even augmenting monitoring protocols resource managers need to understand the tradeoffs and benefits of traditional versus eDNA technologies. Baited Remote Underwater Videos (BRUVs) are commonly used to survey fish communities. Cole et al. (2021) compared BRUV surveys with three eDNA metabarcoding assays to characterize fish communities residing on intertidal oyster reefs and adjacent sandy areas. Overall, eDNA surveys detected a higher species richness compared to the BRUVs. However, there were also significant differences among the three eDNA assays, demonstrating that a multi-assay eDNA metabarcoding approach should be considered to capture overall biodiversity. Furthermore, BRUV and eDNA metabarcoding techniques detected distinct fish communities, with some species recovered only by a single method indicating that a combined approach is desired if the goal is to capture total community diversity in these intertidal habitats. West et al. (2021) and Dugal et al. (2021) focus on an important coral reef network in the Kimberley region of northwestern Australia. Specifically, they showed that eDNA metabarcoding of the ITS2 region resulted in high rates of detection at the species and genus level for the habitat building scleractinian corals (Dugal et al., 2021; West et al., 2021), as well as other benthic invertebrates such as sponges and tunicates (West et al., 2021). Comparing the metabarcoding data with visual surveys, both studies found a high degree of taxonomic overlap and also identified taxa that were detected by only a single method. Employing a multi-assay approach, Saenz-Agudelo et al. (2021) target whole vertebrate communities including fishes, birds, amphibians, and mammals in the biodiverse Rio Cruces Wetlands in Chile using a combination of 12S and 16S rRNA primers. Despite targeting well-known groups, the majority of sequence IDs (ASVs) could only be identified to order or lower taxonomic levels, highlighting that for this geographic region, as for many others, the lack of reference sequence data hinders the accurate identification of species from eDNA surveys. Despite this shortcoming, the eDNA assays broadly detected all groups historically recorded from the system, including rare and exotic species. Furthermore, despite the hydrodynamic complexity of the study region and uncertainties around the decay and transport of eDNA in this system, the authors were able to detect spatial structuring of communities based on distance to the ocean outlet, particularly fishes. As wetlands are critical habitats that provide important ecosystem services, eDNA metabarcoding is a promising tool to support the conservation of these valuable ecosystems. While there has been a great deal of effort allocated to the development of metabarcoding assays for well-studied groups such as marine fishes, there is a paucity of assays for the lesser-studied fractions of marine fauna. Sponges (phylum Porifera) are an ancient lineage of ecologically important metazoans that are particularly difficult to identify taxonomically. Timmers et al. (2020) working in the Pacific Ocean, make significant advancements in community assessments for this troublesome group by comparing species-level identifications made by an expert taxonomist to COI metabarcoding results. While methodological issues remain including primer bias, incomplete DNA reference libraries, and the difficult nature of sponge taxonomy, this is one of the first published papers to use metabarcoding methods for this historically understudied group. Furthermore, Timmers et al. (2020) employed non-traditional sampling devices called Autonomous Reef Monitoring Structures (ARMS), which are standardized 3D sampling platforms that allow for the gradual settling and accretion of benthic organisms as they recruit from the plankton into a new habitat. These devices, first developed over a decade ago for collecting coral reef-associated invertebrates, are now being used across the globe, to assay marine communities. For instance, Nichols et al. (2021) employed ARMS to survey the cryptobiomes, or the unseen cryptic animals that live within the reef structure, on coral reefs in Hawai'i. They sampled the biological material that had settled on the ARMS and using COI metabarcoding compared this to the eDNA in water samples taken from the water column as well as water taken from the crevices within the reef habitat. Interestingly, the authors found negligible overlap between methods. Taxa that were unique to the ARMS were chitinous and calcifying organisms and red algae, which were underrepresented in the water samples. Organisms that reside in the water column (fishes), on the other hand, were detected more often in the ambient water samples. The disparity in species detection across methods highlights the need to carefully consider the research question or the management need prior to starting any metabarcoding experiment. DNA metabarcoding is moving the field of biodiversity monitoring into areas not previously considered. de Bruyn et al. (2021) present a novel application by sequencing DNA fragments bound to gillnets (barrier nets) that were designed to protect bathers from harmful interactions with sharks in Australia. Here, de Bruyn et al. (2021) sampled the mesh surrounding larger holes in the net and extracted DNA directly from the fibers, which were presumed to represent depredation events or instances of animal escape. Using this approach, they were able to detect previously entangled and landed species of sharks and three additional species not previously captured. From a methodological perspective, this study reinforces the need to compare multiple forms of survey data to ground truth eDNA detections when developing novel techniques. In this case, catch data were critical for distinguishing among true detections, residual detections (false positives), missed taxa (false negatives), and potential cross-contamination. This study opens the door for an entirely new approach to better understand species/gear interactions and document mortalities caused by the fishing gear. One of the most anticipated advancements in marine metabarcoding is in moving beyond simple species detection to the use of these technologies to estimate species abundance and biomass. Quantitative species data based on the concentration of eDNA in a sample, although fraught with issues, are an active area of research. West et al. (2021) surveyed benthic communities such as corals, sponges, and tunicates on coral reefs in Australia using the ITS2 region. They found only a weak correlation between hard coral abundance, as determined by visual surveys and metabarcoding read counts, suggesting that in this system at least, relative read abundance from ITS2 is a poor proxy for estimating hard coral coverage. However, in a review of 63 published studies, Rourke et al. (2021) found much stronger correlations for fishes. The authors found that 90% of studies identified a positive correlation between biomass/abundance estimates of the focal species and read count, with most studies using species-specific detection methods (81% of studies across 46 species). They show that results were influenced by several biotic factors including target species, body size, distribution, and reproductive mode whereas influencing abiotic factors were largely the hydrological processes affecting the dispersal and persistence of eDNA (water flow and temperature). While the advantages of eDNA monitoring relative to the time intensive and expensive traditional methods are clear, further development of eDNA approaches is necessary. Detailed dietary information is essential for understanding the functional role of organisms and the dynamics of complex food webs. Yet, our knowledge of the diet of most marine species remains poorly resolved. This is due in part to the fact that most traditional methods can only identify prey to a coarse taxonomic level (e.g., in situ feeding observations or morphological identification of prey items in gut contents). This oversimplification of diet has fueled assumptions of broad trophic overlap and functional redundancy in marine communities. However, recent DNA-based dietary analyses have revealed much more pronounced levels of dietary specialization than previously assumed. Prey-specific DNA fragments remain detectable in gut contents and feces for several hours after ingestion, hence providing a snapshot of what an organism consumed with an unprecedented level of taxonomic resolution. Nalley et al. (2021a, b) build on this methodology to describe the breadth of resources ingested by herbivorous coral reef fishes in Hawai'i and quantify levels of dietary partitioning and overlap. Despite playing an important ecological role on coral reefs the trophic niche of herbivorous fishes remains poorly resolved. Species are typically classified into broad functional groups (grazers, browsers, scrapers, and excavators) based on jaw morphologies, feeding behavior, and on how they affect the underlying substratum. Combining a literature review and DNA metabarcoding of gut contents, Nalley et al. (2021a) identify considerable differences in the level of dietary specialization of species within these functional groups and a significant overlap in diet composition of some species across functional groups. Their results show the limitations of functional groupings that assume trophic equivalency. How herbivorous diets vary across reefs in response to food availability also remain an open question that DNA metabarcoding of gut contents can help address. Nalley et al. (2021b) explore the versatility of the diet of two common congeneric surgeonfish species across more than ten reefs that differ in benthic cover in Hawai'i. Results revealed that these two species do not have equally flexible diets with one species targeting a narrower suite of resources which may make it more vulnerable to coral reef degradation than the more opportunistic congeneric. Species-level identification of prey represents a wealth of information on how animals transfer nutrients between ecosystems. Tropical seabirds redistribute nutrients from vast oligotrophic marine seas to terrestrial ecosystems, but the exact pathways remain poorly known because prey digested while at sea quickly become unidentifiable. Nimz et al. (2021) conduct metabarcoding sequencing of the regurgitations of the Christmas Shearwater (Puffinus nativitatis). Nocturnal reef fish likely captured at the larval stage in the pelagic environment were among the most abundant prey. Other common prey included fish generally living at depths ranging from 200 to 5000 m (mesopelagic zone) during the day, which is well beyond the maximum diving depth of these seabirds. Thus, these seabirds likely consumed mesopelagic fish when they migrated to the surface at night, a feeding behavior that has not been observed for birds in this colony (Nimz et al., 2021). Wells et al. (2020) identified an unsuspected coupling between terrestrial and marine food webs using DNA metabarcoding of the gut contents of the giant plumose anemone (Metridium farcimen). The giant plumose anemone is a dominant suspension feeder in the northeast Pacific Ocean that thrives on subtidal rocky shores and on human made structures such as floating docks where they consume zooplankton suspended in the water column. DNA metabarcoding analysis of its gut contents indicated a diverse and opportunistic diet characterized by a majority of arthropods, annelids, and mollusks. However, approximately 10% of the sequences were from a species of ant that had mating flights at the time samples were collected demonstrating a trophic link that had never been reported. One of the most exciting applications to arise from eDNA metabarcoding approaches is the potential for “real-time” assessments of biodiversity, particularly across vast spatial scales. The Coral Triangle, arguably the world's most diverse marine ecosystem, provides crucial socio-ecological services for millions of people throughout the region. With much of the biodiversity in the Coral Triangle under threat, management and conservation decision making requires a better understanding of the spatiotemporal patterns of biodiversity. In this issue, Marwayana et al. (2021) sampled across the Indonesian Archipelago and demonstrated that 12S rRNA metabarcoding of eDNA could detect regional patterns of fish biodiversity, such as distinguishing between Indian and Pacific Ocean communities. However, they cite the lack of a complete reference database as a significant obstacle to assigning OTUs to species, which likely led to low rates of assignment for some sites. The authors suggest that when working in this hyper-diverse environment that high-intensity eDNA sampling coupled with traditional visual surveys may be required to achieve high species coverage and complete assessments. Sampling intensity, specifically sampling replication, is crucial when aiming to capture community diversity. This was demonstrated by Czachur et al. (2021) who surveyed fish diversity across seven sites spanning over 2,000 km of the South African coastline. As predicted from known patterns of species diversity, richness increased from west to east along the coast, with a maximum of 132 OTUs recovered and a total of of known of fishes The authors included a replication sampling between one to seven to show that particularly for sites with high species diversity, there was of lower abundance or rare across with high read OTUs The authors suggest that three biological are likely to capture fish diversity even in sites with high species However, Czachur et al. (2021) that additional replication may be for groups such as invertebrates, for which higher levels of diversity are and DNA reference are et al. (2021) surveyed fish diversity along a environmental in an array of and marine habitats spanning Using eDNA metabarcoding the authors detected the of to marine species as and both and more species Despite the potential for transport of their analyses showed than more even communities of fishes, suggesting that of eDNA was not to of community This work the combination of studies, for future cross-study analyses, as well as highlighting the use of eDNA as a tool for and marine fish communities. are the marine and their biogeographic are well known compared with other groups of However, their patterns do not of other taxonomic groups, on the of marine more et al. (2021) employed a combination of different of rRNA rRNA metazoans and and 16S to biogeographic patterns known from fishes across km of the are of community showed a significant biogeographic by different functional groups versus rocky reef that was also detected in corals and sponges, as well as the more The data novel into the between environmental and community structuring in the of and the In sea surface and A concentration as for different environmental with the biogeographic patterns et al. (2021) were also able to show that distance to human influenced community structure, although further research is needed to better anthropogenic on communities in the region. DNA metabarcoding is a tool with broad to address biological questions diversity, distribution, and ecology and the growing of applications for metabarcoding is in the diversity of studies they not only highlight the vast potential of these technologies but also some of the now The need for better reference sequence is nearly with the diversity of species in most biodiverse of the developing poorly is an understanding of how environmental and patterns impact eDNA decay rates and Furthermore, while the to read abundance into is we are still a way from if these are even particularly in systems from laboratory studies highlight eDNA metabarcoding as a than tool for monitoring marine biodiversity. the future should be in the of when reference and modeling that into abiotic factors that the persistence and transport of eDNA in marine systems. The authors have of interest to

Topics & Concepts