Environmental DNA最新文献

The conservation of biodiversity is fundamental for the persistence of ecosystems, particularly under climate change. The South African marine environment is characterized by high levels of biodiversity as well as endemism, but species distribution patterns are generally not well characterized. Environmental DNA (eDNA) metabarcoding is a promising tool to help plug the biodiversity information gap, but evidence from previous studies has shown highly variable eDNA signals, even at small temporal–spatial scales (< 300 m, 24 h). Passive samplers, deployed over several hours may circumvent some of the challenges of high eDNA variation by accumulating DNA over time. Using multiple markers (mtDNA 12S rRNA, mtDNA COI), we test both active (using Sterivex) and passive sampling (gauze filled metaprobe) in the detection of kelp forest-associated biodiversity, focussing on fishes and invertebrates. We conducted our experiment across different time periods (6, 12, 24 h) and hypothesized that metaprobes at 24 h would harbor the greatest species richness. We detected 33 ESVs, assigned to 18 different fish families, with 12S rRNA, where active sampling retrieved a larger proportion of diversity (94% vs. 64%, p < 0.001). For COI, we detected 1481 ESVs assigned to 17 different phyla, 99 families, and 55 species, with passive sampling detecting more ESVs (82% vs. 67%, p < 0.001). COI metabarcoding detected an additional four families of fishes, highlighting the importance of multi-marker approaches. For passive sampling only, there was a trend of significant accumulation of numbers of reads and ESVs over time. We found that active samples were more consistent across all statistics: variation in the number of reads, ESVs, and taxa retrieved, which was less in active than passive replicates. Overall, we highlight the need for region-specific approaches and careful project planning before implementing eDNA metabarcoding as a biomonitoring tool.

Microeukaryotes, including protists, microalgae, and small fungi, are fundamental components of marine ecosystems, driving energy transfer, nutrient cycling, and primary production. Despite their ecological significance, they remain understudied due to their small size and taxonomic complexity. Here, we present a comprehensive assessment of eukaryotic diversity across the entire water column of the Eastern Red Sea, integrating surface to deep-sea layers along a latitudinal gradient. Environmental DNA metabarcoding of bottle-net hauls from 266 samples across 128 stations spanning the eastern Red Sea revealed 4298 MOTUs from 40 phyla, with a substantial fraction lacking reliable taxonomic assignments. While traditional diversity indices showed little variation across regions or depths, multivariate analyses revealed pronounced community turnover along the latitudinal gradient and subtler vertical structuring. Temperature, dissolved oxygen, and turbidity emerged as significant correlates of β-diversity, with the remainder likely reflecting unmeasured drivers. Despite a large core of MOTUs shared across epipelagic, mesopelagic, and bathypelagic layers, ecological groups displayed depth-specific abundance patterns, and co-occurrence networks identified phototrophic and microbial hubs with a southward shift toward stronger benthic–pelagic coupling. Notably, we observed a latitudinal transition from fungal- and algal-dominated communities in the north to more complex microbial–benthic–vertebrate networks in the south. Our results indicate that Red Sea eukaryote communities are shaped by both abiotic gradients and biological interactions, and that water column-integrated sampling reduces depth bias and captures biodiversity patterns overlooked by single-depth surveys. This work provides a regional baseline for biodiversity monitoring and conservation in a rapidly changing tropical sea.

Environmental DNA (eDNA) metabarcoding is widely used to detect animals from environmental samples and on the brink of being implemented into routine species monitoring. Surprisingly, birds are among the taxonomic groups which have received comparably little attention in eDNA research regarding primer optimization—particularly reducing the amplification of non-target taxa, the availability of appropriate reference sequence databases for the targeted fragments, and the evaluation of different filter types for their capability to detect avian eDNA from water samples. Here, we present a novel primer pair (BirT) for metabarcoding avian eDNA. We optimized specificity and fragment length with regard to taxonomic resolution and available sequencing platforms. Additionally, we evaluated the availability of 12S rRNA gene reference sequences for birds and filled database gaps by generating novel barcodes. Finally, we tested the applicability of the BirT primer pair using field-collected eDNA samples obtained with three different filter types and compared the eDNA metabarcoding results to visual observations uploaded to eBird (www.eBird.org). Our results confirm the suitability of the BirT primer pair for avian eDNA metabarcoding with no amplification of key non-target groups and improved taxonomic resolution. Albeit there are still substantial gaps in the 12S reference sequence database, the analysis of bird eDNA from water samples resulted in species-level taxonomic resolution for 91% of the detected taxa. All tested filters delivered similar results for total read numbers per sample (mean: 623,990 ± 331,236 SD) and species detected per sample (mean: 5.0 ± 2.0 SD). Ninety-five percent of the bird detections were highly plausible based on eBird data collected at the time of eDNA sampling and 64% were directly confirmed by visual observations. Most detected species were closely associated with aquatic habitats confirming the suitability of water samples for the detection of waterfowl and species inhabiting similar ecological niches via eDNA.

Metabarcoding fecal matter is increasingly common in dietary studies across a variety of taxa. This approach assumes that enough prey DNA remains detectable in the fecal DNA (fDNA) as the prey DNA becomes degraded while passing through the digestive tract. However, as prey DNA is degraded during digestion, diet reconstruction based on metabarcoding of fDNA could become incomplete, i.e., species for which DNA was highly degraded would not be detected. The purpose of this study was to test the use of cloacal swabs and metabarcoding fDNA for shark diet reconstruction. To do this, both stomach contents and cloacal swabs were collected from the same individual sharks and metabarcoded using two previously published primer sets targeting teleost fishes and crustaceans. Samples were collected from four coastal species of sharks in a temperate estuary: bonnethead (Sphyrna tiburo; n = 10), blacknose (Carcharhinus acronotus; n = 4), blacktip (C. limbatus; n = 4), and Atlantic sharpnose (Rhizoprionodon terraenovae; n = 5). We determined how well the fDNA represented the corresponding prey DNA from the stomach contents (stDNA) and found no statistical difference in taxonomic richness or diversity when comparing the two sample types. Our results indicate that the less invasive, non-lethal method of DNA metabarcoding cloacal swabs provided higher taxonomic resolution than more common methods for studying trophic ecology (i.e., morphological stomach contents identification and trophic biomarkers such as stable isotope analysis) with no statistical difference in overall diet description between the two sample types.

Environmental DNA (eDNA) enables sensitive detection of species from environmental samples, particularly water. Large-scale, standardized monitoring of coastal fish communities remains challenging across diverse regions. The ANEMONE Global network was established to address this gap, expanding the workflow developed in Japan to a coordinated worldwide survey using standardized eDNA metabarcoding. Between June and November 2024, 12 countries, including several in Southeast Asia, collected surface water samples from beaches, rocky shores, estuaries, and near coastal protective structures using harmonized protocols for filtration, RNAlater preservation, and metadata recording. Daytime and nighttime sampling captured temporal variation in community composition. All samples were processed with the MiFish metabarcoding protocol, quantitative internal standards, and rigorous contamination controls. Analysis of 90 samples generated over 16.6 million high-quality reads, revealing more than 500 putative fish OTUs across diverse families, genera, and species. Species richness varied geographically, reflecting differences in fish fauna, and assemblages differed across the Atlantic, Indian, North Pacific, and South Pacific Oceans. Diel variation was most pronounced in the North Pacific, and diversity patterns reflected both habitat complexity and ocean basin, with waters adjacent to coastal protective structures and rocky shores supporting the highest diversity. These findings highlight how both habitat complexity and ocean basin geography shape coastal fish assemblages, offering insights for global marine biodiversity monitoring using eDNA. This survey demonstrates that a globally standardized eDNA workflow can generate comparable quality data across ecological and logistical contexts. By combining international collaboration, open data, and locally informed implementation, ANEMONE Global provides a framework for long-term, high-resolution monitoring of coastal biodiversity and sets the stage for expanding coverage to additional aquatic ecosystems worldwide.

Accurate information on species abundances and their distribution in space is key to ecological research and essential for informed decision-making in environmental management. Environmental DNA (eDNA) allows community-wide detection of biodiversity, but its limited ability to estimate species abundance from metabarcoding outputs poses important challenges for its broader application. The eDNA Integrating Transport and Hydrology (eDITH) framework addresses some of these limitations by partly accounting for eDNA transport and decay dynamics in river networks and allows predicting the spatial distribution of taxa at high resolution. However, its capability for providing quantitative estimates of taxon abundances has so far not been empirically validated. Here, we utilized spatially replicated eDNA and kick-net samples collected in spring and summer at 25 sites along a 126 km² river catchment. We contrasted species-level relative abundances of insect communities obtained via eDNA metabarcoding read counts and via eDITH estimates with those obtained from the kick-net samples. We found that eDNA read counts of sampled locations did not correlate with the associated local kick-net (i.e., realized) insect abundance, but that the eDITH estimates did. However, results varied across insect orders and the improvements provided by eDITH were highly species-specific. Our findings corroborate the inadequacy of utilizing raw read counts for quantitative inference and pose forward the potential of the eDITH framework in circumventing some of the limitations of metabarcoding outputs. Together with other recently proposed correction approaches, this framework contributes to ongoing efforts to refine the interpretability of metabarcoding data. As the demand for quantitative biodiversity data continues to grow in both ecological research and environmental management, refining and validating sampling approaches remains a critical priority.

Human-induced global warming has triggered a persistent decline in the health of marine ecosystems, particularly coral reefs, which are experiencing increasingly frequent and severe bleaching and mortality events. Refining cost-effective and precise monitoring tools, such as environmental DNA (eDNA) metabarcoding, is essential to supplement future coral reef monitoring programs, with ongoing efforts focused on improving methods, validating results, and understanding limitations. Although eDNA has been widely used in aquatic ecosystem studies, its application to corals (Anthozoa) remains underexplored. Here, we investigate the use of eDNA metabarcoding with molecular markers targeting the ITS2 region of Anthozoa for monitoring coral communities in a remote and relatively undisturbed atoll system. We integrate three mainstream taxonomic assignment approaches (IDTAXA, BLAST Top Hits, and BLAST LCA), retaining only consensus identifications across methods for downstream analyses. This conservative strategy ensures highly robust and reliable taxonomic resolution, with over 90% of the sequences classified within Anthozoa, encompassing 18 genera and 15 genera of hard corals (Scleractinia). A considerable overlap in coral identification is observed between eDNA and traditional benthic transect surveys, giving support to the ability of eDNA to identify the community composition of Anthozoan taxa. Importantly, cryptic genera, such as Cycloseris, Cyphastrea, Merulina, Oxypora, and Turbinaria were identified by the eDNA approach but not the traditional surveys. Conversely, genera such as Alveopora, Astreopora, Caulastrea, Fungia, Galaxea, Halomitra, Herpolitha, Leptastrea, Platygyra, Plerogyra, and Stylophora were identified by the traditional surveys but not the eDNA approach, likely due to primer bias, taxonomic resolution or incomplete reference databases, supporting the complementary use of both methods. We also observe that the eDNA metabarcoding may capture differences in coral community structure between our lagoonal and seaward reef habitat types and point to potential characteristic taxa. This study underscores the utility of eDNA metabarcoding as a noninvasive, cost-effective tool for coral biodiversity monitoring and provides insights into how to improve eDNA techniques for use as a coral biodiversity monitoring tool.

Aquatic insects are iconic and ecologically highly relevant inhabitants of riverine ecosystems. They are also often the target of monitoring programs to assess the ecological status of these lotic habitats. Environmental DNA (eDNA) techniques have been widely and successfully implemented to investigate freshwater insects and other macroinvertebrates. Commonly, such monitoring is conducted at one or two timepoints per year, despite the known phenology of aquatic insects' life history strongly affecting species presence—or detection—and population dynamics through the seasons. Here, we assessed if and to which extent eDNA can capture the temporal changes of the orders Ephemeroptera (mayflies), Plecoptera (stoneflies), Trichoptera (caddisflies), and Diptera (true flies). We carried out eDNA sampling at roughly monthly intervals from April to October at 25 sites across a whole river catchment in the northeastern part of Switzerland. We found pronounced, cyclic phenological trends in all orders but Trichoptera: the communities diverged from spring to summer and then in fall gradually returned closer to the spring state. The four orders exhibited different predominance in gains or losses of species detection throughout this time interval. Lastly, we found that field replicates, despite showing a relatively high local stochasticity, were able to provide a more complete assessment of aquatic communities. Field replicates, when used as a proxy for the frequency of observation of a species through the seasons, yielded comparable temporal patterns to the ones extracted from the Global Biodiversity Information Facility (GBIF) for about 35% of the investigated species. Overall, our findings demonstrate that eDNA techniques can be used to reveal intra-annual dynamics of aquatic insects. Given the current necessity to assess and monitor the biodiversity status of ecosystems, we therefore show that eDNA methods are a viable option to obtain a deeper understanding of the structuring of freshwater communities over time.

Biodiversity monitoring is essential for conservation and ecosystem management, but current terrestrial survey methods, such as visual observation, trapping, and camera traps, are labor-intensive and time-consuming. Environmental DNA (eDNA) analysis offers a non-invasive alternative and is well-established in aquatic systems, yet terrestrial eDNA methods remain underdeveloped. Existing terrestrial approaches, such as water or soil sampling and air filtration, face limitations in target coverage, DNA persistence, and ease of deployment. Surface-swabbing techniques show promise but often lack scalability for frequent, multisite sampling. We aimed to develop and evaluate a simple, power-free, and scalable terrestrial eDNA collection method capable of sampling wide ground-surface areas for biodiversity monitoring. We designed the Koro-rin sampler, consisting of a rotating body with a single-use nonwoven fabric collector. This device collects surface-associated particles (soil, leaf litter, and fine debris) from diverse substrates without a power source. Over 14 months, we collected 90 ground-surface samples and 29 water samples from a drinking site within the same secondary forest in Japan. Samples were analyzed using MiBird and MiMammal metabarcoding, and results were compared with concurrent camera trap and seasonal observation data. Metabarcoding detected 53 bird and mammal taxa from ground and surface samples. Of these, 92% were also recorded by camera traps in front of the sampling areas, validating detection sensitivity. Seasonal patterns in eDNA detections matched the arrival timing of migratory birds observed visually. The method enabled high-frequency, wide area sampling with minimal labor. The Koro-rin sampler is a practical, sensitive, and time-resolved approach for terrestrial eDNA monitoring. Its portability and disposability make it suitable for large-scale, long-term surveys. Combined with aquatic eDNA monitoring, it enables integrated ecosystem assessments.

Machine learning (ML) has been proposed as a taxonomy-independent approach using environmental DNA (eDNA) for ecosystem biomonitoring. Representations of eDNA amplicons either as clustered sequences termed operational taxonomic units (OTUs) or unique sequences termed amplicon sequence variants (ASVs) are used as inputs in current ML practices with varied successes. The use of eDNA as sole input in ML inherently limits the potential of ML for ecosystem biomonitoring and prediction. Biogeographic data encompassing the physical, climate, and ecological observations provide a repository of potential informative features that can augment ML performance in combination with eDNA data. A multimodal ML workflow using combined eDNA and biogeographic features for ecosystem biomonitoring is introduced in this study. Differentially abundant ASVs were merged with biogeographic data and used as input in an automated ML approach. Using Switzerland's freshwater macroinvertebrate eDNA dataset collected across 163 biomonitoring sites and impact prediction as an example, the multimodal ML approach (83.3% accuracy) significantly outperformed ML using only ASVs (66.7% accuracy) or OTUs (64.6% accuracy). Shapley additive explanation of the best ML model revealed key biogeographic features and species/taxa impacting upon predictions. The proposed workflow can be readily adopted in existing bioinformatics/ML pipelines and will further advance the use of eDNA for ecosystem biomonitoring across large spatiotemporal scales.