Using vertebrate environmental DNA from seawater in biomonitoring of marine habitats

Abstract

Conservation and management of marine biodiversity depends on biomonitoring of marine habitats, but current approaches are resource-intensive and require different approaches for different organisms. Environmental DNA (eDNA) extracted from water samples is an efficient and versatile approach to detecting aquatic animals. In the ocean, eDNA composition reflects local fauna at fine spatial scales, but little is known about the effectiveness of eDNA-based monitoring of marine communities at larger scales. We investigated the potential of eDNA to characterize and distinguish marine communities at large spatial scales by comparing vertebrate species composition among marine habitats in Qatar, the Arabian Gulf (also known as the Persian Gulf), based on eDNA metabarcoding of seawater samples. We conducted species accumulation analyses to estimate how much of the vertebrate diversity we detected. We obtained eDNA sequences from a diverse assemblage of marine vertebrates, spanning 191 taxa in 73 families. These included rare and endangered species and covered 36% of the bony fish genera previously recorded in the Gulf. Sites of similar habitat type were also similar in eDNA composition. The species accumulation analyses showed that the number of sample replicates was insufficient for some sampling sites but suggested that a few hundred eDNA samples could potentially capture >90% of the marine vertebrate diversity in the study area. Our results confirm that seawater samples contain habitat-characteristic molecular signatures and that eDNA monitoring can efficiently cover vertebrate diversity at scales relevant to national and regional conservation and management.

Keywords: ADNa; Arabian Gulf; DNA 条形码技术; Golfo Arábigo; biomonitoreo; biomonitoring; eDNA; fish; meta-código de barras; metabarcoding; peces; 生物监测; 阿拉伯湾; 鱼类.

Figure 1

Location of sampling sites in the Arabian Gulf.

Figure 2

Species accumulation in 6 habitat types in (a) 2016 and (b) 2017 (gray shading, 95% CI based on the unconditional SD).

Figure 3

Taxonomic tree of all vertebrate families detected via eDNA across habitat types and sampling seasons. Node color and size reflect the number of eDNA samples each taxon was detected in (^*, detection should be interpreted with care because the detected fish, Deania sp., is a deep‐sea shark and neither Deania nor fish in the family Centrophoridae have to our knowledge ever been recorded in the Arabian Gulf).

Figure 4

Overlap in fish detections at genus level between eDNA results, visual census results obtained in parallel with eDNA sampling, and results from a trawling expedition in the central Gulf in 1990 (^*, detection should be interpreted with care because Deania is a genus of deep‐sea sharks and neither fish in the genus nor family Centrophoridae have to our knowledge been recorded in the Gulf; ^**, detected with eDNA in a single PCR replicate).

Figure 5

(a, b) Hierarchically clustering trees of eDNA sampling sites based on dissimilarity in the presence or absence of vertebrate taxa between sampling sites and (c, d) results of canonical analysis of principal coordinates ([a, c], 2016; [b, d], 2017). In (c) and (d), samples from the same habitat types are connected in polygons. On average across 4 rarefaction runs, the first 2 PCoA axes provide a mean cumulative classification success of 36% (SE 1) and 55% (1) of the observed variation, respectively, for 2016, and 51% (0) and 71% (0), respectively, for 2017 (rounded to nearest integer).

Figure 6

Species accumulation detected via eDNA in the Arabian Gulf across the 21 sampling sites (a) in 2016, (b) in 2016 with extrapolation of the fitted Lomolino model to 200 sampling sites, (c) in 2017, and (d) in 2017 with extrapolation of the fitted Lomolino model to 200 sampling sites (gray shading, 95% CIs based on unconditional SD). Samples from 30 m depth are excluded.