Monitoring plans using environmental DNA have the potential to offer a standardized and cost-efficient method to survey biodiversity in aquatic ecosystems. Among these ecosystems, coastal wetlands are key elements that serve as transition zones between marine and freshwater ecosystems and are today the target of many conservation and restoration efforts. In this sense, eDNA monitoring could provide a rapid and efficient tool for studying and generating baseline biodiversity information to guide coastal wetland management programs. Here we test an eDNA metabarcoding assay as a tool to characterize vertebrate biodiversity in one of the largest coastal wetlands of Chile, the Rio Cruces Wetland, a Ramsar designated site since 1981. We sampled surface water from 49 sites along the entire wetland. Our eDNA approach detected 91 genera of vertebrates including amphibians, fishes, mammals, and birds, as well as identified several cryptic, exotic, and endangered species. Our results also indicated that the spatial distribution of eDNA from different species is spatially structured despite the complex hydrodynamics inherent in this wetland due to the influence of daily tidal regimes. For amphibians and fishes, the number of taxa detected with eDNA was higher in the periphery of the wetland, and increased with proximity to the ocean, a pattern consistent with small-scale spatial sensitivity for some species and eDNA accumulation downstream for others. Birds and mammals showed somewhat more idiosyncratic distributions. Taken together our results add to the growing body of evidence showing eDNA can serve as a rapid cost-effective tool to characterize vertebrate communities in protected coastal wetlands, where visual surveys are difficult and animal collections are often prohibited. The use of multiple primer sets is also recommended as it facilitates the detection of ephemeral terrestrial organisms and resident aquatic organisms that make use of these wetlands.

Sample collection and laboratory processing

We collected water from the Río Cruces Coastal Wetland including its subsidiaries. Water was collected by boat (45 sites) or from the shore (4 sites) under research permit N SNRC01/2018 issued by the Corporación Nacional Forestal. A total of 49 sampling points at least 1.5 km apart from each other were distributed along most of the wetland and encompassed many of its tributaries (Figure 1). 45 sites were sampled within 4 consecutive days (12-15 February 2019) during a period of neap tide. Four sites (P18, P55, P57 and P58) that could not be accessed by boat were sampled a month later by land (14-15 March 2019). At each sampling site, three 1L water samples were taken 30 cm below the surface. Water samples were stored in glass bottles previously treated for at least 10 minutes with 10% bleach. All water samples were filtered immediately after collection in the field using manual vacuum pumps and 0.45µm pore size (47mm diameter) hydrophilic mixed cellulose esters sterile filters (Pall Corporation, NY, USA). All filtering and handling materials were soaked in the 10% bleach solution before water filtration. A filtration negative control (n = 1) was included by filtering 1L of molecular grade water (Milli-Q^® filtered) in the field following the same procedure as all other water samples. Given that water quality (including the amount of particulate matter) is highly variable in estuaries, the volume of water that we filtered varied among the sites. The general rule that we followed was to filter 1L of water whenever possible or filter water until the membrane became saturated. Filters were then stored in the laboratory in 2mL plastic tubes containing 1 mL of lysis buffer (Thermo Fisher Scientific, Waltham, MA, USA) at 4ºC until extraction within two months of collection. Overall, we filtered 148 samples of water (3 from each of the 49 sites and one Milli-Q^® water control). 

eDNA extractions: DNA extraction was performed by first placing each tube in a Mini-Beadbeater-16 (BioSpec products inc, Bartlesville, USA) for two minutes at 2.5 x1000 stroke/min. Tubes were then opened in the laminar flux chamber where 500 ml of supernatant was taken and transferred to a 1.5 ml microcentrifuge tube for DNA extraction using the GeneJET Genomic DNA Purification Kit (Thermo Fisher Scientific, Waltham, MA, USA) following the manufacture´s protocol with some modifications (50 μL of the Proteinase K solution was added and the sample vortexed, samples were then incubated at 56ºC for 3h). The rest of the procedure followed the manufacture´s recommendations. DNA samples were eluted in 60µl of elution buffer. DNA extractions were processed in batches of 11 samples, and for each batch, a negative extraction control (six in total) was included (using 500 µl lysis buffer and no added sample). Purified extracted DNA samples were stored at -20 ºC until PCR amplification.

PCR amplification: We used a two-step PCR method to amplify eDNA and to add a unique combination of dual barcodes. DNA samples, including the filter and DNA extraction negative controls, were first amplified with three primer sets targeting specific fragments of the mitochondrial 12S and 16S rRNA genes. These primers included an Illumina primer sequence, a 12 base pair (bp) barcode with a 0-4 bp spacer, and the actual target primer sequence. Each sample (n = 147) and all negative controls were amplified in duplicate. In addition, we also incorporated three PCR negative controls. Details of the PCR protocol can be found in the paper. 

The dataset consists of three files: a README file with details, a compress file containing paired demultiplexed sequence files in fastq format and one csv file with metadata for each of the sequence files.