2 Key aspects and challenges in eDNA monitoring of amphibians

2.1 Sample collection and DNA capture

Amphibians exhibit a bifurcated life history, comprising both aquatic and terrestrial phases. In the former, their eDNA is disseminated into the aquatic milieu via exudates from integumentary secretions, excreta, and ova deposition. In the latter, eDNA is presumed to derive from imprints generated during their terrestrial locomotion, manifesting as soil or sedimentary remnants. Consequently, a comprehensive scrutiny of amphibian eDNA mandates the simultaneous procurement of aquatic and terrestrial substrates. The sources of collecting amphibian environmental DNA samples are mainly water, soil, and other natural substrates. Some studies have demonstrated that detecting terrestrial vertebrates from soil samples is challenging and that eDNA testing is more complex under these circumstances. For example, Walker et al. cultured salamanders in a mesocosm ecosystem and found that only 1% of salamander DNA from soil samples could be successfully amplified (Walker et al., 2017). The success of applying the eDNA method for detection depends not only on the presence and concentration of eDNA in the substrate sample but also on capture potency, DNA extraction success, sample interference (such as inhibition), and assay sensitivity (Ficetola et al., 2015; MacKenzie et al., 2002; Schultz & Lance, 2015). Filtration is the standard approach for the capture of DNA from aquatic environments (Goldberg et al., 2016); generally, volumes between 15 mL and 10 L of water are sampled to trap eDNA in a filter. In contrast to fish, amphibians may exhibit comparatively modest eDNA exudation. Some judicious augmentation of filtrate volume may therefore be considered to enhance the sensitivity of detection. Common filtration materials include nitrate cellulose (Caren S Goldberg et al., 2011), glass fiber (Jerde et al., 2011), polycarbonate (Takahara et al., 2012), nylon (Philip Francis Thomsen et al., 2012) and polyether sulfone (Renshaw et al., 2015). Chen et al. evaluated the ability of 12 different types of adsorption materials and filters to capture eDNA in both laboratory and field experiments using quantitative PCR (qPCR) analysis. They found that the ability of a glass fiber filter (GF) to capture eDNA was superior to other materials, and that the amount of eDNA captured increased linearly over 72 hours (Chen et al., 2022).
It has been demonstrated that the period of detectability of eDNA in water samples is limited to between 1 day and 8 weeks (Dejean et al., 2011; Pilliod et al., 2014; Philip Francis Thomsen et al., 2012; P. F. Thomsen, J. Kielgast, L. L. Iversen, C. Wiuf, et al., 2012), and therefore false negative results may be obtained if there is a delay in the collection of water samples relative to the time of species presence. In addition, temperature alterations caused by seasonal changes will also affect the content of eDNA in samples. For example, Wei et al. studied the variation over time of eDNA in sediments over periods of one year, one month and one tidal period, and found that eDNA content was higher in the warm season and lower in the cold season (Wei et al., 2019). Similarly, the concentration and distribution of eDNA can be influenced by spatial factors such as spatial dissipation of eDNA and the flow of the water column. In still water, the eDNA of species decreases significantly within a radius of 5–10 m from the eDNA point source (Brys et al., 2021), while in flowing waters, eDNA may move with the current to new locations hundreds of meters away (Pilliod et al., 2013), leading to false negatives at the original point as well as false positives at the new location.
Therefore, biases in results due to temporal and spatial differences need to be taken into account when designing eDNA experiments, and these factors also need to be considered when conclusions are drawn about the range of species present. For example, conclusions drawn from water samples taken along one side of a river are not directly representative of the whole river.

2.2 DNA extraction

During amphibian DNA extraction, pollution should be avoided and the facilities and equipment employed should be disinfected. Standard autoclave methods are not sufficient to destroy nucleic acid molecules(Unnithan et al., 2014). Among common stain removal methods, sodium hypochlorite solvent is the most effective for removing DNA and PCR products (Champlot et al., 2010; Prince & Andrus, 1992). The choice of eDNA extraction method after filtration can greatly affect the detection of eDNA. Deiner et al. compared the effectivity of six different combinations of sample collection and DNA extraction methods in biodiversity monitoring in water ecosystems. The results showed that different environmental DNA collection and extraction methods had a significant effect on DNA yield and the number of sequences obtained by high-throughput sequencing techniques (Deiner et al., 2015).
Common DNA extraction methods include liquid phase separation and commercial DNA extraction kits. Trials have shown that liquid phase separation methods can in some cases outperform kits in terms of extraction efficiency. Turner et al. showed that a cetyltrimethylammonium bromide (CTAB) method using polycarbonate orbital filters collected more eDNA than a PowerWater DNA isolation kit using GF membranes (1.5 µm pore size, 934-AH grade) (Turner et al., 2014). It has also been shown that the phenol-chloroform-isoamyl alcohol extraction (PCI) method using glass and nitrocellulose membranes extracted more DNA than the DNeasy kit (Deiner et al., 2015; Renshaw et al., 2015). However, liquid phase separation methods are difficult to apply widely due to the utilization of toxic substances such as phenol and chloroform (Tsuji et al., 2019).

2.3 Primer selection and PCR amplification

Species-specific detection is carried out the basis of PCR techniques that use specific primers to amplify and detect short segments of DNA (typically 80–200 bp) of the target species (Bohmann et al., 2014). eDNA was first applied to large organisms using specific primers that amplified only one species during PCR (Jerde et al., 2011; P. F. Thomsen, J. Kielgast, L. L. Iversen, C. Wiuf, et al., 2012). Advances in sequencing technology have made it possible to assess the community composition using universal primers that identify multiple species in a sample; universal primers are required to amplify all species in a target taxonomic unit with a minimal bias to avoid preferential amplification of some species with mismatched primer regions (Ficetola et al., 2010). The mitochondrial genes 12S and Cytochrome b and the mitochondrion’s D-loop region are commonly used in amphibian studies, and the 12S region is often used for metabarcoding eDNA analysis (Table 1) (Li et al., 2021; Wynne E Moss et al., 2022; Sasso et al., 2017). For example, Valentini et al. designed a pair of primers specific for amphibians and scleractinians (batra_F, 5’-ACACCGCCCGTCACCCT-3,’ batra_R, 5’-GTAYACTTACCATGTTACGACTT-3’) and a human DNA blocking primer (batra_blk, 5’- TCACCCTCCTCAAGTATACTTCA-AAGGCA-SPC3I-3’); this primer pair can amplify up to 99% of amphibian species (Valentini et al., 2016). In a previous study, we employed GF filtration membranes in conjunction with the “batra” primers to investigate the amphibian diversity within China’s Chao Lake(Unpublished data). This approach not only facilitated the successful amplification of all species identified through traditional sampling methods but also unveiled two species that had previously eluded detection in historical records and conventional surveys. These outcomes substantiate the efficacy of the ’batra’ primers for detecting Chinese amphibians.
Because it is impossible to pinpoint organisms directly based on eDNA molecules exuded into water, nor to track the movement of specific eDNA molecules in water, the environmental samples collected often contain mixed DNA from many different organisms. When amplifying environmental samples, the specific primers chosen often play an important role in the successful amplification of the target species. Unsuitable primers have the potential to generate false positives, which can lead to overestimates of the presence or abundance of endangered species, or to false negatives, which can lead to insufficient monitoring of invasive species. In his study, Kelly pointed out that using a constant number of PCR cycles, different primers showed different biodiversity structures for the same simulated community (Kelly et al., 2019), which directly led to different analysis results. Wilcox tested the factors affecting the specificity and sensitivity of eDNA detection, and found that specificity was most affected by the mismatch of base pairs in the primers, which was much more prevalent than in the probe (Wilcox et al., 2013). Hence, it is evident that the meticulous selection of appropriate primers is of paramount importance.
\begin{equation} Cq=(\text{Cq}_{\text{samples}}-\text{Cq}_{\text{positive\ control}})\nonumber \\ \end{equation}
It has been demonstrated that diluting the sample alleviates the inhibition, but likewise dilutes the target DNA. Especially when the target DNA is present at very low concentrations, this may lead to failure of the assay (Goldberg et al., 2013; Anna M McKee et al., 2015; Strand et al., 2011; Tsai & Olson, 1992). McKee et al. examined this using a PCR inhibitor removal kit and found some alleviation of inhibition but also potential loss of target DNA (Anna M McKee et al., 2015).

2.4 Sequencing and data analysis

The PCR results are recorded as negative or positive. A negative result indicates that eDNA of the target species is not present in the environmental sample, while a positive result indicates its presence. This result is initially determined by GEP after primer amplification and then further verified by sequence comparison. In qPCR, a fluorescence threshold is set using the positive control as a benchmark; the sample is labelled positive when the amplification exceeds the fluorescence threshold, and negative when it does not (Venkatesan & Bashir, 2011). Metabarcoding uses universal PCR primers to amplify DNA in the environment. A large amount of data containing taxonomic information on genes can be obtained after high-throughput sequencing, which requires sequence comparison and species annotation (including noise removal and assignment of operational taxonomic units), clustering, and assignment of species annotation.
A complete and high quality reference database is the basis for the accuracy of amphibian surveys using eDNA technology (Alexander et al., 2020; West et al., 2022). This can be constructed by consulting established publicly available databases (e.g. NCBI: https://www.ncbi.nlm.nih.gov/) or by building custom versions. For example, Valentini et al. built a reference database containing 53 species of amphibians when using eDNA method to investigate amphibian diversity (Valentini et al., 2016). The accuracy of self-built databases is likely to be higher than that of public databases due to the significant geographic variation in amphibian fauna and the lack of assurance of the quality of DNA barcoding in publicly available databases.