Discussion:
We detected Bd DNA in all positive frog skin swabs collected across multiple sampling locations and amphibian species using rapid in-the-field methods. Mobile handheld real-time PCR thermocyclers are promising tools for rapidly detecting Bd in susceptible amphibian populations. Previously, to detect Bd in individuals, amphibian skin swabs had to be stored to prevent DNA degradation, transferred out of field, and processed in a lab. The difficulty of collecting and processing swabs in a lab is compounded by species that inhabit remote wilderness or otherwise difficult to access locations. The rapid, in-situ method we applied yielded accurate results for all amphibians swabbed in less than 60 minutes and did not require the transfer of samples out of the field for lab analysis. Rapid detection of Bd is critical to predict and/or minimize epizootic outbreaks (e.g., mass die-off events) and initiating intervention / treatment options such as salvaging animals for captive rearing efforts or on site anti-fungal treatments (Harris et al., 2009). The ability to detect Bd presence in a population in less than 60 min will significantly aid to the recovery of MYLF and other Bd susceptible amphibians where high loads are expected if Bd is present.
We also detected Bd DNA in eDNA samples at the same number of sites using the rapid in-the-field method compared to traditional lab methods. We did not detect Bd DNA for Site 3 using field analysis and we did not detect Bd DNA for Site 2 using lab analysis (Table 1); swabs from both sites tested positive as swabs samples have a higher probability of detecting Bd DNA than eDNA samples (Fig. 4). The false negative result in both the lab and field methods suggest that a more sensitive eDNA surveillance strategy should be used (i.e. increase number and quantity of water samples collected and/or increase number of technical replicates), as generally recommended for eDNA surveys (Goldberg et al., 2016).
We detected Bd DNA in more technical replicates for the samples processed in the lab compared to samples processed in the field, suggesting that our field-based methods may not be as sensitive. Approximately 1.5 technical replicates would have to be analyzed using the field protocol in order to have the same mean detection probability of 1 technical replicate using traditional lab protocol. Our findings are consistent with Sepulveda et al. (2018) who found lower detection of northern pike in eDNA samples and subsamples processed with field protocol compared to traditional lab techniques. As a result, the field extraction approach failed to detect DNA in areas collected with low densities of northern pike (Sepulveda et al., 2018). Additionally, we used a liberal positive criterion where a sample was considered positive if ≥ 1 technical replicate detected Bd DNA. A more conservative approach where a sample is considered positive if ≥2 or 3 technical replicates detects the target DNA, as is typically applied to eDNA analysis, would likely increase false negative rates of the field-based protocol. eDNA is typically used as a surveillance tool for rare or elusive species (Rees et al., 2014) or for early detection/ monitoring for invasive species (Jerde et al., 2011; Hunter et al., 2015; Kamoroff et al., 2019); being able to detect trace amount of DNA in a water sample is critical for successful use of eDNA techniques. Prior to the use of rapid field techniques, further assessment should be made to ensure eDNA samples (or other low-quality DNA samples) can be detected at low quantities, and to assess false positive rate at the technical replicate level.
All quantities of DNA detected using rapid in-the-field techniques were below the standard curve, further evidence in the field methods lack sensitivity. Binary detection (i.e. presence/absence) of Bd DNA is an important metric for understanding disease dynamics and host risk. However, DNA quantification of both eDNA and swab samples is critical to the ecological interpretation of the results. Vredenburg et al. (2010) found Bd prevalence increased rapidly and infection intensity increased exponentially with declines of MYLF evident after average infection intensity of ~10,000 zoospores swab-1. Determining when Bd levels and infection intensities rapid/ exponential grown before lethal threshold levels is critical for management to implement conservation strategies. Such determination can only be accomplished through accurate quantification of Bd load on skin swabs and potentially eDNA samples.
Bd detection for conservation and management projects needs to be reliable as well as able to meet budget and time constraints. Typical costs for lab extraction and analysis of swabs and eDNA samples are ~$10-35 and ~$50-150 respectively depending on type of lab, extraction method, and number of samples processed. Typical qPCR machines used in lab analysis have a 96 well capacity and can multiplex up to five targets per well resulting in a high-volume throughput per run. The M1 sample prep kit for field-based DNA and eDNA extraction cost was $15 per sample and the custom Go-Strips™ were $10 per well (Biomeme inc, Philadelphia PA.). Total cost of analysis using the Biomeme field methods is $45 for both eDNA and DNA extractions run in triplicate wells. The two3™ mobile real-time PCR machine has a three well capacity and can multiplex two targets per well (the target species and an internal positive control). The limited wells inherent to a small handheld qPCR machine will take much longer to run a large number of samples compared to a lab-based machine. As a result, for projects requiring high numbers of samples, lab-based extraction and analysis may be more cost and time efficient.