Developing and Assessing an Environmental DNA Protocol for Detecting Amphibian Species in Lentic Systems in Alberta, Canada


Brandon K. Booker


Thesis submitted for Master of Science in Ecology at the University of Alberta. pp. 106; 2016


Global declines of amphibian populations have resulted in increased interest in determining and monitoring the geographic ranges for both endangered and invasive amphibian species. Monitoring programs rely on detection techniques to amass data, techniques that must be dependable in order to create effective conservation plans. Using trace DNA found in the environment (environmental DNA or eDNA) to detect a species may provide better detection rates than conventional detection techniques, such as call and visual surveys, and trapping. A variety of methods have been developed, assessed and compared for detecting eDNA in environmental samples. Determining the most reliable, efficient and cost effective techniques should culminate in the creation of standardized protocols, allowing for more reliable comparison across projects and species. In this study, I designed and compared the efficiency and reliability of multiple methods for collecting eDNA samples and for detecting eDNA for the 10 amphibian species that occur in Alberta, Canada.
I used an Ion Torrent PGM™, a high through-put sequencing detection platform, to investigate the efficacy of three types of primer sets: universal primer set (targets all 10 species), genetically grouped species primer set (targets four groups of one to four species), and species-specific primer set. To amplify eDNA and label it with an identifying tag, I compared a single reaction PCR approach (one-step PCR) with a two reaction PCR approach (two-step PCR). I tested the primer sets on water samples collected from 39 wetlands known to contain one to four species of amphibians. The species-specific primer set combined with the two-step PCR was the most effective for detecting target species with eDNA and produced the lowest rate of false positive detections.
I identified three ponds in central Alberta that contained western toad (Anaxyrus boreas), wood frog (Lithobates sylvaticus) and boreal chorus frog (Pseudacris maculata). At each pond, 10 water and 10 sediment samples were collected and processed for eDNA detections for these three species on three different dates. I investigated: 1) how 10-fold dilutions to reduce potential PCR inhibition influence detection rates; 2) the influence of processing samples with a high through-put sequencing and a quantitative PCR detection platforms on eDNA detection rates; 3) the influence of sample type, water versus sediment, on eDNA detection rates for each detection platform; and 4) how sampling at different times during the breeding and development season affects detection rates for both sampling methods and detection platforms. For each detection platform and sample collection method, detection rates for western toad and boreal chorus frog where low compared to wood frog. By comparing methods I determined that: 1) sample dilution to reduce potential PCR inhibition increased eDNA detections among samples that were negative prior to dilution and reduced eDNA detections among samples that were positive prior to dilution; 2) there was no significant difference in detection rates when samples were processed once with the qPCR and once with the ITPGM platform; however, analysis of samples processed in duplicate with the qPCR platform had a significantly higher detection rate than samples processed once with the ITPGM platform; 3) sediment samples provided higher detection rates than water samples for the ITPGM platform, but not for the qPCR platform; and 4) eDNA detection rates for both sampling methods and detection platforms changed seasonally, generally decreasing later in summer.
This study found that, regardless of the sample collection and processing method, eDNA detection rate varied across amphibian species that share a common habitat. Although I was able to show how detection rates differ across methodologies, I was unable to developing a reliable eDNA detection method for the 10 resident amphibians of Alberta.