A few years ago at a conference I saw a presentation on environmental DNA, or eDNA, that got me very excited. Animals and plants shed DNA into their surrounding environment in the form of feces, urine and skin cells. This DNA can then be collected from environmental sources such as stream or pond water, sediment, and scat. DNA is extracted from the environmental material and tested for the presence of target species using species-specific molecular markers. However, at the end of the talk it was mentioned briefly that they had observed cross-amplification of DNA with other species, including humans! You can imagine the issues this creates when trying to identify the presence of target species. I have been seeing a lot of press lately about eDNA being used to detect rare stream amphibians and decided to look into the robustness of the method and if it was improved. In short, I think this is an exciting but still dangerous new method.
The detection of eDNA in aquatic environments is an exciting area with a wide range of applicability. Aquatic organisms secrete enough DNA to be detected in a water sample, despite dilution and slight degradation. DNA is the water becomes fully degraded after 7-21 days (see here). This allows for the detection of secretive or low abundance species (including some of my favorite amphibians, the hellbender and tailed frogs!) with higher probability of discovery than traditional monitoring. eDNA can also be used for early detection of invasive species in a watershed, including American bullfrogs and Asian carp.
So going back to my earlier reservations, what is the robustness of this method in terms of species specificity? Luckily there is no longer the issue of cross-amplification in humans, but what about closely related species? The use of qPCR (quantitative PCR) increases the sensitivity and specificity of the method, but does it completely solve issues of cross-amplification? An article that came out earlier this year by Wilcox et al. examines this question using closely related bull trout and brook trout (see here). This study found that false negatives (failing to detect trout even though they were present) when species abundance was low. False positives were also detected where one species was misidentified as another due to insufficient specificity of primers and probes. This highlights the need to test for primer specificity and amplification in other taxa, which is done routinely in eDNA studies. It seems intuitive to test against common local species, but what about non-anticipated species? Another big issue is the sensitivity of the method; qPCR can detect species presence with minimal amounts of DNA. This sensitivity makes the method highly susceptible to cross-contamination, which is more difficult to detect than cross-amplification.
The use of eDNA to detect species in aquatic environments is relatively new. However, it is being widely tested, adapted, and applied in many systems across the world (North America, The Netherlands, and France). Yes, this is a potentially a great tool for species detection. But, are we evaluating the precision of the technique sufficiently before implementation? Are we evaluating protocols intensively enough? After researching the state of eDNA, I am excited to learn cross-amplification in humans is no longer an issue! The field of eDNA for species detection is rapidly evolving with continual improvements in the methodology. But I will still be reading eDNA publications with the methods limitations in mind, and thinking carefully about the technique before applying it in my own research. I also can’t wait to see how next-generation sequencing improves the method!