In this guest blog, Dr Cecilia Villacorta-Rath, who leads the TropWATER Environmental DNA Laboratory, introduces how researchers make sense of genetic traces that species leave behind.
Environmental DNA (eDNA) is an increasingly important tool in ecology, conservation, and biosecurity. It allows us to infer species presence from traces of DNA left in water, soil, and other environmental samples and it is sensitive to small or cryptic populations, making fieldwork faster, and non-invasive.
The most important thing is that it allows for non-specialist engagement because anyone can collect a water sample, even school kids. At the same time, the field is evolving quickly as researchers refine sampling designs, laboratory methods, reference databases, and the way results are interpreted.
Unlike traditional ecological surveys – where organisms are trapped, recorded, or directly observed – eDNA analysis detects DNA and not the organisms themselves. Each detection therefore must be interpreted carefully.
Last month, we invited your eDNA questions on social media; here, we explore what eDNA signals can, and cannot, tell us.
How do you make assumptions on how recently individuals from a target species were in the vicinity of a water sample taken from a flowing river?
The first thing to remember is that eDNA methods detect DNA, not organisms. In a flowing river, DNA may be transported downstream, degraded at different rates, or moved indirectly by another animal. This means that a detection does not prove that the source organism was at the sampling point or reveal exactly when it was nearby. For example, our research has shown that eDNA can been detected as far as 22 kilometres downstream from a known population.
Recency can be investigated through signals that break down relatively quickly, such as environmental RNA or longer DNA fragments, but these still need to be calibrated for the river system and environmental conditions. Water discharge in freshwater systems (and currents and tides in marine environments) can carry the eDNA signal well beyond where it was deposited. These processes need to be considered when planning a survey.
If we need a more localised detection in a flowing river, samples should be collected near the end of the dry season when flow is minimal. The strongest inference comes from repeated sampling across locations and times.
This should be interpreted alongside other factors like flow, distance, temperature, DNA decay, and the target species’ ecology.
Not all animals shed skin, scales, or other cells as quickly as others. How do you account for high- or low-volume shedding rates when measuring a signal?
Shedding rate is one of several factors that affect whether, and how strongly, a species is detected. Among other more technical factors, a strong DNA signal may reflect a high-shedding species rather than a larger population while a low-shedding species may be present but missed.
For that reason, eDNA results generally should not be treated as a direct measure of the number of individuals present.
We reduce this uncertainty through a robust sampling design: increasing the number of repeated samples across space and time, collecting consistent sample volumes, sampling at comparable sites and times, and knowing the species’ ecology and behaviour to inform sampling. Where quantitative interpretation is needed, the relationship between the eDNA signal and the target species must be calibrated and validated for that particular species and system.
What are some species that can’t currently be targeted using eDNA but could be a valuable focus of eDNA research in the future?
The greatest gaps are often not species that are inherently impossible to detect, but species where our reference databases and our knowledge of the factors behind detection or non-detection are incomplete.
Undescribed organisms cannot currently be identified reliably to species level because there is no verified reference sequence to match. Those organisms might only be classified down to Class, Order, Family or Genus, depending on the current state of the reference databases.
Known species may also be missed when reference databases are incomplete, meaning that it does not include the genetic makeup of species from a particular geographical area. This would mostly happen in targeted eDNA studies and when there is a strong level of genetic differences between populations across a species’ range.
Low-shedding, rare, or terrestrial species are further challenges, especially when it’s not clear what substrate should be sampled (such as water or soil) for the best chance of detection.
Future research should prioritise filling reference databases and developing species-specific assays following international best-practice guidelines. Future work should also focus on testing the parameters behind detecting species from substrates that are now being considered for eDNA studies such as air, spider webs, and traces collected from footprints, nests, burrows, or trees.
