What Is Environmental DNA? Finding Species Without Finding the Animal

The workflow is standardised and each step introduces its own difficulties:

• •Collection, typically filtering a volume of water through a fine membrane that captures cells and free DNA, with contamination control being the dominant practical concern throughout
• •Extraction of DNA from the filter, using methods that must recover material that is frequently degraded and present in tiny quantities
• •Amplification by polymerase chain reaction using primers targeting a region of DNA that varies between species while being flanked by sequences common to a whole group
• •Sequencing the resulting mixture, which produces enormous numbers of short reads
• •Matching those reads against a reference database of known sequences to assign them to species
• •Interpretation, which is where the real limits lie, since a match tells you the DNA was present and not how much, when, or whether the animal was alive

The method's advantages are specific and substantial. It detects rare and cryptic species that conventional survey misses, and it is now the standard method for monitoring great crested newts in Britain, where a water sample replaces repeated night visits with torches. It detects invasive species early, when populations are too small to find by looking, which is when eradication is still possible, and it is used at ports and in waterways for that purpose. It is non-invasive, so nothing is caught, handled or killed, which matters for protected species and removes the licensing burden. It is fast and cheap per site once established, allowing far more locations to be surveyed than field teams could cover. And metabarcoding, amplifying a region shared across a whole taxonomic group, returns a list of everything in that group from a single sample, so one bottle of water can indicate which fish are in a river rather than requiring a targeted test for each.

Several limitations are inherent rather than technical. Abundance is not reliably measured, since the quantity of DNA depends on body size, shedding rate, behaviour and recent activity as well as on numbers, so more DNA does not straightforwardly mean more animals. Location is uncertain, because DNA moves: in a river it travels downstream for a distance depending on flow and on how fast it degrades, so a detection indicates presence somewhere upstream rather than at the sampling point. Timing is uncertain, since DNA persists for hours to weeks depending on temperature, ultraviolet exposure and microbial activity, so a detection means recently rather than now. Reference databases are incomplete, so a sequence from an unrecorded species returns no match or the wrong one, which biases results towards well-studied groups. Contamination produces false positives easily, and the sensitivity that makes the method valuable is what makes a trace of contaminating DNA sufficient to ruin a result.

The technique has expanded well beyond water. Soil samples reveal fungal and invertebrate communities. Air sampling has been demonstrated to detect terrestrial vertebrates, with studies collecting airborne DNA in zoos and identifying the species housed there, which opens the possibility of surveying land animals from a filter on a pole. Sampling water from a flower can reveal which insects visited it. Ancient environmental DNA extracted from permafrost and lake sediment cores has reconstructed past ecosystems, including a notable study recovering a two-million-year-old community from Greenland sediments, which is far older than any preserved DNA from a body. The direction of travel is towards routine automated monitoring, with samplers deployed continuously and sequencing done locally by portable devices, and towards using the data for regulatory purposes, which requires standardisation and validation that the field is currently building because a method used in enforcement must be defensible in a way an exploratory one need not.

Organisms shed cells constantly, so filtering water and sequencing the DNA reveals which species were recently present without catching anything, which has made it the standard method for detecting rare species and early invasions. Metabarcoding returns a whole group from one sample. It cannot reliably measure abundance, DNA travels downstream so location is approximate, persistence means recently rather than now, and incomplete reference databases and contamination are the main error sources.