What a coastal eDNA survey can—and cannot—tell you

Start with a question the water sample can answer

Environmental DNA, or eDNA, is genetic material that organisms leave in water as cells, mucus, waste, or other biological traces. A filtered water sample can therefore carry evidence of organisms in or connected to a water mass, even when no animal is seen. A metabarcoding survey uses genetic markers and sequencing to detect multiple groups in the sample; a targeted assay asks about selected DNA targets. Neither turns a bottle into a complete census. The useful first decision is the scale of the question: a particular target, a habitat comparison, or a broad inventory each calls for different markers and sampling effort. [1]

Record place, water layer, time, habitat, and question before collection. A signal may move with currents, and surface water is not automatically representative of the bottom. Define the sampled water and treat detected DNA as evidence connected to that sample, not a precise map of where an animal stood.

Replicates answer different questions

A field replicate is a separate water sample collected at a site. It helps reveal whether detections recur across small differences in where water was collected. A laboratory or PCR replicate repeats processing from extracted material and tests another source of variation. More technical repeats cannot fully substitute for collecting more independent water samples or covering more of the habitat. Keep those replication levels labeled separately in the data sheet. [2] [3]

A nearshore study around Juneau collected nine one-liter samples for each site-and-tide combination and included field negative controls. In that particular design, habitat explained important differences in fish detections; the effect of tide was less consistent and varied with habitat. Some rarer detections were sporadic across PCR replicates. Those results show why replicate strategy should follow the survey question, but they are not a universal sample-count rule: the work covered nine sites in one region and a limited sampling period. [2]

For a two-habitat comparison, collect separate samples across both habitats, record depth and tide, and process field controls alongside them. Set sample numbers from expected variability, study scale, and resources, ideally with a pilot. One convenient sample describes its contents; it cannot establish the site’s full community.

Controls make contamination visible

DNA is easy to move unintentionally on hands, containers, filters, or equipment. Clean handling and decontamination reduce this risk, while a field negative control—clean water opened and handled alongside samples—can help identify contamination introduced during collection or filtration. Extraction and PCR controls address later laboratory steps. A control is informative only if its handling and results are reported. The NOAA-hosted marine sampling guide emphasizes fit-for-purpose protocols, rigorous decontamination, and pilot work because no single handling recipe suits every habitat and project. [1]

Record sample ID, site and water layer, collection time, bottle or filter ID, preservation method, and deviations. If a negative control contains a target signal, investigate under predeclared quality rules. Controls do not eliminate every source of error, but they make a key uncertainty inspectable.

A sequence match is not the same as a confirmed sighting

After sequencing, researchers compare recovered DNA sequences with reference databases. The match may support an identification only as specifically as the marker and reference coverage allow; some sequences cannot be resolved to species, or may match multiple candidates. In a coastal metabarcoding analysis, marker performance differed across taxonomic groups, and the authors estimated that some species were missed even after sampling. They cautioned that the study’s estimates should not be generalized beyond its limited sites and design. [3]

A positive result supports the statement that target-like DNA was detected in the processed sample under the stated method and quality criteria. It does not alone establish that a live individual was at the collection point at that moment, how many individuals were present, or whether the species is resident. Transport, DNA persistence, assay performance, and contamination all affect interpretation. Pairing eDNA with visual, acoustic, or conventional sampling can answer different parts of the question; it should be described as complementary evidence rather than an automatic validation step.

Report the limits with the result

Preserve the design and uncertainty in the record: location and habitat; water layer; date, time, and tide context; volume and filtration; preservation and processing; controls; marker or assay; database and identification threshold; replicate structure; and non-detections. State whether the output is targeted detection or a community metabarcoding list. Without these details, comparisons are difficult.

The central lesson is modest but powerful: eDNA can extend observation into water that is hard to survey directly, while remaining a sample-based measurement with imperfect detection and spatial context. A careful report tells readers what material was sampled, what the assay detected, how identification was assigned, and what remains unresolved. That discipline lets molecular evidence inform coastal ecology without asking it to prove more than the design can support.