Trade‐offs between reducing complex terminology and producing accurate interpretations from environmental DNA: Comment on “Environmental DNA: What's behind the term?” by Pawlowski et al., (2020)

Abstract In a recent paper, “Environmental DNA: What's behind the term? Clarifying the terminology and recommendations for its future use in biomonitoring,” Pawlowski et al. argue that the term eDNA should be used to refer to the pool of DNA isolated from environmental samples, as opposed to only extra‐organismal DNA from macro‐organisms. We agree with this view. However, we are concerned that their proposed two‐level terminology specifying sampling environment and targeted taxa is overly simplistic and might hinder rather than improve clear communication about environmental DNA and its use in biomonitoring. This terminology is based on categories that are often difficult to assign and uninformative, and it overlooks a fundamental distinction within eDNA: the type of DNA (organismal or extra‐organismal) from which ecological interpretations are derived.


| EDNA S HOULD B E US ED TO REFER TO THE TOTAL P OOL OF DNA ISOL ATED FROM THE ENVIRONMENT
Clear and unambiguous scientific terminology is important to communicate science, particularly when misunderstanding or miscommunications can lead to costly ramifications (Gouran et al., 1986;Jerde, 2019;Mahon et al., 2013). Hence, we applaud Pawlowski et al. (2020) for highlighting inconsistencies in the use of the term "environmental DNA" (eDNA) and their implications for biomonitoring. As described by the authors, these inconsistencies stem from some researchers using the term to refer to any DNA collected from an environmental sample without first isolating targeted organisms (e.g., Stat et al. (2017)), while others use it to refer only to extraorganismal DNA released by macro-organisms into the environment (e.g., Fraija-Fernández et al. (2020)). Although some of us have previously advocated for eDNA to be defined as extra-organismal DNA, the value of which is effectively refuted by Pawlowski et al. (2020), we agree with Pawlowski et al. (2020) that environmental DNA should be defined in the broadest sense.
However, the recommendation to employ a standard two-level terminology in eDNA studies, first indicating the environmental origin of the DNA collected (e.g., water, sediment, biofilm, soil) and second indicating the taxa (e.g., fish, diatom, bacteria) targeted by polymerase chain reaction (PCR), does not align with the overall purpose of improving clarity in eDNA biomonitoring.
The reason is that it does not account for the distinction between the different types of eDNA (organismal and extra-organismal), which is the level of classification that can have a strong impact on eDNA data interpretation. While Pawlowski et al. (2020) discount this, we argue there is a need to be clear about the type of eDNA that is being evaluated in any given study and this is the reason for why the term has been described in the broad and narrow sense.

| EDNA IS COMP OS ED OF ORG ANIS MAL AND E X TR A-ORG ANIS MAL DNA
Environmental DNA can be classified into two types ( Figure 1a): organismal DNA and extra-organismal DNA, the latter also including extracellular DNA (Barnes & Turner, 2016;Bohmann et al., 2014;Taberlet et al., 2012;Torti et al., 2015). Organismal DNA is sourced from whole individuals most probably alive at the time

| EDNA C AN B E ENRI CHED FOR D IFFERENT SOURCE S AND T YPE S
Generally, not all DNA present in a studied environment is required to address a given research question or is used for an application, and successive steps of enrichment for specific types or sources of eDNA are usually applied. For example, eDNA from a large variety of taxonomic groups can be found as organismal or extra-organismal DNA (types) in the environment ( Figure 1a) and can be obtained in many ways from aquatic, aerial and terrestrial environments (Deiner et al., 2017). The first step is performed at the sampling level, where typically the collected material is passed through filters, meshes or nets to retain organisms, organismal debris or particles of a desired size ( Figure 1b). Notably, this step does not imply a separation of DNA types or taxonomic groups because different sources and types of DNA overlap in size ( Figure 1a) and because of the "sticky" nature of eDNA to bind other particles (Barnes et al., 2020).
A subsequent enrichment can be performed during laboratory work through PCR or sequence capture using taxon-specific primers or probes (Jensen et al., 2020). However, this step is not perfect; a fraction of nontarget taxa DNA can also be amplified, and target taxa DNA can be missed. Finally, DNA sequences from particular taxa can be selected at the analysis/interpretation step by considering only those sequences belonging to a given taxonomic group.
The particular methods applied at each of these enrichment steps will determine the final data set used for ecological inferences, but these methods evolve and are not in themselves completely deterministic. For example, "water eDNA amplified for metazoans" could refer either to organismal DNA collected through a plankton net containing fish larvae and zooplankton, or to extra-organismal DNA collected through a 0.45μm pore size filter containing tissue, scales or cellular debris from fish and zooplankton.

| ECOLOG I C AL INTERPRE TATI ON S S HOULD CONS IDER DNA T YPE
While it is currently impractical to separate and independently analyse organismal and extra-organismal DNA, the distinction between the two types is nonetheless crucial for ecological hypothesis-testing and data interpretation. Organismal DNA is often targeted when a living community of organisms is studied, asking questions about specific habitat, the functional role of communities or community assembly processes driven by abiotic factors and biotic interactions. Here, the chances of misleading data (i.e., the species was not in that environment at that time and place) are likely to be minimal. Instead, work focusing on extraorganismal DNA is more prone to misinterpretations about organismal distribution due to potential long-distance transport from source we hypothesize govern the transport between the temporal and spatial bounds of detected DNA and what inferences we can therefore make from its detection.

| CON CLUS IONS
In summary, we agree with Pawlowski et al. (2020) that eDNA should be defined in the broadest sense, but do not agree that the formal adoption of their additional proposed nomenclature will improve clarity in communication or reduce confusion around the use of the term eDNA. We suggest instead that scientists carefully and clearly identify the type of DNA being targeted for analysis (Figure 1) based on the existing terminology of organismal and extra-organismal DNA. This explicit stated intention would then clearly inform study design, sampling strategies, analytical choices and data interpretation to avoid potential biases and promote valid inferences. Because none of these choices and strategies are perfect in their detection of a particular type of DNA and in the place of a field-specific nomenclature, we suggest that in the methods sections of studies, authors should clearly describe the sampling strategy including the targeted size classes and taxa and whether taxa were targeted in any way during sampling, laboratory analysis (PCR, capture), data analysis (sequence selection) or some combination thereof. We feel that improvement of the field is a shared responsibility among researchers, reviewers, editors and managers and support the development and application of best practices in the acquisition and reporting of eDNA data (Goldberg et al., 2016) as the best way to improve clarity.