DNA metabarcoding is increasingly used for the assessment of aquatic communities, and numerous studies have investigated the consistency of this technique with traditional morpho‐taxonomic approaches. These individual studies have used DNA metabarcoding to assess diversity and community structure of aquatic organisms both in marine and freshwater systems globally over the last decade. However, a systematic analysis of the comparability and effectiveness of DNA‐based community assessment across all of these studies has hitherto been lacking. Here, we performed the first meta‐analysis of available studies comparing traditional methods and DNA metabarcoding to measure and assess biological diversity of key aquatic groups, including plankton, microphytobentos, macroinvertebrates, and fish. Across 215 data sets, we found that DNA metabarcoding provides richness estimates that are globally consistent to those obtained using traditional methods, both at local and regional scale. DNA metabarcoding also generates species inventories that are highly congruent with traditional methods for fish. Contrastingly, species inventories of plankton, microphytobenthos and macroinvertebrates obtained by DNA metabarcoding showed pronounced differences to traditional methods, missing some taxa but at the same time detecting otherwise overseen diversity. The method is generally sufficiently advanced to study the composition of fish communities and replace more invasive traditional methods. For smaller organisms, like macroinvertebrates, plankton and microphytobenthos, DNA metabarcoding may continue to give complementary rather than identical estimates compared to traditional approaches. Systematic and comparable data collection will increase the understanding of different aspects of this complementarity, and increase the effectiveness of the method and adequate interpretation of the results.

Anthropogenic activities are changing the state of ecosystems worldwide, affecting community composition and often resulting in loss of biodiversity. Rivers are among the most impacted ecosystems. Recording their current state with regular biomonitoring is important to assess the future trajectory of biodiversity. Traditional monitoring methods for ecological assessments are costly and time-intensive. Here, we compared monitoring of macroinvertebrates based on environmental DNA (eDNA) sampling with monitoring based on traditional kick-net sampling to assess biodiversity patterns at 92 river sites covering all major Swiss river catchments. From the kick-net community data, a biotic index (IBCH) based on 145 indicator taxa had been established. The index was matched by the taxonomically annotated eDNA data by using a machine learning approach. Our comparison of diversity patterns only uses the zero-radius Operational Taxonomic Units assigned to the indicator taxa. Overall, we found a strong congruence between both methods for the assessment of the total indicator community composition (gamma diversity). However, when assessing biodiversity at the site level (alpha diversity), the methods were less consistent and gave complementary data on composition. Specifically, environmental DNA retrieved significantly fewer indicator taxa per site than the kick-net approach. Importantly, however, the subsequent ecological classification of rivers based on the detected indicators resulted in similar biotic index scores for the kick-net and the eDNA data that was classified using a random forest approach. The majority of the predictions (72%) from the random forest classification resulted in the same river status categories as the kick-net approach. Thus, environmental DNA validly detected indicator communities and, combined with machine learning, provided reliable classifications of the ecological state of rivers. Overall, while environmental DNA gives complementary data on the macroinvertebrate community composition compared to the kick-net approach, the subsequently calculated indices for the ecological classification of river sites are nevertheless directly comparable and consistent.

Metabarcoding of environmental DNA (eDNA) is a powerful tool for describing biodiversity, such as finding keystone species or detecting invasive species in environmental samples. Continuous improvements in the method and the advances in sequencing platforms over the last decade have meant this approach is now widely used in biodiversity sciences and biomonitoring. For its general use, the method hinges on a correct identification of taxa. However, past studies have shown how this crucially depends on important decisions during sampling, sample processing, and subsequent handling of sequencing data. With no clear consensus as to the best practice, particularly the latter has led to varied bioinformatic approaches and recommendations for data preparation and taxonomic identification. In this study, using a large freshwater fish eDNA sequence dataset, we compared the frequently used zero‐radius Operational Taxonomic Unit (zOTU) approach of our raw reads and assigned it taxonomically (i) in combination with publicly available reference sequences (open databases) or (ii) with an OSU (Operational Sequence Units) database approach, using a curated database of reference sequences generated from specimen barcoding (closed database). We show both approaches gave comparable results for common species. However, the commonalities between the approaches decreased with read abundance and were thus less reliable and not comparable for rare species. The success of the zOTU approach depended on the suitability, rather than the size, of a reference database. Contrastingly, the OSU approach used reliable DNA sequences and thus often enabled species‐level identifications, yet this resolution decreased with the recent phylogenetic age of the species. We show the need to include target group coverage, outgroups and full taxonomic annotation in reference databases to avoid misleading annotations that can occur when using short amplicon sizes as commonly used in eDNA metabarcoding studies. Finally, we make general suggestions to improve the construction and use of reference databases for metabarcoding studies in the future.

Large tropical and subtropical rivers are among the most biodiverse ecosystems worldwide, but also suffer from high anthropogenic pressures. These rivers are hitherto subject to little or no routine biomonitoring, which would be essential for identification of conservation areas of high importance. Here, we use a single environmental DNA multi-site sampling campaign across the 200,000 km² Chao Phraya river basin, Thailand, to provide key information on fish diversity. We found a total of 108 fish taxa and identified key biodiversity patterns within the river network. By using hierarchical clustering, we grouped the fish communities of all sites across the catchment into distinct clusters. The clusters not only accurately matched the topology of the river network, but also revealed distinct groups of sites enabling informed conservation measures. Our study reveals novel opportunities of large-scale monitoring via eDNA to identify relevant areas within whole river catchments for conservation and habitat protection.

The current advances of environmental DNA (eDNA) bring profound changes to ecological monitoring and provide unique insights on the biological diversity of ecosystems. The very nature of eDNA data is challenging yet also revolutionizing how biological monitoring information is analysed. In particular, new metrics and approaches should take full advantage of the extent and detail of molecular data produced by genetic methods. In this perspective, machine learning algorithms are particularly promising as they can capture complex relationships between the multiple environmental pressures and the diversity of biological communities. We investigated the potential of a new generation of biomonitoring tools that implement machine‐learning techniques to fully exploit eDNA datasets. We trained a machine learning model to discriminate between reference and impacted communities of freshwater macroinvertebrates and assessed its performances using a large eDNA dataset collected at 64 standard federal monitoring sites across Switzerland. We show that a model trained on eDNA is significantly better than a naive model and performs similarly to a model trained on traditional data. Our proof‐of‐concept shows that such a combination of eDNA and machine learning approaches has the potential to complement or even replace traditional environmental monitoring, and could be scaled along temporal or spatial dimensions.