Strenghts and limitations of the dDNA metabarcoding approach
The dDNA approach offers several notable advantages over traditional methods used for monitoring tropical freshwater communities. First, it is cost-effective and easy to implement in the field, requiring minimum equipment. Using just a few baited traps and net hauls, we were able to collecte large numbers of shrimp (>30 specimens for M. olferssi and M. brasilience ) within just a few days along riverbanks and in rapid zones. Encapsulated in the stomach, dDNA samples can be easily preserved in ethanol with minimal contamination risk until extraction in the laboratory (Siegenthaler et al., 2019). This makes it especially suitable for tropical environments, where logistical challenges often hinder extensive sampling campaigns. This method avoids the capture and handling of fish, which is particularly relevant in protected areas or when monitoring threatened species. The shrimp used as biosamplers are highly abundant and occupy basal levels of food webs, meaning that collecting a few dozen individuals likely has a limited impact on population dynamic and ecosystem stability.
Another significant strength of this method is its remarkable efficiency and low selectivity. The metabarcoding analysis shrimp dDNA enabled the detection of as many species as an intensive 10-days deep inventory of the study area. This approach uncovered nearly three times more species than the 2022 WFD survey and exceeded the total number of species recorded over the 11 years of WFD campaigns. As expected, given the generalist and opportunistic feeding behaviours of these detritivorous organisms, dietary DNA analysis detected a broad spectrum of species in terms of size. Indeed, our analyses revealed no significant difference in the size ranges of species identified through using dDNA and those recorded during our deep inventory. This suggests that the dDNA metabarcoding of detrivorous crustaceans does not present any particular size-related detection bias, in contrast to WFD gillnet-based campaigns that preferentially captured large fish. Dietary DNA analysis also detected elusive species that live in specific microhabitats and are rarely captured when using traditional sampling methods. For instance,Cyphocharax biocellatus and Jupiaba abramoides are small, abundant species inhabiting shallow waters along riverbanks or vegetation in fast-flowing zones were detected using dDNA but never during WFD campaigns using gillnets. However, nearly 20% of the species from the deep inventory were absent from the shrimp dDNA samples. Some of theses species (e.g. Anchovia surinamensis or Moenkhausia aff grandisquamis ) are relatively common but undetected due to gaps in the reference barcoding databases. Other species are rarer, with low capture rates and increasing the sampling effort with additional biosampler specimens might help detect them. The use of easily accessible, generalist invertebrate species for large-scale monitoring has already been demonstrated in other ecosystems. For instance, dDNA analysis of brown shrimp (Crangon crangon ) detected twice as many species than traditional net surveys in coastal marine areas (Siegenthaler et al., 2019). Similarly, molecular gut content analysis of mussels has proven effective in monitoring planktonic communities, identifying a broad range of dietary taxa that mirror the diversity in surrounding water samples (Weber et al., 2023).
The dDNA approach is also notable for its high taxonomic resolution combined with its minimal reliance on taxonomic experts. Despite the small size of the DNA barcodes, we successfully identified most fish taxa at the species level, with only minor corrections needed, primarily due to recent taxonomic updates. This is a critical point because taxonomic identification can be challenging in highly diverse environments, especially when dealing with cryptic species complexes. However, this success was greatly supported by the well-documented fish fauna in French Guiana (Le Bail, 2012) and the availability of nearly exhaustive barcoding reference databases. Accurate taxonomic assignment through metabarcoding depends largely on the completeness and accuracy of reference databases (Hilário et al., 2023; Keck et al., 2022) and significant gaps still exist for tropical fish species (Marques et al., 2021; Sales et al., 2018). This underscores why the development of open-acess, highly-quality and complete reference databases has been identified as the top priority for the field (Blackman et al., 2024).