Unveiling biogeographical patterns of the ichthyofauna in the Tuichi basin, a biodiversity hotspot in the Bolivian Amazon, using environmental DNA

local environmental conditions by damming (sediment and flow patterns modifications, per- manent flooding of large area) will likely further compromise the usefulness of the identified spawning area. It will profoundly modify the structure of fish communities and the fisheries that depend on them, as observed in other dammed rivers of the Amazon basin (review in Duponchelle et al., 2021 [87]). Contributions and limitations of the eDNA approach in the biogeographic study of the region. The eDNA approach permitted a study of fish biodiversity in the Tuichi sub-basin and broader Beni basin within and immediately adjacent to the Madidi NP with a high species diversity thanks to a high taxonomic resolution, since 93.3% of the assigned OTUs (5,246) are at the spe- cies level. A majority of OTUs (8,205) although aligned with at least one sequence in the COI fish database could not be assigned taxonomically due to a too low percentage of identity with the reference (<97%). This is closely related to the fact that of 2,406 known Amazonian fish species [2], only a third (694) have a COI barcode available. The effectiveness of the method would therefore be considerably enhanced if the reference databases used for taxonomic attri- bution (Bold, GenBank) were improved over time, thereby improving our species identifica- tion capabilities. Markers other than COI have been used to record fish biodiversity, such as a 12S marker in French Guiana [26]. But the smaller size of the latter (65 bp versus 185–285 bp for the COI bar-

codes used here) and an even more limited database (181 available 12S barcode in Cilleros et al., 2018 [26]) than that of the COI, makes this marker less efficient than the COI, with more difficulty in assigning OTUs below family or gender level. The combination of several markers [88] and the availability of more exhaustive reference databases would allow a more accurate appreciation of species diversity. The eDNA approach, already widely used for species invento- ries, also seems to be a promising tool for localised studies to assess the impacts of anthropo- genic disturbance on biodiversity, such as dams, factories or mines, but also for carrying out a rapid census of biodiversity on more extensive regional scales. Ecological and conservation studies will require the complementarity of traditional sampling approaches to provide infor- mation on recruitment, age, size and precise abundance, as well as to identify the specimens needed to complete the DNA reference databases. Nevertheless, eDNA approaches to access large-scale species inventories in a rapid and non-destructive manner can complement these studies. Quantitative approaches based on eDNA sampling are being developed using qPCR or digital PCR techniques, in order to gain access to relative biomasses [89, 90]. Although many technical or biological biases still need to be overcome, access to biomass would provide infor- mation on the demographic patterns of species required to put in place conservation strategies essential for sustainable ecosystem management. Fishes are among the freshwater biota most

threatened by global changes [91, 92]. The evolution of fish biodiversity in the Tuichi study area could be monitored over several years in relation to climate change and dynamic anthro- pogenic threats in the region, and serve as a reference for the study of other river basins in the Amazon. Supporting information S1 Fig. Rarefaction curves at library and site level. (PDF) S2 Fig. Frequency spectrum of taxa. (PDF) S3 Fig. Taxonomic diversity at order level per site. (PDF) PLOS ONE Fish diversity and distribution in the Tuichi basin PLOS ONE | https://doi.org/10.1371/journal.pone.0262357 January 4, 2022 19 / 25