Insights into the representativeness of biodiversity assessment in large reservoir through eDNA metabarcoding

than across stations (Fig 4B). For instance, the C12 station presented richness in one of the replicates that was almost two times greater than the other two replicates (12 compared to five and three). Stations C01, C02, and C03 exhibited the greatest richness, whereas the highest Shannon’s H values were observed in stations C03, C01, and C04, respectively. A quarter of the sampling effort was sufficient to reach the biodiversity plateau of the TI reservoir A PCA was built with only the fish MOTUs identified to the species level. The individual posi- tion in the graph indicated a high similarity in composition among the samples, except for those from station C01 (C01R1, C01R2, and C01R3) and a few from stations C02 (C02R3), C03 (C03R1) and C07 (C07R2) (Fig 5A). These samples were collected at stations considerably distant from all others, potentially suggesting the presence of species exclusive to these sites. In contrast, no distinct clustering pattern was observed for all other sites, indicating that despite the pilot sampling being conducted one year earlier, the composition remained very similar, preventing the segregation of these samples in the ordination space. However, the maximization of differences between these groups reinforces the pattern observed in the PCA, Fig 5. Beta diversity results for ichthyofauna taxa detected in the Três Irmãos (TI) reservoir. Based on fish Molecular Operational Taxonomic Units (MOTUs)

identified through eDNA per sampling station, A) Principal Component Analysis (PCA) illustrates the variance in community composition. Each dot represents a sample from a station, with colors designating stations, as in the legend. Stations abbreviations as in S1 Table. Samples plotted close to each other are zoomed in the minor graph on the left side. B) Discriminant Analysis of Principal Components (DAPC) exhibits the maximization of differences among groups (6 PCs, K = 5 according to the BIC method). The colors correspond to those in PCA. C) Boxplot shows the contribution of each station and their samples to the segregation into five clusters. D) Assign plot illustrates each individual sample’s assignment, considering K = 5. Darker shades of red indicate a higher chance of the sample belonging to that cluster. https://doi.org/10.1371/journal.pone.0314210.g005 PLOS ONE eDNA metabarcoding reservoir PLOS ONE | https://doi.org/10.1371/journal.pone.0314210 January 24, 2025 12 / 22