Choice of markers and biosampler species
Our results emphasize the importance of combining several markers to enhance species detection and confidence in the results (Alberdi et al. 2019; Polanco et al. 2021; Tournayre et al. 2024). As expected,MiFish and Teleo showed substantial overlap, detecting more than 70% of species in common and yielding highly similar frequencies of occurrences in biosamplers.  Both markers identified nearly the same number of fish species, though 10 species were detected exclusively by one or the other. These unique species were found in a small number of digestive contents, suggesting that these differences likely reflect the rarity of these species in the environment, or at least their lower consumption by crustaceans, rather than variations in primer robustness or discriminatory power. The third marker tested (MG2 ) preferentially amplified invertebrates and detected significantly fewer fish species than the 12S rRNA markers. While it appears less effective for fish monitoring, it may still prove helpful for validating the identity of the biosampler species or to inventory other taxonomic groups (Tournayre et al., 2024). Although using multiple markers can increase the time and cost of molecular analyses, laboratory-processing time can be reduced by pooling stomach contents while ensuring reliable detection through large sequencing depth.
Another important practical consideration is the choice of natural samplers based on their abundance and distribution range. For neotropical rivers, Macrobrachium brasiliense and Macrobrachium olfersii which are widely distributed across South America and can be easily collected in large numbers, seem to be ideal candidates, collectively detecting 94% of the total dDNA fish diversity. Understanding the spatial ecology and feeding habits of potential biosamplers can help guide the selection of species (or species combinations) that will yield the most comprehensive inventory (Calvignac-Spencer et al., 2013). This knowledge can help assess the spatial and temporal resolution of the dietary data, which depends on the home range size and digestion rate of organisms (Feller 2006; Prog, Pihl, and Rosenberg 1984). In this context, an interesting perspective is to compare scavenger crustacean dDNA and the increasingly used water eDNA approaches to determine which one provides the best insight into local assemblages in large rivers (Cantera, Decotte, et al., 2022; Cantera et al., 2023b, 2023a; Coutant et al., 2023; Zinger et al., 2020). Indeed, while water eDNA provides a snapshot of biodiversity, dietary DNA may offer a more temporally integrated signal due to gut retention time and circumvents several issues commonly encountered with water eDNA in turpid river systems, such as high sediment loads that limit filtration capacity and high concentrations of organic matter that can cause PCR inhibition.