Conclusion
Our study demonstrates that a portfolio of metabarcoding markers provides sufficient taxonomic coverage to detect and identify all of the >100 marine and freshwater taxa collected worldwide and combined here into heterogeneous and complex mixtures. Although multi-locus metabarcoding has become increasingly common for biodiversity assessments, where studies rely on different loci (COI, 18S, 16S, 12S) to resolve different taxonomic groups, our focus was to employ multiple markers to provide greater taxonomic coverage and increased taxonomic resolution across a single group: fishes (and tangentially, other aquatic taxa). We recommend that investigators seek to optimize their own portfolios of markers suited to the taxonomic and geographic scope of their work, but our findings demonstrate that this approach can enable powerful inferences about biodiversity. Unfortunately, in order to achieve quantitative inferences about relative abundance of each species in a complex mixture, we must recommend further exploration of capture-based approaches (e.g., Mariac et al., 2018) or leveraging genome-wide data (e.g., skim-seq; Bohmann, Mirarab, Bafna, & Gilbert, 2020; Chua et al., 2021; Kobus et al., 2020). Accurately detecting and then quantitatively accounting for the relative abundance species in complex mixtures is a critical next step for ecological and forensic studies using eDNA.