Integrating eDNA-inferred species diversity and amplicon-based phylogenetic alpha diversity to assess fish assemblages in the Araguaia River

Environmental DNA (eDNA) is an efficient, non-invasive approach for monitoring aquatic biodiversity, enabling species detection from genetic material released into the environment and offering greater scalability than traditional surveys. The Araguaia River basin, a biodiversity-rich system in Central Brazil within the Cerrado biome, has been increasingly exposed to anthropogenic pressures. Here, we applied eDNA metabarcoding to investigate fish taxonomic and phylogenetic alpha diversity, as well as phylogenetic community structure, in floodplain lakes along the middle Araguaia River. Water samples (1,500 mL per site) were collected from 132 lakes during the flood season (February 2023), filtered through 0.45 µm membranes, and processed for DNA extraction. The 12S rRNA gene was amplified using MiFish primers and sequenced on an Illumina MiSeq platform. Amplicon sequence variants (ASVs) were inferred using DADA2, taxonomically assigned with the MIDORI2 database, and used to reconstruct an amplicon-based phylogeny under a maximum-likelihood framework. A total of 1,402 fish ASVs were detected, of which 384 were assigned to species level, including 67 species previously recorded for the Araguaia basin and 45 not listed in occurrence databases. Assemblages were dominated by Characiformes, followed by Cichliformes. Species alpha diversity, estimated using Shannon and Simpson indices, exhibited pronounced spatial heterogeneity, with higher values in the southern portion of the study area. Phylogenetic diversity (Faith’s PD) showed a similar spatial pattern and was strongly correlated with species richness (p < 0.001). Phylogenetic structure was predominantly random across lakes, although some communities showed significant clustering or overdispersion. Net Relatedness Index values indicated phylogenetic clustering at deeper evolutionary levels, whereas Nearest Taxon Index values were centered around zero, suggesting variable assembly processes near the phylogenetic tips. Variance partitioning revealed that spatial processes related to hydrological connectivity and geographic structure explained most variation in taxonomic and phylogenetic diversity, while local environmental variables had limited influence.