Large-scale DNA-based survey of frogs in Amazonia suggests a vast underestimation of species richness and endemism

Aim: Mapping Amazonian biodiversity accurately is a major challenge for integrated conservation strategies and to study its origins. However, species boundaries and their respective distribution are notoriously inaccurate in this region. Here, we generated a georeferenced database of short mtDNA sequences from Amazonian frogs, revised the species richness and the delimitation of bioregions of the Eastern Guiana Shield and estimated endemism within these bioregions. Location: Amazonia, with a focus on the Eastern Guiana Shield. Taxon studied: Amphibia: Anura. Methods: We used an extensive DNA-based sampling of anuran amphibians of Amazonia using next-generation sequencing to delimit Operational Taxonomic Units (OTU) and their distribution. We analysed this database to infer bioregions using Latent Dirichlet Allocation modelling. We then compared endemism within these bioregions based on our results and the current IUCN database, and inferred environmental variables that contributed the most to the biogeographic pattern. Results: The recognized anuran species richness within the focal area increased from 440 species currently listed by the IUCN Red List to as much as 876 OTUs with our dataset. We recovered eight bioregions, among which three lie within the Eastern Guiana Shield. We estimated that up to 82% of the OTUs found in this area are endemic, a figure three times higher than the previous estimate (28%). Environmental features related to seasonal precipitations are identified as playing an important role in shaping Amazonian amphibian bioregions. Main conclusions: Our results have major implications for defining future conservation priorities of this vast area given that endemism in most Amazonian bioregions is vastly underestimated, and might therefore hide a large portion of threatened species. Moreover, these findings raise concern about meta-analyses based on public databases within Amazonia dealing with poorly known groups