Hide and seek: species detectability in a biodiversity hotspot increased by environmental DNA - Supplementary Material

Data on species distribution is fundamental for conservation planning. However, acquisition of high-quality data demands great sampling effort and time, especially in biodiverse regions. Furthermore, species detectability is frequently not considered during species inventories based on conventional methods and this could potentially compromise biodiversity assessments when detectability is low. In this context, environmental DNA (eDNA) metabarcoding is a new promising method to study species distribution and community composition though the detection of DNA left by animals in environmental samples (e.g. water). We aimed to compare the detectability of eDNA metabarcoding in comparison to conventional field inventory methods in tropical, species-rich frog communities with previously documented field-inventory-based low detectability for most species (Serra do Cipó, southeasern Brazil). Field inventories recorded 23 species, whereas eDNA sequences were successfully assigned to 18 taxa (14 to species level), this number being limited by the incomplete reference library (only 19 out of 40 species in the study site were represented). Among 11 species detected by both methods, eDNA metabarcoding provided higher detectability for four species and similar performance for the remaining ones. eDNA metabarcoding also detected three species not observed in field surveys. Despite the incomplete sequence database, in seven out of ten streams eDNA metabarcoding detected more species than field inventories assessing either adult individuals or all life stages. We strongly recommend that eDNA surveys are performed in combination with sequencing efforts to include missing species into reference libraries. It is also important to further investigate variables likely to influence detectability (e.g., inhibitors, DNA transport and degradation, effects of life history and behavior on DNA release) to improve the efficacy of this promising method and speed acquisition of species distribution data in the conservation race against the pace of human impacts on the planet.