Environmental DNA metabarcoding for monitoring fish biodiversity in remote lakes

Abstract Objective Environmental DNA (eDNA) metabarcoding provides an attractive option for monitoring biodiversity in remote freshwater ecosystems, where the deployment of conventional gears encounters major logistical constraints. We evaluated eDNA metabarcoding for monitoring fish communities and early detection of nonnative species in three remote lakes on Isle Royale, Michigan, USA. Methods At each of the three lakes, we collected surface, midwater, and lake bottom samples from 10 sites during spring and fall sampling events. We performed metabarcoding on all the water samples, targeting the 12S region of all fish species. Results Despite a relatively small sample size (N = 60 samples per lake across two visits; 10 locations with three depths per location), we recovered 70% of all the species that were previously observed using conventional methods. We recovered several detections of putative Cisco Coregonus artedi, a vulnerable coldwater species, providing evidence that Cisco have persisted in these lakes. However, we found disentangling likely false positives from rare species challenging, which we overcame by employing multiple types of detection thresholds and a species-specific quantitative PCR assay. Conclusions Although we were able to successfully characterize the fish communities using eDNA metabarcoding, more attention needs to be given to the detection thresholds and communication protocols that provide guidance in interpretating new eDNA detections and using eDNA detections to inform management decisions. Although eDNA metabarcoding has limitations that should be accounted for at the outset of the project, the ease of sample collection makes eDNA metabarcoding an option for monitoring freshwater biodiversity in remote systems.