Evaluating the effects of sampling design, sequence coverage, and bioinformatic filtering toward an optimal protocol for documenting freshwater fish communities in northern temperate lakes.

Background/Question/Methods

For more than a decade, environmental DNA technology has successfully demonstrated its effectiveness as a tool to assess biodiversity. While methodological and technical improvements have significantly enhanced the field in the last years, factors such as sampling effort, survey designs, and sequence coverage needed for accurate biodiversity monitoring remain relatively little explored. In lotic ecosystems characterized by very low eDNA horizontal transport, sampling design should be tailored and coupled with optimal sequence coverage and data filtering to provide a realistic portray of biodiversity. Considering these aspects, our goal is to explore how to optimize our protocol for assessing biodiversity in such ecosystems. Toward this end, we sampled water for fish community eDNA analysis in a total of 12 freshwater lakes distributed from southwestern Quebec to eastern Ontario, Canada. The lakes were selected to encompass a wide range of surface sizes and fish species richness. We followed a grid-based sampling design, with cell size set proportionally to the lake area. We sampled water from a total of 558 sampling stations (25 to 50 per lake). Three mitochondrial DNA regions (12S rRNA; 16S rRNA; and cytb) were analyzed.

Results/Conclusions

We followed three steps: First, we evaluated each individual marker and marker combinations to identify which one(s) better capture(s) community composition of each lake. Second, using the best marker or marker combination, we evaluated the filtration parameters to provide the best estimation of species richness. Once the best marker and filtering parameters were selected, we evaluated different sub-sampling strategies to identify sampling methods that are both easily implemented in the field and provide optimal recovery of species diversity for a given sampling effort. For each of these objectives, we present what this extensive dataset can tell us about selecting optimal experimental, sampling, and analytical designs in the context of fish community studies in temperate lakes.