Bee better: Leveraging multiple primer sets improves efficacy of eDNA metabarcoding to detect native bees

The global biodiversity crisis and concomitant “Insect Apocalypse” has manifested in an increasing number of pollinating species petitioned for or listed under the United States Endangered Species Act. Consequently, initiatives to protect and restore pollinator habitat and monitor biodiversity have become a conservation imperative. However, conventional pollinator monitoring methods are often labor intensive, require lethal collections, rely on dwindling taxonomic expertise, and are difficult to scale. Environmental (e)DNA methods have gained traction as a complement to conventional surveys. eDNA methods reveal species diversity by isolating genetic evidence deposited on environmental substrates (e.g. flowers). However, previous studies have had varying success in detecting native bees. Numerous hypotheses for why eDNA has been less successful have been proposed (e.g. low DNA deposition rates, primer biases, etc.). Here, we explored this issue in a living laboratory, sampling pollinator eDNA from 24 plant species at a native prairie plant nursery. Using eDNA metabarcoding, we deployed two primer sets used extensively in published arthropod eDNA studies, and a bumble bee-optimized primer set. We found that the Bombus primer set revealed greater bumble bee diversity, including two rare species conventional methods and other, more general primer sets failed to detect. These results suggest that it is unlikely that eDNA deposition by native bees is driving false negatives in eDNA metabarcoding studies. Rather, primer bias, perhaps due to mitogenome GC content, may drive false negatives from Bombus species. Ultimately, these results advance the efficacy of eDNA as a tool for pollinator biodiversity monitoring.