The importance of conserving invertebrate diversity is becoming increasingly appreciated, and with recent declines of some pollinators, including native bees, many land managers and producers are looking to implement restoration plans that enhance native bee habitat. However, there is limited information on which plant species are important food sources for native bees. DNA metabarcoding techniques can be used on pollen loads to rapidly identify the plant species that native bees visit, with the potential to reveal a more detailed record of foraging than traditional methods. In this exploratory study, we isolated pollen from four different species of flowering plants and from three foraging native bees. Our main questions were 1) Can quantitative results be obtained from samples containing pollen mixtures of known quantities? and 2) Can DNA metabarcoding detect rare plant species in a pollen sample? We created three artificial pollen mixtures of known concentrations from the four flowering plant species: even mixed, highly skewed, and slightly skewed. We also created pure pollen samples from each different flowering plant species. We had 16 samples sequenced at the Center for Genome Research and Biocomputing (CGRB) in Corvallis, OR: four pure pollen, three even mixtures, three highly skewed mixtures, three slightly skewed mixtures, and three native bee pollen loads.
Results/Conclusions
We were able to detect three out of four expected species in our single species samples, but other species also appeared. It is unknown whether this is from contamination in the lab or from foraging bees moving pollen from one plant to another in the field. We were able to detect most of the plant species in our mixtures, and we were able to detect rare plant species in the highly skewed mixtures. Our evenly mixed samples do not show an even abundance of sequence reads from each plant species, ruling out any quantitative interpretation. With regard to future work, great care must be taken to prevent cross-pollen contamination in the field and in the laboratory and creating a reference library of plant species from the study area of interest will reduce the possibility of mistakes during taxonomic assignment. Sequence read data from DNA metabarcoding of pollen loads should be interpreted as presence-absence data. DNA metabarcoding can detect rare plant species in a pollen sample and has great potential to detect rare plant species in an insect pollen load.