5: Using DNA metabarcoding techniques to improve plant-pollinator interaction networks

Plant-pollinator networks have typically been composed using insect-visitation data that may not accurately depict true pollination events. Recently, there has been an increase in the number of networks created via insect pollen-load samples. These networks are often determined through light microscopy identification, requiring specialised knowledge and reference collections; yet even with these, lower levels of identification can be hard to achieve. DNA metabarcoding has been used as an alternative to identifying species-specific pollen samples from honey, hives and insects. However, this novel approach has not yet been used to test ecologically driven questions, like how varying landscape factors may be altering resource collection patterns and floral fidelity in pollinating insects. We use DNA metabarcoding and the newly developed DADA2 bioinformatics pipeline to assess the impacts of a mass-flowering crop, Brassica napus, on the floral fidelity of pollinating insects in wildflower communities adjacent to these fields. We collected 120 pollinating insects from remnant patches of York Gum-Jam woodland wildflower communities in SW Western Australia that were varying distances from the edge of a blooming canola field. Using DNA metabarcoding we identified the individual species composition of insect pollen-loads and created pollination networks and pollen-load composition analyses to determine how increased proximity to canola crops alters pollen-load consistency. With this research, we highlight the practical applications of using this technique and the potential for its use in ensuring improved accuracy for plant-pollinator network analyses and pollen-load analyses that will give valuable insights towards answering central questions in ecology.