Date of Award

Summer 2019

Project Type


Program or Major

Molecular and Evolutionary Systems Biology

Degree Name

Doctor of Philosophy

First Advisor

Jeffrey T Foster

Second Advisor

Jeff R Garnas

Third Advisor

Adirenne I Kovach


A food web is a model of the feeding relationships among organisms in an environment. The fidelity of this model is limited principally by the ability to detect these interactions. Researchers who study cryptic interactions such as nocturnal insectivory in bats typically rely on fecal samples to identify trophic connections. Historically these diet analyses were limited to morphological inspection of arthropod fragments, however modern metabarcoding techniques have improved the richness and specificity of consumed prey: rather than bats foraging for a few arthropod orders, we observe hundreds of species among guano samples. Animal metabarcoding is not without bias; nevertheless, a decade of improvements upon such biases have focused largely on molecular portions while bioinformatic considerations remain unresolved. When researchers use distinct software to perform their analyses—tools that have not yet been compared in animal metabarcoding studies—it is unclear if distinct perspectives between two experiments represent meaningful biological differences, or if they arise because of the alternative programs and parameters deployed. We investigated three fundamental bioinformatic tasks that impact a metabarcoding experiment: sequence processing, database construction, and classification (Chapter I). These comparisons offer guidance regarding which steps are most sensitive to parameterization and are therefore in need of optimizing for individual experiments, as well as highlight areas that are in need of critical improvement. We applied these bioinformatic lessons to a molecular diet analysis of Indiana bats, the first ever for this endangered species (Chapter II). While management decisions currently focus on protecting roosting habitat, our molecular analyses provide evidence that site-specific data is needed to more effectively inform conservation practices. For example, while these bats forage a broad swath of the arthropod community, the molecular data suggests they rely on particular aquatic habitats that are not currently protected. Finally, we investigated the diets of New Hampshire bats by collaborating with citizen scientist volunteers throughout the state to perform an extensive sampling regime in that spanned 20 locations over 2015 and 2016, and sequenced more than 900 guano samples (Chapter III). Molecular analysis of these data suggested these bats are foraging hundreds of arthropod species, including some turf and forest pests, demonstrating that our local bats provide ecosystem services. Individual diets varied across season and site, providing evidence of highly flexible and local foraging behaviors.