Research
Can fungi connect plants?
I used Sorghum and dark septate endophytes to test whether non-mycorrhizal fungi can form networks between plants. We found that DSE can connect neighboring plants and facilitate water movement between them, expanding the kinds of fungi we need to consider when thinking about common fungal networks.
Can a computer read 20,000 papers and tell me something useful?
I built machine-learning and natural language processing workflows to extract ecological information from the scientific literature. One project classified more than 20,000 abstracts to investigate the global distribution of fungal endophytes. Other work uses BioBERT to identify fungal trophic modes from scientific text.
What does a fungal symbiosis cost a tree under drought?
I studied native Populus hybrids across greenhouse and field experiments to ask how drought changes the relationship between trees and their root-associated fungi. We found that fungal communities and plant traits respond differently across cottonwood genotypes and water availability, highlighting the importance of belowground interactions when thinking about drought response.
What's hiding in a microbial collection?
At USDA-ARS, I work with microbial genomes and the NRRL culture collection, using phylogenomics and comparative genomics to improve the identification and characterization of microbial genetic resources. My current work spans groups including Streptomyces, Bacillus, Aspergillus, Penicillium, Trichoderma, and Fusarium.
Machine learning beyond biology
I've also worked on graph neural networks for predicting microbial community coalescence. The same framework led to some unexpected applications outside biology, including modeling soccer transfers and interactions in online games.
​
Paper upcoming :)
What is the environmental footprint of computational research?
People often use computers for their research. Using computers costs energy. So, how much energy does research cost?
​
We did a review of the literature published in the journal Environmental Data Science to assess how often researchers named the environmental impacts of their research. Then we made an R package to help researchers track environmental impact of their code in R.