Meet Ryan Bartelme: Digital Agriculture Science Analyst

Ryan Bartelme joined the CyVerse Science Analyst team in December, 2020. His projects include leading a National Institute of Food and Agriculture (NIFA) High-Intensity Phenotyping Systems grant, working with Agricultural Genome to Phenome Initiative (AG2PI) on a U.S. Department of Agriculture (USDA) grant to improve communication between biologists, data scientists, and engineers, contributing to publications, and applying for grants. He develops software packages for research and automated home gardening systems on the side.

Image

Bartelme is helping coordinate a future workshop with the USDA and AG2PI, introducing basic computational programming for biologists doing crop phenotyping, and basic biology for computer scientists and informaticians. The goal is to facilitate future project management and communications between biologists and computer scientists working on crop resilience.

Also this spring, Bartelme will co-teach the CyVerse Foundational Open Science Skills course. FOSS Online is a 10-week virtual workshop teaching the principles, practices, and how-tos for doing collaborative open science using cutting-edge, open-source cyberinfrastructure.

Bartelme earned a Bachelor's degree in microbiology from the University of Wisconsin-Madison, where he studied pathogenic fungi. He completed his PhD in freshwater sciences at the University of Wisconsin-Milwaukee, examining high-density microbial community structures in aquaculture and aquaponic systems.

Bartelme came to the University of Arizona in 2018 as a postdoctoral researcher with Paul Carini in Environmental Science, studying soil bacteria cultivation and bacterial stress response. Bartelme later transferred to the lab of Bryan Heidorn in the School of Information, where he developed methods of predicting plant phenotypes from genomic data and environmental variables.

In 2020 he received a Data Science Fellowship through UArizona Health Sciences, hosted by CyVerse and the UArizona Data Science Institute. During his fellowship he worked in an area he calls the "front end of healthcare," specializing in metagenomics and digital health under the mentorship of Duke Pauli in the School of Plant Sciences.

Working with data generated by UArizona's TERRA-REF project, which combines advanced robotic sensing with crop science to improve agricultural outcomes, Bartelme developed a machine learning framework to investigate how genes, environment, and phenotype are interrelated.

With regard to crops for example, he said: "I can use information to test how well my model predicts a mutation in a gene with the height of a plant. Or conversely, the model can take genetic data and then determine what the ideal growth conditions would be."

Key to his framework is its flexibility across research areas. For example, the same model can easily be adapted to medical applications, such as predicting predisposition to cardiopulmonary disease based on information such as stress metrics, family history, or health and behavior, or to predicting outcomes for cancer.

"It doesn't matter if it's a plant in the field or a metastasized cell," he said. "Linking genetic information to an actual phenotype is a similar process."

"One of the reasons we're delighted to have Ryan Bartelme join our Science Analysts team is the breadth of his scientific background," said CyVerse co-principal investigator Eric Lyons, an associate professor in UArizona's School of Plant Sciences and School of Information. "The processes we apply to data analysis can be used to solve problems in all scientific domains, and Ryan's skills and enthusiasm demonstrate that perfectly."

Bartelme readily admits his approach to a career in plant phenomics was less than linear. "It was something that I wanted to do without knowing that I wanted to do it," he said. Follow him on Twitter @MicrobialBart.