A recent article, published in April 2023 in Nature’s Scientific Data, provides public access to the past 20 years of fumigation data collected at the SoyFACE (Free Air Concentration Enrichment) experimental facility. ‘Fumigation data’ refers to the carbon dioxide and ozone that are released and the speed and direction of wind that spreads these gases across experimental plots. This data is captured by a computer every four seconds across the duration of the summer growing season. As one can imagine, scaling this up to all experimental plots across 20 different growing seasons generates a huge amount of data. Once all this information is collected, it needs to be quality controlled and analysed into usable metrics, such as seasonal averages. Luckily, we have a computational biologist on the team to help us tackle this.
Dr. Elise Aspray, the first author of the article, began as a postdoctoral associate in the Ainsworth lab at the beginning of 2022 after getting her doctoral degree in Applied Mathematics from the Florida Institute of Technology. As a computational biologist, Elise uses mathematics, data analysis, and simulations to better understand biological processes. Elise’s Ph.D. work used discrete forms of partial differential equations to create computer models that help understand how tiny blood cell fragments, called platelets, stick together and move around, providing valuable information about how blood clots form. When asked about her decision to pursue an applied mathematical career, she reflected on her university coursework, “It was always pretty abstract. And it was interesting, but it wasn’t always clear how it could be used in a useful way besides just in academia.”
Although a far cry away from blood clot mechanics, Elise’s desire to apply her mathematical skills to big-picture scientific projects led her to SoyFACE. Speaking on this transition, Elise explained-
“Even though the fields are so different, I really love being able to see how math and analysis can be used to actually help further those fields. To me, I’d rather see the math used in a real concrete way, mixed in with other fields. It’s not just a standalone thing.”
Now, Elise’s work at SoyFACE is publicly available to help inform climate change research, such as improving the accuracy of models that predict the impacts of climate change on agriculture. The work was published as a ‘Data Descriptor’, which pairs open access of large datasets (e.g., in the Illinois Data Bank) with a more traditional narrative. Explaining her reasoning for writing a Data Descriptor, Elise said, “It was really nice to be able to explain each of the steps in detail and not feel restricted the way I would in a little README file in a dataset. So it’s not just- here’s the data, use it how you want. It’s here’s all the details behind it and also potential for what it could be used for.”
The published datasets include hourly and seasonal averages of all the fumigation data. Even for a computational biologist, figuring out how to approach such a massive set of data was difficult. There was also the added challenge of removing faulty data points, such as when the fumigation equipment wasn’t working properly. Elise explained the challenges she faced in attempting to scale-up her analysis process- “I was doing it very manually for the first few weeks. I was basically just having all the fumigation data for each minute during the growing season and then just using an Excel formula to calculate the hourly averages.” Elise quickly realized this manual approach was not going to cut it for covering two decades worth of data in a reasonable timeframe and without accidentally making errors. She made the decision to switch over to the coding language MATLAB, after which she says, “It helped a lot when I started using my coding knowledge. I still have some manual work to do, but it’s a lot less. I can do it probably within a day or less for [any] year.”
Another challenge Elise faced was the background learning needed to understand the context in which her code was being used. Even with a biological background, transitioning into a different and already well-established research field was daunting. Speaking on her transition out of her Ph.D. in blood clot mechanics to the work at SoyFACE, Elise said, “The biggest similarities were this kind of period of learning the science behind the actual concept that I would be writing code to help. And then the actual coding itself and some sort of visualization. But that’s kind of where the similarities ended.”
Although facing her own learning curves, Elise was able to share her knowledge as the resident expert in computational biology with her colleagues, including two undergraduate students she mentored. Jelena Herriott, a co-author on the paper, completed an internship in the Ainsworth lab in the summer of 2022, identifying trends in wind speed and direction at SoyFACE. Jelena went on to present her work at the 2022 Society for the Advancement of Chicanos/Hispanics and Native Americans in Science (SACNAS) conference in Puerto Rico. Elise also served as a mentor to Munirat Abraham, who gained coding experience with Elise as part of the Center of Digital Agriculture’s Research Experiences for Undergraduates program. In an interview with ACES News, Munirat spoke on her experience in the lab, “My mentor, she was very, very helpful…For my project, we created 3D graphs using MATLAB. That was something that I was completely new to, and my mentor really like broke it down and helped me truly understand it.”
In addition to keeping up with current fumigation data, Elise is now analyzing other metrics collected throughout the history of SoyFACE such as yield and soil moisture data. She is currently learning more about machine learning techniques to help understand what future agricultural responses to climate change could look like.
Article Source: Aspray EK, Mies TA, McGrath JA, Montes CM, Dalsing B, Puthuval KK, Whetten A, Herriott J, Li S, Bernacchi CJ, DeLucia EH, Leakey ADB, Long SP, McGrath JM, Miglietta F, Ort DR, Ainsworth EA (2023) Two decades of fumigation data from the Soybean Free Air Concentration Enrichment facility. Scientific Data, doi: 10.1038/s41597-023-02118-x.
Story Source: Elise Aspray
Written by Megan Allen