This past June, CABBI undergrad Jack Mason won a hackathon competition at Plant Biology 2024, a yearly conference held by ASPB, the American Society of Plant Biologists.

A hackathon is a short, competitive event where people work in teams to solve computational or data-based problems. They are a relatively new but increasingly important element of biological research conferences because addressing major societal issues such as clean energy, sustainable agriculture, and food security often necessitates the use of artificial intelligence and big data.

In this case, the ASPB-ARN (African Researchers Network) challenged teams to design a computational framework that uses modern tools like artificial intelligence and data analytics to fast-track a breeding pipeline for enhanced flavor in a fruit or vegetable of their choosing.

Tackling challenges such as these requires a combination of skills that traditional training programs do not teach. However, cross-campus collaboration at the University of Illinois has created the opportunity for Mason and his peers to graduate with a unique blend of skills.

Mason is pursuing a bachelor’s degree in Computer Science + Crop Science — a program jointly hosted by the College of Agricultural, Consumer, and Environmental Sciences and the Grainger College of Engineering.

Within CABBI and AIFARMS, Mason is applying convolutional neural networks — a form of machine learning — to do image analysis that allows the in-field performance of emerging bioenergy crops to be quickly evaluated by uncrewed aerial vehicles (i.e. drones). This work, which he performs as a member of Andrew Leakey’s research group, is helping to more rapidly breed dedicated biomass crops that can be used as the foundation of a sustainable bioeconomy.

Mason’s research experience was, in part, made possible by a Summer Undergraduate Research Fellowship from ASPB. One attractive but required element of the fellowship was attending the ASPB annual conference.

Mason was eager to attend. “[Going to Hawaii for Plant Biology 2024] was a really exciting experience,” he said. “It was great to connect with so many professionals from around the world and to see the broad domain of plant sciences.” And, given his background, signing up for the conference’s hackathon was a must.

The hackathon challenge put Mason and his teammates’ interdisciplinary skills to the test. To improve a fruit or vegetable’s flavor, they’d need to select for genetic traits associated with characteristics like high sugar content and balanced acidity.

“My teammates — who both had backgrounds in plant breeding — were able to determine the flavor compounds we should focus on,” Mason said. “And then with my experience with modeling, I worked on finding a way to optimize for those target compounds.”

The team proposed a model that, when given the genetic profiles of potential parent crop lines, would iteratively predict which combinations of parent crops would produce the best hybrids.

Rather than using a traditional deep network to approach this task, Mason opted to leverage a transformer network. A transformer is the mathematical engine behind Chat GPT and other powerful computational tools.

“You would take the known parent genetics and flavor profiles of preexisting hybrids and train a network to predict flavor profiles from those input genetics,” Mason said. “And from there the model would test every possible cross of lines. And once we’ve found the crosses that we’d like to try, we could give those crosses to breeders to see if they do indeed improve the flavor.”

Mason and his teammates gave a presentation on this framework at the end of the hackathon, and their win was announced the very next day.

As he enters his final year of undergrad, Mason is looking toward the future. “Currently I’m thinking about pursuing a Ph.D.,” he said. “And I’m also applying to jobs in agriculture and the tech industry. I really love working with these cutting-edge models.”

“I’d like to thank Andrew and CABBI for taking me on as a freshman,” Mason said. “I owe an incredible thanks to Jeremy Ruhter, who taught me how to think like a computer scientist; to Dylan Allen, who inspires me every day with his passion for plant biology; and Sebastian Varela, for his continued research mentorship. I stand on the shoulders of giants, and I wouldn’t have crossed that stage without their investment in me — before I had anything to give back. All I can say is thank you.”

Director Andrew Leakey is hopeful for CABBI’s role in training the next generation of researchers. “Together we are enabling many more students like Jack by building pipelines for transdisciplinary training toward a wide range of careers in the bioeconomy, including the Research Internship in Sustainable Bioenergy (RISE),” he said. “If you get a chance to meet the young people we have in these programs, you can only come away optimistic for the future.”

CABBI and the Leakey research group are co-hosted by the Carl R. Woese Institute for Genomic Biology; Institute for Sustainability, Energy and Environment; College of Liberal Arts and Sciences; College of Agricultural, Consumer, and Environmental Sciences; Grainger College of Engineering; and the Center for Digital Agriculture.