A University of Arizona research team is exploring a new solution to an old agricultural challenge: how to supply crops with nitrogen efficiently and at scale.

Rather than refining how fertilizer is used, the team is working toward a fundamentally different approach to how crops access nitrogen—one that looks beyond a century-old method and draws on the relationships between plants and soil microbes. By better understanding how plants and bacteria interact, the researchers aim to reduce reliance on energy-intensive nitrogen fertilizers and open the door to more adaptive, efficient growing systems that will help farmers get the most out of every acre.

The effort brings together expertise from plant science, artificial intelligence, molecular biology and the humanities. Together, the researchers are studying how plants and microbes exchange chemical signals in the soil and how those interactions could be guided in ways that align with real-world agricultural practices. 

The team will pair new capabilities in AI and advanced biological research, an approach that applies design principles to optimize how living systems interact. The researchers plan to decode these two-way plant-bacteria communication signals to enhance the efficiency of the nitrogen cycle.

The project begins by focusing on Arizona cotton, using one of the state's most iconic crops to develop a biological toolkit for high-efficiency farming. By optimizing nutrient cycles in the desert Southwest, the team aims to create a scalable model that can be adapted for diverse crops and climates worldwide. 

“At the heart of this dialogue is nitrogen,” said Mark Beilstein, an associate professor in the School of Plant Sciences. “It’s a critical farming bottleneck because nitrogen is essential to plant growth yet scarce in natural soils.”

Critical and growing need

Beilstein is the principal investigator of the transdisciplinary project, heading one of six teams that received seed funding through the U of A Office of Research and Partnerships’ Big Idea Challenge in 2025. The challenge is designed to accelerate projects with the potential to transform lives, shape policy, drive economic impact and provide training for the next generation of talent. 

“Plants and microbes have been exchanging chemical signals for millions of years, and we’re only beginning to understand that language. Decoding it and turning that understanding into something farmers can use means crossing disciplines, bringing scientists, technologists and humanists together from discovery through application,” said Tomás Díaz de la Rubia, senior vice president for research and partnerships. “This team is combining AI, advanced biology and the humanities to understand those interactions and develop a new approach that could help farmers make every pound of fertilizer go further. That’s the kind of ambition with a clear path to real-world use that the University of Arizona is known for.”

Global food production will need to increase significantly by 2050 to meet projected demand for nearly 10 billion people, according to a World Resources Institute report. Meeting that demand will require higher yields and new approaches to how vital resources like nitrogen are produced and used. 

The nitrogen challenge

Since the industrialization of the Haber-Bosch process in the early 1900s, farmers have relied on synthetic nitrogen fertilizer to drive high yields. The breakthough process that produces nitrogen-based fertilizer transformed global food production but is energy intensive. The fertilizer also requires careful management, as nitrogen – a highly mobile element in soil – and nitrogen forms such as ammonium and nitrate salts can move beyond the root zone before the plant fully uses it. 

Maximizing the efficiency of nutrient inputs is a primary goal for growers. For Arizona cotton farmers, nitrogen fertilizer represents a major investment, often second only to water, according to Beilstein.

To help farmers grow higher volumes of food and other crops, the researchers aim to customize a next-generation probiotic soil additive, a treatment that cultivates beneficial bacteria in the soil, to optimize the efficient use of nitrogen fertilizer. 

“Altering the microbiome is an active field of research in agriculture. Our unique perspective is to try and engineer portions of that microbiome,” Beilstein said.  

Existing commercial soil treatments rely on naturally occurring bacteria or broad mixtures that aren’t tailored to a plant’s needs. The U of A team’s approach positions the university to work with industry partners to develop a novel product.

The tools to crack a mysterious language 

Plants release chemical signals from their roots to attract beneficial microbes, which help them access nutrients, fend off pathogens and adapt to environmental stresses. 

“It’s like the plants are telling the right bacteria, ‘Come live with me, because I need the food and protection you provide,’” Beilstein said.  

The bacteria also talk to plants, sending signals that help them secure nutrients and maintain a stable environment around the roots. Researchers know these exchanges happen through chemical messages, but the exact molecules behind the dialogue remain largely unknown.  

Using the latest AI large language models, the researchers aim to decode these signals. Applying biotechnological tools will allow the team to use that language to optimize soil microbes, ensuring a more effective response to plant signals. The bio-optimized bacteria will be applied as a probiotic soil additive, positioning them around roots to make fertilizer use more efficient. 

“We want to drive that interaction between the plant and the bacteria so that they're together, exactly when and where we want them to be,” said Rebecca Schomer, co-principal investigator and assistant professor in the School of Plant Sciences within the College of Agriculture, Life and Environmental Sciences.  

Society, science and grand ambitions 

The team is building stakeholder communication strategies early to guide development toward solutions that will be accepted and used.  

“We hope to address potential adoption from the onset, instead of putting time and effort into development only to find that there's pushback, or that people are not interested in having the cost-benefit conversation after the fact,” Beilstein said. 

College of Humanities professor Matthew Mars is leading the effort to develop and pilot a set of descriptive metaphors that enhance stakeholders’ understanding of the technology. 

Concurrently with communication efforts, Schomer is leading the molecular biology research to refine microbial candidates. As the project progresses, Claire McWhite, assistant professor of molecular and cellular biology in the College of Science, will guide AI decoding of plant-microbe communication networks to enhance these biological interactions. Following these steps, the team will co-cultivate modified microbes with plants in greenhouse conditions before eventually moving to field trials. 

Beilstein will oversee plant testing and ensure the research is connected to practical needs. Beilstein said that as the group looks far ahead – beyond Arizona cotton and additional crops – the collaborators dream about the applications that could stem from decoding the language interactions of single-celled bacteria and multicellular plants.  

 The fundamental rules could extend to medicine, he said. Just as plants rely on microbial partners to access nutrients, humans are multicellular organisms with their own microbiomes. One example could be treating people with antibiotics who have difficulty retaining beneficial bacteria.  

 “Could we engineer something that's more stable and more likely to stick around, because we understand these signaling systems?” Beilstein asked.