PhD Fulbright Scholar Manana Dlangalala Applies Cutting Edge AI Tools to Her Work on South African Farms.
On farms where crops and livestock are grown together, invisible microbial worlds can shape both food safety and public health. Animal waste, which carries bacteria and other microbes, can be washed by rain into soils and irrigation water used for crops. What happens to those microbes, how they interact with each other and evolve, how they are controlled by other microbes or multiply to become health risks for people eating the crops, has largely remained a mystery since humans first began farming.
But modern technology is allowing researchers to trace microbes in agricultural systems through their genes and use artificial intelligence to find patterns that reveal those hidden dynamics. Among the important questions researchers are concerned with is how and when microbes develop resistance to antibiotics—those found naturally in the soil as well as those introduced through human use. The results will lead to safer farming practices around the world.
That’s why PhD student Manana Dlangalala has brought the research she conducted on South African farms across the globe to the lab of Nutrition and Food Science Assistant Professor Ryan Blaustein. She is a Fulbright Scholar from the University of Pretoria, and like the researchers on Blaustein’s team, Dlangalala’s work centers on tracing microbes through farming systems, from soils and water to various types of produce.
But unlike Dlangalala, Blaustein and his team have been using AI to build a clearer picture of how microbes move and how resistance to different pathogens and antibiotics might emerge and spread through the food system. Dlangala has come to the US to learn their approach and apply it to her research back home.
For the last few months, she has been strengthening her data analysis skills, which she will later incorporate into her research toolkit for predictive modeling using machine learning. It is a natural extension of her work to understand how patterns of antimicrobial resistance in E.coli differ between the small-scale and commercial farming operations on “integrated farms,” agricultural systems where crops and animals are raised together. She is hoping to build predictive models that could predict antibiotic resistance based on microbial data.
“If we can predict where antimicrobial resistance is likely to occur, we can implement early and more targeted interventions,” she says.
In South Africa, integrated farm systems span two very different scales: small, backyard operations and large commercial farms. Before coming to UMD, her research focused on comparing the presence of bacteria through both types of farm ecosystems—from soil and irrigation water to pig feed, fresh produce, and even animal waste.
“We wanted to look at everything together,” she explains. “Not just one piece, but the whole system.”
To do that, she collected samples from multiple points across six farms—three small-scale and three commercial. She tested them for indicator microorganisms like E. coli, which signal potential contamination and thus food safety risk. The bacteria appeared across both types of farms, but important differences emerged.
On small farms, E. coli levels were significantly higher in pig feed and fresh produce. She suspects that storage conditions or handling practices may play a role, especially since small-scale farmers typically purchase feed rather than produce it themselves and have limited formal hygiene training. In contrast, commercial farms—where feed is often made from homegrown maize—showed lower contamination in feed but higher E. coli levels in irrigation water, likely tied to the use of surface water sources like rivers or dams which may be contaminated.
“These systems are complex,” she says. “You can’t always compare them directly because conditions are different.”
Her work identified important patterns in resistance to antibiotics. For example, she found that while resistant bacteria were present in both systems, multi-drug resistance (defined as resistance to three or more classes of antibiotics) was more common in samples from small-scale farms. That raises concerns about antibiotic use and regulation.
“In some cases, antibiotics are used for growth promotion, not just treatment,” she notes. “And there aren’t always clear guidelines.”
The ultimate goal of her work is not just to identify problems, but to help solve them. By mapping where contamination and resistance occur, she hopes to provide farmers with practical recommendations to reduce risk, such as improving feed storage or treating irrigation water.
Beyond the technical skills, the experience has reshaped how she thinks about her work. Exposure to new research styles, collaborative approaches, and diverse perspectives have helped her refine not only her analysis, but also how she tells the story of her data.
“I was struggling to package my work before,” she admits. “Now I can see the bigger picture—what story the data is telling.”
As she prepares to return home and complete her Ph.D., she carries with her not just findings, but a broader vision: one that connects environmental health, animal health, plant health, and human well-being in a single, integrated farm production system.
