Included in the May 22, 2025, biweekly update

This week’s articles by MSU faculty, specialists and students making a difference feature foodborne pathogen behavior during storage and precision agriculture tools that detect crop disease risks in real-time.

Dynamics of Physiological Changes of Shiga Toxin-Producing Escherichia coli O157:H7 on Romaine Lettuce During Pre-Processing Cold Storage, and Subsequent Effects on Virulence and Stress Tolerance

Sharma et al. (2025) examined how cold storage affects the survival, stress tolerance, and virulence of Shiga toxin-producing Escherichia coli O157:H7 (STEC) on Romaine lettuce. Given the frequency of lettuce-linked outbreaks and the long transport times between harvest and processing, the researchers studied whether cold temperatures trigger physiological changes in the pathogen that could influence food safety.

Using greenhouse-grown lettuce injected with three outbreak-associated STEC strains, the study simulated commercial conditions. Lettuce was stored at 2.2°C for five days following brief exposure to typical harvest temperatures (9°C or 17°C). The researchers tracked transitions in bacterial states, and measured changes in acid tolerance, chlorine resistance, and virulence using a Galleria mellonella larvae model.

The researchers found that harvesting lettuce at colder temperatures increased the likelihood that E. coli O157:H7 would enter a dormant state, reducing its ability to cause illness. As cold storage time increased, the bacteria became less tolerant to acidic conditions as well, while chlorine tolerance remained largely unchanged.

The study highlights how cold storage not only preserves lettuce but also alters the behavior of pathogenic bacteria on its surface. These physiological shifts could inform microbial risk assessments and support the development of safer handling practices for leafy greens. The researchers emphasized that strain-specific responses and storage conditions should be considered when applying these findings more broadly.

Sharma, D., Owade, J. O., Kamphuis, C. J., Evans, A., Rump, E. S., Catur, C., Mitchell, J., & Bergholz, T. M. (2025). “Dynamics of physiological changes of Shiga toxin-producing Escherichia coli O157:H7 on romaine lettuce during pre-processing cold storage, and subsequent effects on virulence and stress tolerance.” Applied Microbiology, 5(2), 45. https://doi.org/10.3390/applmicrobiol5020045

Corresponding author: Teresa M. Bergholz, tmb@msu.edu.

Proteus: Enhanced mmWave Leaf Wetness Detection with Cross-Modality Knowledge Transfer

Liu et al. (2025) developed Proteus, an advanced sensing system that uses mmWave radar and machine learning to detect leaf wetness with greater accuracy. Leaf wetness is key to predicting and managing crop disease, but existing detection tools often struggle with accuracy—especially under difficult field conditions like low light, wind, or complex plant structures. Prior approachesrelied on Synthetic Aperture Radar (SAR) or used SAR with RGB cameras in limited ways, leading to noisy images, poor surface detail, and inconsistent results across different environments.

Proteus improves on these limitations in three ways. First, the researchers designed a noise reduction method to eliminate speckle interference in SAR images, resulting in clearer radar data. Second, they incorporated phase angle information to capture fine surface details that indicate the presence of moisture. Third, they introduced a “Teacher-Student” machine learning model, where a network trained on RGB images guides a radar-based model to better detect subtle signs of wetness.

The system was tested using commercial radar hardware across a wide range of conditions, including indoor trials and real-world farms. Proteus accurately identified wet leaves 96.3% of the time and was able to estimate how long leaves remained wet with an error margin of less than five minutes. It consistently outperformed existing systems, maintaining accuracy even in challenging environments.

This study shows that combining advanced radar imaging with machine learning across different sensors can improve how we monitor crop conditions. Proteus provides a reliable tool for detecting disease risk and tracking plant health in real-time, making it suitable for use in precision agriculture.

Liu, Y., Gan, M., Zeng, H., Ren, Y., Li, G., Lin, J., Dong, Y., Tan, X., & Cao, Z. (2025).“Proteus: High-Accuracy mmWave Leaf Wetness Detection via Cross-Modality Learning and Signal Enhancement.” Proceedings of the ACM on Interactive, Mobile, Wearable and Ubiquitous Technologies, 9(1), Article 26. https://doi.org/10.1145/3715014.3722052

Corresponding author: Yimeng Liu, liuyime2@msu.edu.