Integrating fluorescence microscopy and Generalized Additive Models to predict Escherichia coli colonization dynamics on lettuce leaves

Leafy greens are susceptible to contamination by human pathogens, and because ready-to-eat salads cannot undergo thermal decontamination, even initial contamination poses a food safety risk. To understand early bacterial colonization, we combined high-resolution fluorescence microscopy with Generalized Additive Models (GAMs) to investigate development of Escherichia coli micro-colonies on the leaves of romaine lettuce ( Lactuca sativa L. var. longifolia ) and wild lettuce ( Lactuca serriola L.). Green Fluorescent Protein labelled bacteria were used as experimental tracers to visualize colonization dynamics under controlled conditions. Bacterial proliferation was quantified through fluorescence area, representing spatial expansion, and fluorescence intensity, reflecting aggregation; their combined metric provided a reliable proxy for micro-colony development. Over the 24 h incubation period, romaine lettuce consistently supported higher colonization than wild lettuce, reflecting differences in leaf surface micro-morphology. Maximum surface roughness emerged as a structural factor associated with bacterial proliferation. GAMs captured nonlinear growth patterns and species-specific responses, with romaine models explaining up to 91% of the observed variance. Comparison with Colony Forming Units (CFU) counts showed that fluorescence imaging resolved early spatial aggregation and colony expansion, whereas CFU measurements assessed culturability but lost spatial information due to tissue homogenization. Although fluorescence imaging based on genetically labelled bacteria is not directly applicable to post-harvest monitoring, it provides a platform for establishing predictive relationships between leaf surface traits and microbial growth. GAMs represent the transferable outcome of the study, enabling quantitative evaluation of temporal and morphological predictors of colonization dynamics and supporting approaches for improving microbiological safety in fresh-cut salad production systems.