Food safety challenges from emerging contaminants such as antibiotics and antibiotic resistance genes (ARGs) have received increasing attention due to rapid increases in their abundance in agroecosystems. This is particularly true in soil-vegetable systems as microbiomes and antibiotic resistomes of vegetables are important to their quality and safety and could be influenced by crop production with contaminated soil and water. Additionally, the food safety of vegetables may also drive consumers' preference and demand for certain food products (especially for labeled products such as USDA Organic, Raised Without Antibiotics, etc.). Using a soil-lettuce (Lactuca sativa) model system, the first study in this dissertation assessed how irrigation with antibiotics-contaminated water via overhead or soil-surface irrigation could influence bacterial communities and ARG profiles in lettuce shoots, roots, and soil, using 16S rRNA amplicon sequencing and high throughput qPCR techniques, respectively. The overall abundance and diversity of ARGs and bacteria associated with soil-surface irrigated lettuce shoots were lower than those under overhead irrigation, indicating soil-surface irrigation may have lower risks of producing food crops with high abundance of ARGs. ARG profiles and bacterial communities were sensitive to pharmaceutical exposure, but no consistent patterns of changes were observed. The second study examined the fate and transport of selected antibiotics through bulk soil, rhizosphere soil, and lettuce roots and shoots under soil-surface irrigation. Root concentration factors based on the antibiotic concentrations in bulk soil (RCFbs) were significantly higher than those based on antibiotic concentrations in rhizosphere soil (RCFrs) for ciprofloxacin, lincomycin, oxytetracycline, sulfamethoxazole, and tetracycline, similar for trimethoprim and tylosin, and lower for monensin. The third study investigated bacterial community assembly and ARG profiles in lettuce shoots, roots, rhizosphere soil, and bulk soil upon exposure to antibiotics. Bacterial communities were driven by stochastic processes upon exposure to low level antibiotics, and were more resilient in roots and rhizosphere soil than in bulk soil and shoots. The fourth study explored the importance of demographics, food-relevant habits, and foodborne disease perception to consumers' buy and pay preferences to labeled products by using conventional statistical and novel machine learning methods to analyze survey data. Consumers' willingness to buy or to pay more for certain labeled food products is dependent on certain demographic traits (e.g., urban living) and food-relevant habits (e.g., cooking fresh produce). Machine learning methods achieved sufficient prediction accuracy scores for estimating consumers' willingness to buy or to pay for labeled products, and thus could be useful tools for evaluating survey data and facilitating the development of strategies promoting healthy food production and consumption.
Read
