New Mexico Consortium
Ingredients commonly used within a particular group of products could potentially be targeted by terrorists to intentionally contaminate many foods simultaneously. Additionally, specialty ingredients, or those having only a few suppliers, could be preferentially chosen because of their potential to find their way into foods distributed across a vast geographic area. Such attacks could lead to major public health incidents having tremendous shock effect and potentially severe consequences. This project demonstrates a methodology to establish a broader understanding of ingredient-product relationships and exploit that understanding to guide informed decisions about where to focus traceability efforts. This project has two aims that are being pursued simultaneously. The first aim is to develop an informatics approach to better characterize food product-ingredient relationships. Additionally, query and analysis methods that aid reasoning are being developed that will ultimately allow users to investigate connections between individual ingredients, foods, and food groups. The summary will provide quantification of the food and ingredient interrelationships and create representations of the solutions to promote understanding. The second aim is to develop reasoning models that use enhanced traceability information to solve problems. To accomplish this, Bayesian inference and statistics can be used to develop reasoning models that can help in contexts like applying epidemiology to contain outbreaks of food-borne illness. For example, by connecting the contaminating agents to individual ingredients, and combining this information with food diaries from affected and unaffected consumers, one could quantify the likelihood of ingredients and food groups being the source of the outbreak.