Food allergies are an important health problem and affect up to 2% of the adult population and 8% of children worldwide. Under the Food Allergen Labeling and Consumer Protection Act (FALCPA) of 2004, foods that contain or derive from the "Big 8" allergens (milk, egg, finfish, crustacean shellfish, tree nuts, peanuts, wheat, and soybeans) must be declared and the "common or usual name" of the allergen source must be printed on the label of the food product. Currently, the most common used detection methods for food allergens are enzyme-linked immunosorbent assay (ELISA) based. ELISA, a protein-based method, targets specific allergen(s) and detects by colorimetric reaction following binding with a specific-enzyme labeled antibody. However, studies have demonstrated that matrix interference and heat treatment can interfere with the detectability of commercial ELISA kits. An alternative approach to targeting the allergen in soy is to use deoxyribonucleic acid (DNA) as a unique marker that can be used to indicate the presence of soy in food. According to FALCPA the source of an allergen should be declared on the label, therefore identifying an allergen, such as soy, by DNA detection could be a valid means of meeting FALCPA requirements. Real-time polymerase chain reaction (real-time PCR), a DNA-based method, can identify the presence of soy through amplification of specific sequences of DNA through the use of primers. However, the sensitivity of real-time PCR can be influenced by the amplification protocol, primer design and DNA extraction methods. Thus, the main objectives of this study were to 1) verify the specificity of primers designed to detect soy DNA from different soy products, 2) optimize the previously developed real-time PCR protocol to detect soy DNA, 3) investigate the application of two commercially available DNA recovery systems (column and magnetic beads) to recover soy DNA from different forms of soy products using real-time PCR and 4) determine the effect of food matrices and thermal processing on soy detection using DNA and ELISA methods. In this study, Wizard Magnetic DNA Purification system kit (Promega, Madison, WI) was selected as the column DNA recovery system while DNeasy mericon Food Kit (Qiagen, Valencia, CA) was selected as the magnetic beads system. Neogen Veratox for soy allergen was selected as the ELISA system. The evaluations of both DNA recovery systems were conducted on soy protein isolates (SPI), powdered soybean and soymilk. The effect of thermal processing in soy detection was conducted on four different food matrices (protein, fat, carbohydrate and water). Each food matrix was spiked with 10% soy protein isolates and heated at 95ºC for an hour. Both DNA (column and magnetic beads DNA recovery system) and ELISA detection methods were used to detect soy in heated and non-heated food matrices. The limit of detection for column DNA recovery method in soybean, SPI and soymilk can be as low as 20 ppm, while magnetic beads DNA method was matrix dependent. The magnetic beads methods demonstrated a lower detection for soybean sample (1.33 ppm) but higher for soymilk (133.3 ppm). The soy percent recovery for non-heated food matrices was higher in ELISA methods and lower in magnetic beads DNA method. For heated food matrices, percent recovery for both DNA methods was higher than ELISA method. Overall, heat treatment can significantly reduce the ability of the ELISA method to detect soy in all food matrices. However, for DNA methods (column and magnetic beads), water and ranch matrices were the only two that were significantly affected by thermal processing. In terms of food matrices, water matrix (heated and non-heated) has the highest percent recovery of soy for all detection methods. However, percent recovery of soy in flour matrix (non-heated) was the lowest using both DNA methods. / Graduation date: 2013