Soybeans are one of the major allergenic foods in many countries. Soybeans are commonly processed into different types of soy ingredients to achieve the desired properties. The processing, however, may affect the protein profiles and protein structure, thus affecting the detection of soy proteins. Mass spectrometry (MS) is a potential alternative to the traditional immunoassays for the detection of soy-derived ingredients in foods. This study aims to develop a liquid chromatography−tandem MS method that uniformly detects different types of soy-derived ingredients. Target peptides applicable to the detection of six commercial soy ingredients were identified based on the results of MS label-free quantification and a set of selection criteria. The results indicated that soy ingredient processing can result in different protein profiles. A total of six soy ingredients were then individually incurred into cookie matrices at different levels. Sample preparation methods were optimized, and a distinct improvement in peptide performance was observed after optimization. Cookies and dough incurred with different soy ingredients at 100 ppm total soy protein showed a similar level of peptide recovery (90% mean signal relative to unroasted soy flour), demonstrating the ability of the MS method to detect processed soy ingredients in a uniform manner.

Cashews are one of the most prevalent causes of tree nut allergies. However, the cashew proteome is far from complete, which limits the quality of peptide identification in mass spectrometric analyses. In this study, bioinformatics tools were utilized to construct a customized cashew protein database and improve sequence quality for proteins of interest, based on a publicly available cashew genome database. As a result, two additional isoforms for cashew 2S albumins and five other isoforms for cashew 11S proteins were identified, along with several other potential allergens. Using the optimized protein database, the protein profiles of cashew nuts subjected to different oil-roasting conditions (139 and 166 °C for 2−10 min) were analyzed using discovery LC-MS/MS analysis. The results showed that the cashew 2S protein is most heat-stable, followed by 11S and 7S proteins, though protein isoforms might be affected differently. Preliminary target peptide selection indicated that out of the 29 potential targets, 18 peptides were derived from the newly developed database. In the evaluation of thermal processing effects on cashew proteins, several Maillard reaction adducts were also identified. The cashew protein database developed in this study allows for comprehensive analyses of cashew proteins and development of highquality allergen detection methods.