食品污染、药物滥用、非法添加是我国食品安全领域存在的突出问题，基于质谱裂解特征的食品化学危害物筛查技术可为解决这一问题提供可能。在前期工作中，我们对磺胺、喹诺酮等四类外源性药物的质谱裂解规律进行了研究，并对基于质谱特征碎片的危害物筛查及锁定技术进行了初步探索。在此基础上，本项目拟系统研究食品化学危害物及其生物转化产物的质谱裂解规律，采用计算机深度学习技术提取每类化学危害物的质谱特征裂解信息，开发可用于大型数据匹配检索的新型智能检索程序，构建基于质谱裂解特征的UHPLC-HRMS筛查技术；同时，利用人工神经网络技术深度挖掘化合物质谱碎片与其特征结构及保留时间之间的关联，实现高效、精准的质谱碎裂规律解析，提升未知化学危害物的筛查速度和准确度。预计通过本项目的实施，可突破食品未知危害物筛查的技术难点，为食品安全性评价和风险预警提供必要的技术手段和支持。

Food contamination, drug abuse, illegal addition are prominent problems in the field of food safety in China. The screening techniques for chemical hazards in foods based on MS fragmentation characteristics can provide a possibility to solve this problem. In the preliminary work, MS fragmentation characteristics of 4 kinds of drugs were summarized, screening and locking techniques for chemical hazards in foods based on MS fragmentation characteristics were tentatively investigated. On this basis, This project will focus on the systematic study of MS fragmentation characteristics of chemical hazards and their simulated metabolites, computer deep learning technology will be applied to extract MS fragmentation characteristics information of each type of chemical hazards and their simulated metabolites, C++ language and exhaustive algorithm will be used to develop new intelligent application of large data matching retrieval. At the same time, the artificial neural network will be applied to excavate deeply compounds pieces of spectrum and its corresponding relation between the characteristics of structure, which will increase screening speed and accuracy of unknown chemical hazards. This study will provide the necessary technical support for safety regulation and risk warning of foods.