本研究初步探讨了矿物元素指纹分析对我国地理标志猪肉产地溯源的可行性。采用电感耦合等离子体质谱对四川巴山的青峪黑猪肉，山东莱芜黑猪肉和北京黑六猪肉三种不同地域来源的地理标志猪肉中33种矿物元素含量进行测定。通过元素含量筛选，排除猪肉样品中含量低于或接近检出限的元素，筛选出13个元素进行研究。结合主成分分析、多重比较分析和判别分析对数据进行统计分析。其中Na、Fe、Co、Cu、Zn、Se、Rb、Sr共8种元素在地域之间差异显著（p<0.05）。通过对比偏最小二乘法判别分析（PLS-DA）、基于正交信号校正的偏最小二乘判别分析（OPLS-DA）、支持向量机、朴素贝叶斯、决策树和神经网络六种分类模型，得出OPLS-DA和决策树分类模型较适合基于矿物元素指纹分析的特色猪肉产地溯源。矿物元素指纹分析在我国地理标志猪肉产地溯源领域具有长远的应用前景。

The feasibility of tracing the origin of geographical indication-featured pork in China by analysis of mineral element fingerprints was invastigated. The contents of 33 mineral elements in Qingyu Black Pork from Bashan in Sichuan, Laiwu Black Pork from Shandong and Heiliu pork from Beijing were determined by inductively coupled plasma mass spectrometry (ICP-MS). Through element content screening and after the exclusion of the elements with contents below or close to the detection limit, 13 elements in the pork samples were screened out for further investigations. Statistical analysis of data was carried out through combining principal component analysis, multiple comparative analysis and discriminant analysis. Among them, the contents of the eight elements, Na, Fe, Co, Cu, Zn, Se, Rb and Sr, showed significant regional differences (p<0.05). Based on the comparison among Partial Least Squares Discriminant Analysis (PLS-DA), Orthogonal Signal Correction Based Partial Least Squares Discriminant Analysis (OPLS-DA), Support Vector Machine, Naive Bayesian, Decision Tree and Neural Network, OPLS-DA and decision tree classification models are more suitable for origin traceability of characteristic pork based on mineral element fingerprints. Mineral element fingerprint-based analysis has a long-term application prospect in origin traceability of geographical indication-featured pork in China.

国家重点研发计划重点专项（2017YFC1601704）