奶茶粉脂肪酸指纹特征及真实性判别模型建立

Fatty acid fingerprint characteristics and establishment of authenticity discrimination model for milk tea powder

中文关键词:  奶茶粉  脂肪酸  真实性  主成分分析  偏最小二乘

英文关键词:milk tea powder  fatty acid  authenticity  orthogonal partial least squares discriminant analysis  partial least square

中文摘要:

      目的 评价以脂肪酸(fatty acid, FAs)指纹建立奶茶粉真实性判别模型的可行性。方法 采集生牛乳喷雾干燥(湿法)和植脂末、乳粉、炼乳粉粉末原料混合制备(干法)奶茶粉138种产品, 并采集植脂末、奶油粉和奶粉为参照样品, 气相色谱法检测FAs, 进行差异性比较, 并进行正交偏最小二乘判别分析(orthogonal partial least squares discriminant analysis, OPLS-DA)和层聚类分析(hierarchical cluster analysis, HCA), 观察各种类奶茶粉样品聚类特征, 评价模式相似度。构建奶茶粉掺植脂末OPLS-DA和偏最小二乘(partial least square, PLS)预测模型。结果 生牛乳湿法制备的奶茶粉与其他3种干法制备的奶茶粉FAs指纹显著不同, 与奶粉有同质性, 两者模式相似度为94.9%; 3种干法制备的奶茶粉与植脂末、奶油粉存在同质性, 可归入同一个大类, 模式相似度为27.2%。OPLS-DA和PLS模型有良好的预测能力, 可以区分0%植脂末(奶粉)与添加12.5%植脂末的人工奶茶粉样品。3种市售干法制备奶茶粉植脂末掺入量在50.0%~100.0%区间, 生牛乳制备奶茶粉也有填充粉末油脂现象。结论 利用FAs指纹可以判别纯正的生牛乳湿法制备与干法工艺生产奶茶粉, 并可建立奶茶粉真实性的定性和定量判别模型, 为奶茶粉真实性判别提供了创新策略和方法。

英文摘要:

      Objective To evaluate the feasibility of establishing milk tea powder authenticity discrimination model based on fatty acid (FAs) fingerprints. Methods The 138 kinds of products in milk tea powder were collected by spray drying of raw milk (wet method) and mixing powder raw materials of non-dairy creamer, milk powder and condensed milk powder (dry method), and non-dairy creamer, cream powder and milk powder were collected as reference samples. FAs was detected by gas chromatography, and the differences were compared. Orthogonal partial least squares discriminant analysis (OPLS-DA) and hierarchical cluster analysis (HCA) were used to observe the clustering characteristics of various milk tea powder-like samples and evaluate the pattern similarity. OPLS-DA and partial least square (PLS) prediction model of milk tea powder mixed with non-dairy creamer was constructed. Results Milk tea powder prepared by wet method of raw milk was significantly different from milk tea powder FAs fingerprints prepared by other 3 kinds of dry methods, and it was homogeneous with milk powder, and the similarity between them was 94.9%. Milk tea powder prepared by 3 kinds of dry methods was homogeneous with non-dairy creamer and cream powder, which could be classified into the same category, and the pattern similarity was 27.2%. OPLS-DA and PLS models had good prediction ability, and could distinguish 0% non-dairy creamer (milk powder) from artificial milk tea powder samples with 12.5% non-dairy creamer. The mixing amount of 3 kinds of commercially available dry methods for preparing milk tea powder non-dairy creamer ranged from 50.0% to 100.0%, and raw milk for preparing milk tea powder was also filled with powdered oil. Conclusion FAs fingerprint can be used to distinguish the milk tea powder produced by wet method and dry method of pure raw milk, and the qualitative and quantitative models for judging the authenticity of milk tea powder can be established, which provides innovative strategies and methods for judging the authenticity of milk tea powder.