Center for Food Safety and Applied Nutrition, Division of Food Processing Science and Technology, U.S. Food and Drug Administration , 6502 South Archer Road, Bedford Park, Illinois 60501, United States.
J Agric Food Chem. 2014 Jun 4;62(22):5198-206. doi: 10.1021/jf404924x. Epub 2014 May 20.
Chromatographic profiles of skim milk powder (SMP) and mixtures of SMP with soy (SPI), pea (PPI), brown rice (BRP), and hydrolyzed wheat protein (HWPI) isolates were obtained by ultrahigh-performance liquid chromatography (UHPLC) with 215 nm detection. Two data analysis approaches were compared for their utility to classify samples as authentic or adulterated. The t test approach evaluated data points exceeding the 99% confidence limit of the mean authentic SMP chromatogram and used data points from the entire chromatogram. The other approach used the multivariate Q statistic from a SIMCA model of authentic samples to determine adulteration and used a selected retention window to obtain best classifications. Q-Statistic and t test correctly classified adulteration of SMP with SPI at the 1% and 3% levels, respectively, while minimizing false classifications of authentic SMP. Detection of SMP adulterated with PPI, BRP, and HWPI was possible at higher adulteration levels.
采用超高效液相色谱法(UHPLC)结合 215nm 检测,获得了脱脂乳粉(SMP)以及 SMP 与大豆(SPI)、豌豆(PPI)、糙米(BRP)和水解小麦蛋白(HWPI)混合物的色谱图。比较了两种数据分析方法在对样品进行真伪分类方面的应用。t 检验方法评估了超过真实 SMP 色谱图平均值 99%置信限的数据点,并使用了整个色谱图的数据点。另一种方法使用来自真实样品 SIMCA 模型的多元 Q 统计量来确定掺假情况,并使用选定的保留时间窗口来获得最佳分类。Q-统计量和 t 检验分别正确地将 SMP 与 SPI 的 1%和 3%水平的掺假情况进行了分类,同时最小化了对真实 SMP 的误分类。SMP 与 PPI、BRP 和 HWPI 的掺假检测可以在更高的掺假水平下进行。