基于代谢组学技术的神农百花蜜真实性鉴别

Authentic identification of Shennong multifloral honey based on metabolomics technology

中文关键词:  神农百花蜜  代谢组学  鉴别  多元统计分析

英文关键词:Shennong multifloral honey  metabolomics  discrimination  multivariate statistical analysis

中文摘要:

      目的 通过非靶向代谢组学鉴定真实神农百花蜜特征标志物, 并建立基于偏最小二乘(partial least squares, PLS)的神农百花蜜判别模型。方法 采用基于超高效液相色谱-四极杆串联飞行时间质谱(ultra performance liquid chromatography-quadrupole time-of-flight mass spectrometry, UPLC-Q-TOF-MS)的非靶向代谢组学方法分析神农百花蜜样本, 利用V+S图、聚类热图和受试者工作特征(receiver operating characteristic, ROC)曲线确定真实神农百花蜜特征标志物, 利用神农百花蜜特征标志物构建PLS判别模型。结果 多元统计分析结果显示, 主成分分析(principal component analysis, PCA)和正交偏最小二乘-判别分析(orthogonal partial least squares-discriminant analysis, OPLS-DA)可以实现真实神农百花蜜和对照组蜂蜜的区分; 分别在正、负离子模式下筛选出了19种和27种差异代谢物, 其中13种差异代谢物在真实神农百花蜜中的含量较高, 可作为其潜在标志物; 通过ROC曲线评估, 确定了女贞甙、鸟苷、半胱氨酰-苯丙氨酸和4’-O-甲基甘醇W等13种化合物可以作为真实神农百花蜜的特征标志物; 利用神农百花蜜特征标志物建立PLS判别模型, 对未知样本的鉴别准确率可以达到100%。结论 非靶向代谢组学技术在神农百花蜜的鉴别中具有可行性。

英文摘要:

      Objective To identify the characteristic markers of real Shennong multifloral honey by non-targeted metabolomics, and establish the discrimination model of Shennong multifloral honey based on partial least squares (PLS). Methods A non-targeted metabolomic method based on ultra performance liquid chromatography-quadrupole time-of-flight mass spectrometry (UPLC-Q-TOF-MS) was used to analyze Shennong multifloral honey samples. V+S plot, cluster heat map and receiver operating characteristic (ROC) curve were used to determine characteristic markers of real Shennong multifloral honey, and PLS discriminant model was constructed using characteristic markers of Shennong multifloral honey. Results The results of multivariate statistical analysis showed that principal component analysis (PCA) and orthogonal partial least squares-discriminant analysis (OPLS-DA) could distinguish the real Shennong multifloral honey from the control group honey. Nineteen and 27 kinds of differential metabolites were screened out under the positive and negative ion modes, respectively. Among them, the content of 13 kinds of differential metabolites were higher in the real Shennong multifloral honey, which could be used as its potential markers. Through the evaluation of ROC curve, 13 kinds of compounds including ligustroside, guanosine, cysteinyl-phenylalanine and 4’-O-methylkanzonol W were identified as the characteristic markers of real Shennong multifloral honey. PLS discriminant model was established by using characteristic markers of Shennong multifloral honey, and the identification accuracy of unknown samples could reach 100%. Conclusion Non-targeted metabolomics technology is feasible in the Shennong multifloral honey authentication.