基于代谢指纹图谱对扁形茶的产地溯源及品质预测

发布时间：2018-06-08 19:23

本文选题：代谢组学 + 产地溯源　；参考：《中国农业科学院》2016年硕士论文

【摘要】：茶叶有着悠久的历史,是中国重要的经济作物。在传统的名优绿茶中,龙井茶,特别是西湖龙井的地位举足轻重。目前,除西湖龙井外,新昌大佛龙井、嵊州越乡龙井、磐安生态龙井、淳安千岛湖龙井、萧山湘湖龙井等龙井茶“区域子品牌”已呈现快速发展的势头,形成了茶叶界独一无二的龙井茶“品牌板块”,也引起了业界高度关注。同时,以西湖龙井为例的部分名优茶市场依然广泛存在“以假乱真”的现象,为了给消费者一个明确的判断标准以及完整的科学依据,本研究以对扁形茶产地判定为目的进行化学指纹图谱的分类和判别研究。绿茶是世界主要的健康饮品之一,其内在品质是评判绿茶质量的主要因子,专家感官审评是当前主要的品质等级评定方法,然而感官审评或多或少受审评者的身体精神状态或人为主观等因素影响,不同的审评者对同一茶叶的品质评判可能会有所差异。在研究了龙井茶原产地判别问题的基础上,建立扁形茶品质的预测模型和扁形茶等级的判别模型,对比影响扁形茶品质的特征化合物和影响扁形茶产地的特征化合物,最终确定了不同产地的扁形茶样品并未与其品质有明显的相关性。究其根本原因是影响扁形茶品质的代谢化合物主要为儿茶素、氨基酸、植物碱、茶黄素以及其他儿茶素类聚合物等茶叶中含量较高的化合物,而不同产地间因海拔、纬度与小气候因素所形成的特征化合物主要为黄酮类、多糖类与肽类等茶叶中含量较低的化合物。本实验主要研究结果如下:1对浙江、四川、贵州、山东四个产茶省的扁形茶样品的分析通过对浙江、四川、贵州、山东四个产茶省的扁形茶样品的代谢指纹图谱进行多元统计分析,PCA分析结果R~2X=0.643,Q~2X=0.363,散点图中各产茶省样品各自呈现聚合的趋势;PLS分析结果R~2X=0.396,R~2Y=0.874,Q~2X=0.898。共鉴定了32种不同产茶省间的特征化合物。包括Cysteinyl-Glycine、3-Nitrotyrosine、Mevalonic acid-5P、N-Carbamoyl-2-amino-2-(4-hydroxyphenyl)acetic acid、3-O-p-Coumaroylquinic acid、Gallic acid4-O-(6-galloylglucoside)等。2对西湖、钱塘、越州、缙云四个产茶区的扁形茶样品的分析通过对西湖、钱塘、越州、缙云四个产茶区的扁形茶样品的代谢指纹图谱进行多元统计分析,PCA分析结果R~2X=0.730,Q~2X=0.388,散点图中各产茶省样品各自呈现聚合的趋势;PLS分析结果R~2X=0.437,R~2Y=0.784,Q~2X=0.845。共鉴定了32种不同产茶省间的特征化合物。包括5-Amino-6-(5'-phosphoribitylamino)uracil、4-Methyl-2-phenyl-2-pentenal、Ethyl aconitate、1-(2-Hydroxyphenylamino)-1-deoxy-beta-D-gentiobioside 1,2-carbamate、Imidazoleacetic acid ribotide等。3对西湖龙井茶保护区内个主要产茶村的扁形茶样品的分析通过对西湖龙井茶保护区内个主要产茶村的扁形茶样品的代谢指纹图谱进行多元统计分析,PCA分析结果R~2X=0.786,Q~2X=0.464,散点图中各产茶省样品各自呈现聚合的趋势;PLS分析结果R~2X=0.453,R~2Y=0.848,Q~2X=0.895。共鉴定了44种不同产茶省间的特征化合物。包括4-O-Methylgallic acid、3,4,5-Trimethoxycinnamic acid、Ethyl 2-phenyl-3-furancarboxylate、3-(3,4-Dihydroxyphenyl)lactic acid、Dimethyl tetrasulfide等。4影响扁形茶品质的特征化合物的鉴定通过三个等级来鉴定了水提取结合UPLC-Q-TOF/MS测到的1568种代谢化合物中的60种化合物。通过t检验得到其中23种化合物在A、B、C三个区域间具有显著性差异且三个区间的样品审评得分递减,其中L-Arginine与L-Glutamine等氨基酸类化合物在区域A与区域C之间具有显著性差异,得出该类化合物与扁形茶品质成正相关。与氨基酸类化合物相反,Serinyl-Serine等三种多肽类化合物在A区域最低C区域最高,表现出了与品质的负相关性。在儿茶素类化合物中Epigallocatechin表现出了与绿茶品质的负相关性,Catechin仅在B区域的含量显著低于C区域,其余Gallocatechin,Gallocatechin gallate,Catechin gallate,Epigallocatechin gallate等四种化合物都在不同程度上表现出了与品质成正相关性。包括Theaflavin与Catechin-(4beta-8)-Epigallocatechin在内的茶黄素类物质和黄烷醇类化合物的聚合物均表现出了与绿茶品质的负相关。因此,可初步认为,这些特征化合物可用来进行扁形绿茶品质的预测。5甲醇提取条件与水提取条件下扁形茶品质与等级预测模型的对比分析在高通量色谱质谱技术中,水提取条件下得分预测模型(PLS回归模型)的RMSEP值小于甲醇提取条件,而R~2、Q~2值大于甲醇提取条件。在等级预测模型中,决策树、神经网络和贝叶斯网络等模型的准确度都在在水提取条件下达到90%以上,而预测度均达到80%以上,远高于甲醇提取条件下准确度72%的均值和预测度62%的均值。这表明,相对传统的甲醇提取方法,在UPLC-Q-TOF/MS平台中水提取方法更适合进行绿茶审评得分或等级的预测。
[Abstract]:Tea has a long history and is an important economic crop in China. In the traditional famous green tea, Longjing tea, especially West Lake Longjing, is very important. At present, besides West Lake Longjing, Xinchang Grand Buddha Longjing, Shengzhou Yue Township Longjing, Panan ecological Longjing, Chunan Qiandao Lake Longjing, etc. The momentum of rapid development has formed the unique "brand plate" of Longjing tea industry in the tea industry, which has also aroused great concern in the industry. At the same time, some famous tea markets in Longjing, West Lake, still exist widely in the "false and true" phenomenon, in order to give consumers a clear standard of judgment and a complete scientific basis, the research is a complete scientific basis. The purpose of the study is to classify and discriminate the chemical fingerprint with the purpose of determining the producing area of the flat tea. Green tea is one of the main health drinks in the world. Its intrinsic quality is the main factor to judge the quality of green tea. The expert sensory evaluation is the main quality evaluation method at present, but the sensory evaluation is more or less subject to the body essence of the reviewers. Different judges may vary the quality of the same tea. Based on the study of the identification of Longjing tea origin, the prediction model of the quality of the flat tea and the discriminant model of the level of the flat tea are established, and the characteristic compounds and the flat tea which affect the quality of the flat tea are compared. The fundamental reason is that the metabolic compounds that affect the quality of the flat tea are mainly catechins, amino acids, alkaloids, theaflavins, and other catechin polymers, and the different compounds in the tea. The main research results are as follows: 1 the analysis of flat tea samples from four tea producing provinces in Zhejiang, Sichuan, Guizhou and Shandong was analyzed through the analysis of four tea producing tea in Zhejiang, Sichuan, Guizhou and Shandong. The metabolic fingerprint of the sample of the flat tea of the province was analyzed by multivariate statistical analysis. The results of PCA analysis were R~2X=0.643, Q~2X=0.363, and each tea producing province in the scatter plot showed the trend of polymerization. The results of PLS analysis, R~2X=0.396, R~2Y=0.874, Q~2X=0.898., identified the characteristic compounds of 32 different tea producing provinces, including Cysteinyl-Glycine, 3-Nitrotyro Sine, Mevalonic acid-5P, N-Carbamoyl-2-amino-2- (4-hydroxyphenyl) acetic acid, 3-O-p-Coumaroylquinic acid, Gallic acid4-O- (6-galloylglucoside), etc. Analysis of the flat tea samples from four tea producing areas in West Lake, Qian Tang, Yue Zhou and Jinyun through the metabolic fingerprint of the flat tea samples from four tea producing areas in West Lake, Qian Tang, Yue Zhou and Jinyun. Multivariate statistical analysis of the atlas, PCA analysis results R~2X=0.730, Q~2X=0.388, each tea producing province sample in the scatter plot showed the trend of polymerization; PLS analysis results R~2X=0.437, R~2Y=0.784, Q~2X=0.845. have identified 32 different tea producing provinces, including 5-Amino-6- (5'-phosphoribitylamino) uracil, 4-Methyl-2-phenyl-2-pent Enal, Ethyl aconitate, 1- (2-Hydroxyphenylamino) -1-deoxy-beta-D-gentiobioside 1,2-carbamate, Imidazoleacetic acid Ribotide and so on, the analysis of the flat tea samples from the main tea producing village in West Lake Longjing tea protection area was carried out by the metabolic fingerprint of the flat tea samples from the main tea producing village in the West Lake Longjing tea reserve. Multivariate statistical analysis, PCA analysis results R~2X=0.786, Q~2X=0.464, the scatter plot of the tea production samples each showed the trend of polymerization; PLS analysis results R~2X=0.453, R~2Y=0.848, Q~2X=0.895. identified 44 different tea producing provinces, including 4-O-Methylgallic acid, 3,4,5-Trimethoxycinnamic acid, Ethyl Boxylate, 3- (3,4-Dihydroxyphenyl) lactic acid, Dimethyl tetrasulfide and other.4 that affect the quality of flat tea, identified 60 compounds of the 1568 metabolic compounds detected by water extraction combined with UPLC-Q-TOF/MS by three grades. By t test, 23 of these compounds were found in A, B, and three regions. The score of sample review in the three interval is decreasing, in which the amino acid compounds such as L-Arginine and L-Glutamine have a significant difference between the regional A and the regional C. It is concluded that the compounds are positively related to the quality of the flat tea. Contrary to the amino acids, the three kinds of polypeptide compounds such as Serinyl-Serine are the lowest in the A region C. In the catechin compounds, Epigallocatechin showed a negative correlation with the quality of green tea, and the content of Catechin only in the B region was significantly lower than that in the C region. The remaining Gallocatechin, Gallocatechin gallate, Catechin gallate, Epigallocatechin gallate and other four compounds were different. The quality is positively correlated with the quality. The theaflavins and the flavanin compounds, including Theaflavin and Catechin- (4beta-8) -Epigallocatechin, are negatively correlated with the quality of green tea. Therefore, it is preliminarily believed that these compounds can be used to predict the quality of flat green tea,.5 a Comparison and analysis of the quality and grade model of flat tea under the conditions of alcohol extraction and water extraction, in high flux chromatography-mass spectrometry, the RMSEP value of the prediction model (PLS regression model) under the water extraction condition is less than the methanol extraction condition, while the R~2, Q~2 value is greater than the methanol extraction strip. The accuracy of the Bias network model is more than 90% under the water extraction condition, and the prediction degree is above 80%, which is far higher than the mean value of the accuracy of 72% and the mean of 62% in the methanol extraction condition. This shows that the water extraction method in the UPLC-Q-TOF/MS platform is more suitable for green tea review than the traditional methanol extraction method. A prediction of a score or grade.
【学位授予单位】：中国农业科学院
【学位级别】：硕士
【学位授予年份】：2016
【分类号】：S571.1

【相似文献】