Method and system for differentiation of tea type

Patent metadata

Field	Value
Publication number	US-12590934-B2
Application number	US-202218078188-A
Country	US
Kind code	B2
Filing date	Dec 9, 2022
Priority date	Aug 24, 2020
Publication date	Mar 31, 2026
Grant date	Mar 31, 2026

Abstract

Disclosed are a tea type differentiation method and system, belonging to the technical field of detection. The method comprises: building a differentiation function by using ionic strengths of 20 compounds as evaluation indexes to discriminate tea types. According to the disclosure, the tea types are discriminated by using relative abundance of 20 compounds in tea, problems in sensory differentiation can be solved, the tea is classified more objectively and scientifically, and the reliability and accuracy of differentiation results are improved. By using three algorithms, the feasibility and accuracy of using 20 discovered compounds for tea type differentiation in a combined manner are validated.

First claim

What is claimed is:

1 . A tea type differentiation method, comprising: building a differentiation function by using ionic strengths of 20 compounds as evaluation indexes to differentiate tea types, wherein mass-to-charge ratios of the 20 compounds being as follows: 116.0648-116.0764, 267.1206-267.1474, 268.0906-268.1174, 280.1252-280.1532, 289.0561-289.085, 307.0657-307.0964, 308.0757-308.1065, 309.0814-309.1123, 364.0819-364.1183, 381.0604-381.0985, 425.0658-425.1083, 485.0833-485.1318, 518.2984-518.3503, 537.2765-537.3302, 554.1509-554.2063, 579.1207-579.1786, 607.2611-607.3218, 677.3378-677.4055, 744.2234-744.2978, and 869.1124-869.1993; wherein, for the 20 compounds, different tea type sample data is subjected to OPLS-DA, then, variables with VIP (Variable Importance in Projection) values greater than 1.5 are selected as candidate variables, and the candidate variables are further selected by a stepwise discriminant analysis method to finally obtain 16 compounds; further, through a parameter of FC (Fold Change) introduced, compound variables with FC>2, FC<0.5 and high ironic response intensity between green tea and yellow bud tea are selected, and 4 compounds are finally obtained through selection; and the compounds obtained through twice selection are aggregated to finally obtain 20 compounds for discriminating different tea types.

2 . The tea type differentiation method according to claim 1 , specifically comprising:
(1) performing preprocessing on tea samples, wherein the preprocessing comprises grinding the tea samples into powder, performing centrifugation to obtain a supernatant, and setting an internal standard; (2) detecting the tea samples to obtain each group of data matrixes comprising peak area, retention time and mass-to-charge ratio information of each of the samples, performing internal standard normalization by respectively dividing an ionic response intensity value of an obtained compound peak by an internal standard compound ionic response intensity value, and performing variable selection through OPLS-DA (Orthogonal Partial Least-Squares Discriminant Analysis) and stepwise discriminant analysis to obtain 20 characteristic compounds; (3) building tea type differentiation models by using internal standard normalization ionic response intensity values of the 20 compounds obtained from the tea samples; and (4) putting the internal standard normalization ionic response intensity values of the 20 compounds in tea samples to be differentiated into the built tea type differentiation models to obtain the types of the tea samples to be differentiated.

3 . The tea type differentiation method according to claim 1 , wherein, before the putting the internal standard normalization ionic response intensity values of the tea samples to be differentiated into different built tea type differentiation models, the method further comprises: collecting tea samples of each type, preprocessing and detecting the samples, processing and analyzing obtained data, and obtaining the internal standard normalization ionic response intensity values of the 20 compounds of the samples between the set mass-to-charge ratios.

4 . The tea type differentiation method according to claim 1 , mathematical methods for building the tea type differentiation models comprise a random forest method, a support vector machine method or a Fisher differentiation method.

5 . The tea type differentiation method according to claim 1 , wherein the building different tea type differentiation models further comprises: using data of the internal standard normalization ionic response intensity values of 20 compounds of different types of collected tea samples between the set mass-to-charge ratios as a data set of sample tea; and randomly dividing the data set of the sample tea into a training set and a validation set, wherein data of the training set is used for building the tea type differentiation models, data of the validation set is used for validating the built tea type differentiation models, and a ratio of the quantity of the sample tea samples of each tea type to the quantity of the tea samples to be differentiated is not smaller than 3:1.

6 . The tea type differentiation method according to claim 1 , wherein, according to the different built tea type differentiation models, the tea samples to be differentiated are detected to obtain data of the internal standard normalization ionic response intensity values of 20 compounds between the set mass-to-charge ratios, the data is put into the built tea type differentiation models to obtain a classification result of the tea samples to be differentiated.

7 . The tea type differentiation method according to claim 1 , wherein the tea types comprise one or more of green tea, yellow tea, dark green tea, white tea, black tea and oolong tea.

8 . A tea classification system, comprising:
a sampling module, configured to obtain tea mass spectrometric data corresponding to tea to be differentiated by using an LC-MS (Liquid Chromatography-Mass Spectrometry) technology; a classification module, configured to build a differentiation function by using ironic intensities of 20 characteristic compounds as evaluation indexes to perform classification processing on the obtained tea mass spectrometric data for obtaining a classification result of the tea to be differentiated, mass-to-charge ratios of the 20 compounds being as follows: 116.0648-116.0764, 267.1206-267.1474, 268.0906-268.1174, 280.1252-280.1532, 289.0561-289.085, 307.0657-307.0964, 308.0757-308.1065, 309.0814-309.1123, 364.0819-364.1183, 381.0604-381.0985, 425.0658-425.1083, 485.0833-485.1318, 518.2984-518.3503, 537.2765-537.3302, 554.1509-554.2063, 579.1207-579.1786, 607.2611-607.3218, 677.3378-677.4055, 744.2234-744.2978, and 869.1124-869.1993; a model building module configured to build tea classification models, wherein the model building module specifically comprises: a model building data obtaining submodule, configured to obtain sample tea mass spectrometric data corresponding to different types of sample tea and use a data set formed by the obtained sample tea mass spectrometric data as a sample tea mass spectrometric data set; a model building processing submodule, configured to randomly divide the obtained sample tea mass spectrometric data into a training set and a validation set, and perform model building processing on the training set by using a random forest method, a support vector machine method or a Fisher differentiation method to build and obtain tea type differentiation models; and a validation submodule, configured to validate a random forest model by using the validation set.

9 . An automatic tea separation device, comprising the tea classification system according to claim 8 .

10 . A tea type differentiation method, comprising: building a differentiation function by using ionic strengths of 20 compounds as evaluation indexes to differentiate tea types, wherein mass-to-charge ratios of the 20 compounds being as follows: 116.0648-116.0764, 267.1206-267.1474, 268.0906-268.1174, 280.1252-280.1532, 289.0561-289.085, 307.0657-307.0964, 308.0757-308.1065, 309.0814-309.1123, 364.0819-364.1183, 381.0604-381.0985, 425.0658-425.1083, 485.0833-485.1318, 518.2984-518.3503, 537.2765-537.3302, 554.1509-554.2063, 579.1207-579.1786, 607.2611-607.3218, 677.3378-677.4055, 744.2234-744.2978, and 869.1124-869.1993; wherein the building different tea type differentiation models further comprises: using data of the internal standard normalization ionic response intensity values of 20 compounds of different types of collected tea samples between the set mass-to-charge ratios as a data set of sample tea; and randomly dividing the data set of the sample te

Assignees

• Univ Anhui Agricultural
• Anhui Agricultral Univ

Inventors

• Wang Yijun
• Wan Xiaochun
• Kan Zhipeng
• Hu Lizhen
• Ning Jingming
• Li Daxiang

Classifications

• G01N30/8696

  Details of Software · CPC title

• G01N30/8693

  Models, e.g. prediction of retention times, method development and validation · CPC title

• G01N30/861

  Differentiation · CPC title

• G01N30/7233

  interfaced to liquid or supercritical fluid chromatograph (interfaces in general for introducing or extracting samples to be analysed with specially adapted mass spectrometer, see H01J49/04) · CPC title

• G01N30/14

  by elimination of some components · CPC title