CN108398511B

CN108398511B - Quality detection method for dendrobium huoshanense

Info

Publication number: CN108398511B
Application number: CN201710064180.3A
Authority: CN
Inventors: 赵田
Original assignee: Beijing Lanbiao Yicheng Technology Co ltd
Current assignee: Beijing Lanbiao Yicheng Technology Co ltd
Priority date: 2017-02-04
Filing date: 2017-02-04
Publication date: 2021-03-16
Anticipated expiration: 2037-02-04
Also published as: CN108398511A

Abstract

The invention discloses a quality detection method of dendrobium huoshanense, which comprises the following steps: 1) sequencing by taking ITS-26SE and ITS-17SE as primers to identify the variety of the dendrobium medicinal material to be detected; 2) performing chromatographic detection on a sample with the sample capacity of n to obtain detection data with chemical small molecular components of schaftoside and/or naringenin as reference components; 3) respectively carrying out fingerprint detection on the samples to obtain fingerprint peak area values of all chemical components of dendrobium huoshanense; 4) the method comprises the steps of taking the content value of the schaftoside and/or naringenin component in chromatographic data as a response variable, taking the peak area value of other components in a fingerprint as an independent variable to establish an analysis model, and screening variables by a Lasso method to establish a fingerprint model of related characteristics of chemical micromolecule components. The quality of the dendrobium huoshanense medicinal material is accurately identified and controlled through first-generation sequencing and characteristic fingerprint spectrum.

Description

Quality detection method for dendrobium huoshanense

Technical Field

The invention belongs to the field of detection and analysis of traditional Chinese medicine components, and particularly relates to a quality detection method of dendrobium huoshanense.

Background

Dendrobium huoshanense (Dendrobium huoshanense C.Z.Tang & S.J.Cheng) is also known as Dendrobium huoshanense. The stem is upright, the meat quality is high, the stem grows densely, the stem grows 3 cm-9 cm, the stem becomes thin upwards obviously from the base, the diameter above the base is 0.3 cm-1.0 cm, the stem does not branch, the stem has 3 nodes-7 nodes, the internode length is 0.3 cm-0.8 cm, the stem is light yellow and green, and the stem is light purplish red speckles when dry and is light yellow. The leaf base has a sheath for holding the stem, the leaf sheath is membranous and lodged. 1-3 total inflorescences are formed, and the inflorescences are sent out from the upper part of the old stem of the fallen leaves and have 1-2 flowers; the inflorescence handle is 0.2 cm-0.3 cm long, and the base part is provided with 1-2 sheaths; the sheath is made of paper, the shape of the ovum is like a needle, the length is 0.3cm to 0.4cm, and the tip is sharp; the buds are light white with chestnut color, egg shape, length of 0.3-0.4 cm, sharp tip; the flower stalks and the ovaries are light yellow green and 2 cm-2.7 cm long; the flower is light yellow and green, and the development is carried out; the middle sepal is in the shape of an egg with a needle shape, the length is about 1.2cm to 1.4cm, the width is 0.4cm to 0.5cm, the tip is blunt, and 5 veins are provided; the side sepals are in the shape of sickle and are in the shape of needles, the length is 1.2 cm-1.4 cm, the width is 0.5 cm-0.7 cm, the tip is blunt, and the base is inclined; the sepal sac is nearly rectangular, 0.5 cm-0.7 cm long and nearly round at the tail end; the petal is oval, usually 1.2 cm-1.5 cm long, 0.6 cm-0.7 cm wide, blunt tip, 5 pulse; the labial lobe is nearly rhombic, the length and the width are approximately equal, the length is 1 cm-1.5 cm, the base part is wedged and provided with a corpus callosum, the upper part is slightly cracked, short hair is densely grown between the lobes at the two sides, and white hair is densely grown near the base part; the middle cleft sheet is in a semicircular triangle shape, the tip of the middle cleft sheet is close to a blunt tip, white hair grows densely at the base part of the middle cleft sheet, and the middle cleft sheet is provided with a yellow transverse elliptic plaque; the pistil is light green, is about 0.4cm long and has pistil feet with the length of 0.7 cm; the basal part of the stamen poda is yellow, white hair grows densely, and the two sides are provided with odontoseisis occasionally; the medicated cap is green white, nearly hemispherical, about 0.15cm long, and slightly concave at the top.

Dendrobe is always regarded as a precious Chinese herbal medicine and has very important nourishing efficacy. Clinically, dendrobium is used for treating various diseases and has the pharmacological effects of enhancing immunity, resisting oxidation, reducing blood sugar, inhibiting cancers and the like. Due to the fact that the dendrobium is mined artificially and unknowingly for a long time and is unreasonably utilized, wild resources of the dendrobium are gradually reduced, and phenomena of falseness and poor quality appear in the market. In addition, because the varieties of dendrobium are more, the characters of the related varieties are crossed due to the hybridization of the varieties of dendrobium, and the classification is difficult. Therefore, it is necessary to establish a characteristic fingerprint spectrum of dendrobium to evaluate the quality of dendrobium medicinal materials.

The chromatographic fingerprint is a comprehensive and quantifiable identification means, and is used as a panoramic mode of global analysis to reflect the overall condition of a sample. However, in the process of analyzing the chromatogram fingerprint, many data are high-dimensional, that is, the data contain many attributes or features, for example, regarding the chromatogram fingerprint of dendrobium huoshanense, the dendrobium huoshanense can be better described, but in practical application, the direct operation of the high-dimensional data will face the problem of "dimension disaster", which will cause the number of samples required by the modeling process to grow exponentially as the dimension increases. In the face of high dimensional data, the conventional least squares method is no longer applicable and variable selection becomes important in order to improve model interpretability and prediction accuracy. How to efficiently screen out a plurality of variables which play an important role in dependent variables from a plurality of variables is a problem which needs to be solved urgently when fingerprint spectrum is analyzed.

The quality of the traditional Chinese medicine is evaluated by quantitatively measuring the content of a certain active ingredient or effective ingredient, namely a micromolecular ingredient in the traditional Chinese medicine in the current national pharmacopoeia. However, the research proves that the curative effect of the traditional Chinese medicine comes from the synergistic effect among various active ingredients, even the commonly recognized effective synergistic effect or the 'Shengke effect' between the active ingredients and the inactive ingredients can achieve the curative effect of the traditional Chinese medicine, but not the result of the single action of a certain active ingredient. In the traditional Chinese medicine guided by the theory of traditional Chinese medicine, any one active ingredient can not comprehensively reflect the overall curative effect of the traditional Chinese medicine.

Disclosure of Invention

The invention provides a quality detection method of dendrobium huoshanense, which identifies varieties of dendrobium medicinal materials through first-generation sequencing, screens variables by using a Lasso method to establish a related characteristic fingerprint model of chemical micromolecule components in the dendrobium huoshanense, and accurately evaluates the quality of the dendrobium huoshanense medicinal materials through the first-generation sequencing and the related characteristic fingerprint model.

The purpose of the invention is realized by the following technical scheme:

a quality detection method of dendrobium huoshanense comprises the following steps:

1) with ITS-26 SE: 5 'GAATTCCCCGGTTCGCTCGCCGTTAC 3';

ITS-17 SE: 5 'ACGAATTCATGGTCCGGTGAAGTGTTCG 3' is used as a primer to carry out PCR amplification sequencing so as to identify the variety of the dendrobium to be detected as a dendrobium huoshanense sample;

2) performing chromatographic detection on a dendrobium huoshanense sample with the sample volume of n to obtain detection data with chemical small molecular components of schaftoside and/or naringenin as reference components;

3) respectively carrying out fingerprint detection on the samples to obtain fingerprint peak area values of all chemical components of dendrobium huoshanense;

4) taking the content value of the schaftoside and/or naringenin in chromatographic data as a response variable, taking the peak area value of other components in a fingerprint as an independent variable to establish an analysis model, screening variables by a Lasso (the Least adsorbed library scattering and selecting operator) method to establish a fingerprint model of related characteristics of chemical small molecular components, wherein the basic model is as follows:

y＝X^Tβ+ε

wherein y is a response variable, and y ═ y₁,y₂,...,y_n)^T(ii) a X is a matrix, X ═ X₁,x₂,...,x_n)^T；E(ε)＝0；Var(ε)＝σ²I_n(ii) a Epsilon is a random error term of the model; σ is the standard deviation of the random error term; n is the sample size; i is_nIs an n × n unit array.

The random term is assumed to obey classical assumptions, namely:

(1) the random term has a zero mean, E (ε_i|x_i)＝0；

(2) The random terms have the same variance, Var (ε)_i|x_i)＝σ²；

(3) Random term no sequence dependence, Cov (ε)_i,ε_j)＝0,i≠j；

(4) ε obeys a normal distribution_i～N(0,σ²)。

The variance matrix of the random term is one diagonal of sigma²And 0 elsewhere, as follows:

wherein, I_nIs a unit array of n x n, n being the sample size of the data,

further, the Lasso method is implemented by calculating according to formula i:

in formula I, n is the sample size; p is a radical of^*Is a variable number; p is the dimension of the sample; y ═ y₁,y₂,...,y_n)^T∈RⁿIs a response variable; x ═ x₁,x₂,...,x_n)^TA design matrix of n × p, containing all candidate independent variables having an influence on the response variable; lambda is an adjusting parameter;

is a penalty function; beta is a₀Is the intercept term of the formula, i.e., the value of the response variable y when all the independent variables x are 0; beta is a_jIs the independent variable x_jCoefficient of (2), i.e. argument x_jThe degree of influence on the response variable y.

Further, the selection method of the lambda is a K-fold cross-validation method:

K-fold CV：

wherein K is 5 or 10.

Further, the analysis model is a sub-model with the minimum CV value.

Further, the selection of λ follows the GCV criterion, which is defined as:

wherein, SSE_kIs the sum of the residual squares of the CV submodels containing k variables, df is trace { P (λ) }; trace represents the trace of the matrix. In linear algebra, the sum of the elements on the main diagonal (diagonal from top left to bottom right) of an n × n matrix a is called the trace (or trace number) of the matrix a, and is generally denoted as tr (a). That is, df is equal to the sum of all elements on the main diagonal in the matrix P (λ).

Further, the analysis model is a sub-model with the minimum GCV value.

Further, when the chromatographic data shows the ultra-high dimensional condition, firstly, the following SIS (sure Independence screening) method is adopted to screen variables, and then the Lasso method is utilized to process the variables;

SIS：M_γ＝{1≤i≤p:|ω_iis the first | γ n | larger }

Wherein M ═ {1 ≦ i ≦ p ≦ β_iNot equal to 0 represents a subscript set of non-zero coefficients in the true model; s ═ M_*L represents the number of nonzero coefficients; ω ═ ω (ω)₁,ω₂,...,ω_p)^T＝X^Ty；For any given γ ∈ (0,1), the p elements of ω are arranged and defined from large to small in absolute value; at this time, if gamma n is less than n, M is selected_γThe independent variable corresponding to the middle subscript is reduced from the ultrahigh dimension to the dimension d (d is less than or equal to n); wherein d is n or d is [ n/logn ═ n]。

The invention also provides application of the method in quality control of dendrobium huoshanense.

Compared with the prior art, the invention has at least the following advantages:

(a) according to the method, ITS-26SE and ITS-17SE are used as primers to determine the characteristic sequence of the dendrobium medicinal material so as to determine that the dendrobium medicinal material is a dendrobium huoshanense variety; then, taking chromatographic data of schaftoside and/or naringenin as independent variables, and taking chromatographic data of other components in the detection data as dependent variables to establish a linear regression model of micromolecule components and the fingerprint spectrum, so that the quality evaluation of the dendrobium medicinal material is more accurate;

(b) the method adopts the Lasso method to perform variable selection on the dendrobium huoshanense fingerprint, thereby effectively solving the problem of dimension disaster;

(c) the invention reduces the dimension of the original fingerprint, establishes the fingerprint of the relative characteristics of the dendrobium huoshanense chemical micromolecule components, and has stronger pertinence and applicability to the content explanation of the single chemical micromolecule component;

(d) according to the invention, the correlation analysis of the chemical small molecular component content is realized through the related characteristic fingerprint spectrum of the dendrobium huoshanense chemical small molecular component, and the quality of the dendrobium huoshanense medicinal material can be effectively identified and controlled;

(e) when chromatographic data shows the condition of ultrahigh dimension, the SIS method is adopted to reduce the dimension, and then the Lasso method is used for processing.

Drawings

FIG. 1 is a fingerprint chromatogram of the whole components of Dendrobium huoshanense;

FIG. 2 is a chromatogram of a standard sample of schaftoside chemical small molecule substance;

FIG. 3 is a chromatogram of a standard sample of naringenin chemical small molecule substance;

FIG. 4 is a chromatogram of a chemical small molecule of schaftoside in Dendrobium huoshanense;

FIG. 5 is a chromatogram of naringenin chemical small molecules in Dendrobium huoshanense.

Note: 1 peak schaftoside; naringenin peak 2.

Detailed Description

The present invention will be further described with reference to the following drawings and examples, which are illustrative only and not intended to be limiting, and the scope of the present invention is not limited thereby.

Example 1 Primary sequencing of Dendrobium huoshanense

First generation sequencing primer sequence:

ITS-26SE：5’GAATTCCCCGGTTCGCTCGCCGTTAC 3’；

ITS-17SE：5’ACGAATTCATGGTCCGGTGAAGTGTTCG 3’。

the amplification sequencing parameters were: performing PCR circulation after denaturation at 98 ℃ for 2min, wherein the PCR circulation parameter is 98 ℃ for 20 s; 30s at 52 ℃; 1min at 68 ℃, 38 cycles, 7min at 68 ℃, setting the temperature preservation at 4 ℃ after the amplification is finished, and performing first-generation molecular sequencing.

And identifying the variety of the dendrobium to be detected as dendrobium huoshanense by first-generation sequencing.

Example 2 extraction method of Dendrobium huoshanense

Taking a dendrobium huoshanense dry sample, crushing the dendrobium huoshanense dry sample by using a crusher, sieving the dendrobium huoshanense dry sample by using a pharmacopoeia sieve (the aperture is 0.335mm), precisely weighing 1.000g of dendrobium huoshanense powder (the weighing error cannot exceed 0.2%), putting the dendrobium huoshanense dry sample into a 100mL conical flask, respectively adding 50mL of 75% methanol (V water: V methanol is 25:75), performing ultrasonic treatment at room temperature for 30min, taking out the dendrobium huoshanense dry sample, filtering the dendrobium huoshanense dry sample, performing rotary evaporation and concentration on the filtrate until the filtrate is dry, dissolving the dendrobium huoshanense dry sample by using a 75% methanol solvent (V water: V methanol is 25:75), finally transferring the dendrobium huoshanense.

Example 3 chromatography detection method of Dendrobium huoshanense extract

Preparation of control solution

Precisely weighing schaftoside 4.10mg and naringenin 4.08mg respectively, placing into 10ml volumetric flasks respectively, adding 75% (V/V) methanol to dissolve and dilute, and shaking up to obtain stock solutions. Refrigerating at 4 deg.C in refrigerator.

And precisely absorbing a certain amount of reference substance stock solutions respectively, diluting with 75% methanol, and accurately preparing a schaftoside and naringenin mixed reference substance solution. The ingredients were diluted to prepare 7 concentration points by different dilution ratios. Injecting into high performance liquid chromatograph.

② a sample extraction and treatment method for measuring the content of small molecular components of dendrobium huoshanense:

weighing 1.00g of the powder (sieved by a third sieve), precisely weighing, placing in a 100ml volumetric flask, precisely adding 50ml of methanol-water (75:25), carrying out ultrasonic treatment (power 250W and frequency 40kHz) for 30 minutes, cooling, filtering, carrying out rotary evaporation and concentration on the filtrate until the filtrate is dry, dissolving the filtrate with 5ml of methanol-water (75:25), filtering the supernatant with a 0.45 mu m microporous filter membrane, and taking the subsequent filtrate to obtain the product.

③ measuring the chromatographic conditions of the small molecular component content of dendrobium huoshanense:

chromatographic conditions are as follows:

content determination chromatographic conditions: gracealitima C18 chromatography column (250mm x 4.6mm, 5 μm); the mobile phase adopts a binary gradient elution system, and the A phase: 0.2% acetic acid-water, phase B: acetonitrile; gradient elution procedure as in table 1; measuring naringenin at wavelength of 290nm, and measuring schaftoside at wavelength of 334 nm; the reference wavelength is 500nm, and the column temperature is 30 ℃; the flow rate was 1.0mL/min, and the amount of sample was 20. mu.L.

Fingerprint chromatogram conditions: a GraceAllitima C18 chromatography column, preferably a 250mm × 4.6mm, 5 μm format chromatography column; mobile phase: phase A: 0.4% acetic acid +20mmol/L ammonium acetate in water, phase B: acetonitrile; gradient elution: 0-12 min: 2% -15% of phase B, 12-35 min: 15% -24% of phase B, 35-45 min: 24% -36% of phase B, 45-60 min: 36-75% of phase B, 60-80 min: 75-95% of phase B; the flow rate is 1.0 mL/min; the column temperature is 30 ℃; the sample volume is 20 mu L; the detection wavelength is 280 nm.

TABLE 1 elution gradient for determination of content of small molecular components of Dendrobium huoshanense

FIG. 1 is a fingerprint chromatogram of the whole components of Dendrobium huoshanense, with a detection wavelength of 280 nm;

FIG. 2 is a chart of the detection map of a standard sample of schaftoside (peak 1);

FIG. 3 is a chart of a detection map of a naringenin (peak 2) standard sample;

FIG. 4 is a detection spectrum of schaftoside (peak No. 1) in Dendrobium huoshanense;

FIG. 5 is a detection spectrum of naringenin (peak No. 2) in Dendrobium huoshanense.

Example 4 establishment of fingerprint spectrum of relevant characteristics of dendrobium huoshanense schaftoside

1. Preparation of dendrobe sample solution

Taking a dried dendrobium sample, crushing the dendrobium sample by using a crusher, sieving the dendrobium sample by using a pharmacopoeia sieve (the aperture is 0.335mm), precisely weighing 1.000g of dendrobium powder (the weighing error cannot exceed 0.2%), placing the dendrobium powder into a 100mL conical flask, respectively adding 50mL of 75% methanol (V water: V methanol: 25:75), performing ultrasonic treatment at room temperature for 30min, taking out the dendrobium powder, filtering, performing rotary evaporation and concentration on the filtrate until the dendrobium powder is dried, dissolving the dendrobium powder by using a 75% methanol solvent (V water: V methanol: 25:75), finally transferring the dendrobium powder into a 10mL volumetric flask to fix the volume, shaking the dendrobium powder uniformly, and filtering the dendrobium powder by using a 0.45 mu m microporous filter membrane to obtain the. Table 2 shows the linear relationship of the obtained control schaftoside.

TABLE 2 Linear relationship table of the reference schaftoside

2. Method for establishing fingerprint spectrum of relevant characteristics of dendrobium huoshanense schaftoside

The first step is as follows: calculating correlation coefficients of all covariates x and y;

the second step is that: arranging the absolute values of the correlation coefficients from large to small, and selecting the first 2 √ n covariates, which are marked as x _1, x _2, … and x _ p;

the third step: and performing linear regression on y and x _1, x _2, … and x _ p, and performing variable selection by adopting a Lasso method.

The first part of the Lasso (LeastAbsource Shringkge and Selection operator) function represents the goodness of model fitting, and the second part can be considered as penalty. The method compresses small coefficients toward 0, and once a certain coefficient is compressed to 0, the corresponding variable is deleted. It is just like filtering with a "sieve", and the variables that have little influence are sieved off at a time. The smaller the λ, the more variables in the model the larger the λ, the larger the contraction, and the fewer variables are selected. While the Lasso method is a continuous, ordered process with small variance. When the tuning parameters are large enough, the penalty term has the effect of forcing the estimated values of some of the coefficients to be 0, so that LASSO can make variable selection, enabling a sparse model to be obtained.

When the independent variable is p and the sample size is n, and when p > n, the SIS method is adopted to reduce the dimension, and then the Lasso method is adopted to screen the variable.

SIS：M_γ＝{1≤i≤p:|ω_iIs the first | γ n | larger }

Wherein M ═ {1 ≦ i ≦ p ≦ β_iNot equal to 0 represents a subscript set of non-zero coefficients in the true model; s ═ M_*L represents the number of nonzero coefficients; ω ═ ω (ω)₁,ω₂,...,ω_p)^T＝X^Ty; for any given γ ∈ (0,1), the p elements of ω are arranged and defined from large to small in absolute value; at this time, if gamma n is less than n, M is selected_γThe independent variable corresponding to the middle subscript reduces the ultrahigh dimension to d (d is less than or equal to n) dimension; wherein d is n or d is [ n/logn ═ n]。

The linear model screened by the Lasso method is:

wherein, y_iIs the ith response variable, y ═ y₁,y₂,...,y_n)'；X_iIs P_nX1 order covariate, X ═ X₁,x₂,...,x_n)'；ε_iIs a mean of 0 and a variance of σ²I.d, E (E) ═ 0, and Var (E) ═ σ²I_n。

The random term is assumed to obey classical assumptions, namely:

(1) the random term has a zero mean, E (ε_i|x_i)＝0；

(2) The random terms have the same variance, Var (ε)_i|x_i)＝σ²；

(3) Random term no sequence dependence, Cov (ε)_i,ε_j)＝0,i≠j；

(4) ε obeys a normal distribution_i～N(0,σ²)。

wherein, I_nIs a unit array of n x n, n being the sample size of the data,

for simultaneous variable selection and parameter estimation, the Lasso method is implemented by penalizing the minimization of the least squares objective function formula i.

Wherein y ═ y₁,y₂,...,y_n)^T∈RⁿIs a response variable vector. Taking dendrobe data as an example, each dendrobe has two response variable sequences (schaftoside (mu g/g) and naringenin (mu g/g)). The response variable is affected by the independent variable, typically y is a continuous variable.

x＝(x₁,x₂,...,x_n)^TA matrix is designed for n × p, containing all candidate independent variables that have an effect on the response variable.

p is the dimension of the sample and n is the sample volume. In the dendrobe data, the dimension p is far larger than the sample capacity n, so the least square estimation is not applicable any more, and a variable selection method is required to be adopted for model estimation.

β＝(β₁,β₂,...,β_p)^TIs a p-dimensional parameter.

For the penalty function, λ is the adjustment parameter. In the variable selection method, the balance between the model fitting goodness and the punishment degree for the number of the selected variables is embodied by different criteria, and is realized by directly selecting the adjusting parameters, and different lambda values correspond to different punishment degrees. The larger the lambda is, the stronger the compression degree is, the fewer the non-zero parameters obtained by final estimation are, and the most common method for selecting the lambda is a K-fold cross-validation method:

K-fold CV：

in general, K can be 5 or 10.

The GCV criterion is an approximation of the CV criterion when K is taken to be n, and is defined as:

wherein, SSE_kIs the sum of the squared residuals of the CV submodel containing k variables, df is trace { P (λ) }. For the selection of the final optimal model, the sub-model with the minimum CV value or GCV value can be taken.

With the linear model, since the initial independent variables p are 333, the sample size n is 9, and p > n, the variable screening is an important task. After SIS dimensionality reduction, a Lasso method is adopted to screen variables.

The Lasso method resulted from 4 of the variables, with an R-square of 0.9848.

The variables from which the Lasso method screens and their corresponding coefficients are listed in table 3 below, where the first column is the number of the selected variable, the second column is the corresponding coefficient, the third column is the coefficient variance, and the fourth column is the test P value.

TABLE 3 Lasso method selected variables and their corresponding coefficients

X _ model (selected argument X)	Beta (coefficient of independent variable X)	Varbeta (variance of beta)	P-value (P value)
				Constant term	7.757795	1.398185	2.88E-08
120	0.016332	0.003653	7.8E-06
				83	-0.00419	0.002873	0.145201
225	0.011372	0.006713	0.090248
				243	0.007317	0.001999	0.000252

The results in table 3 show the results of the screening with schaftoside as the response variable: column 1 is the variable selected by the Lasso method, that is, the variables 120, 83, 225, and 243 are selected, and the corresponding p values (column 4) are all less than the significance level 0.05, and have significance difference. The meanings of the above independent variables: and aligning the fingerprint peaks of n batches of dendrobium huoshanense samples (n is not less than 10) according to the retention time.

Column 2 is the beta parameter value for each variable. The beta value is positive, which indicates that the variable has positive influence on the dendrobium schaftoside micromolecule; the beta value is negative, which indicates that the variable has negative influence on the dendrobium schaftoside micromolecule. The absolute value of the beta value shows the influence degree of the variable on the dendrobium schaftoside micromolecule. Specifically, in table 3, the effect of variables 120, 225, and 243 on dendrobe schaftoside small molecules is positive, with the positive effect of variable 120 being the greatest; the influence of the variable 83 on the dendrobe schaftoside small molecules is negative.

Example 5 establishment of fingerprint spectrum of relevant characteristics of Dendrobium huoshanense naringenin

1. Preparation of control solutions

Precisely weighing naringenin 4.08mg, placing in a 10ml volumetric flask, adding 75% methanol for dissolving and diluting, shaking up to obtain stock solution. Refrigerating at 4 deg.C in refrigerator. Precisely sucking a certain amount of reference substance stock solution, adding 75% methanol for dilution, and accurately preparing naringenin reference substance solution. The components are diluted to prepare 7 concentration points through different dilution ratios, and the concentration points are injected into a high performance liquid chromatograph. Table 4 shows the linear relationship of naringenin, which is the control product obtained.

TABLE 4 reference naringenin linear relationship table

2. Establishment of naringenin related characteristic fingerprint spectrum

With the linear model, since the initial independent variable p is 333, the sample size n is 10, and p > n, the variable screening is an important task. The invention adopts a Lasso method to screen variables.

The Lasso method resulted from 3 of the variables, with an R-square of 0.9203.

The variables from which the Lasso method screens and their corresponding coefficients are listed in table 5 below, where the first column is the input selected variable number, the second column is the corresponding coefficient, the third column is the coefficient variance, and the fourth column is the test P value.

TABLE 5 Lasso method selected variables and their corresponding coefficients

X _ model (selected argument X)	Beta (coefficient of independent variable X)	Varbeta (variance of beta)	P-value (P value)
				Constant term	3.226154	1.499972	0.031491
231	0.0075	0.003113	0.015995
				65	0.009792	0.001398	2.49E-12
214	-0.16808	0.145634	0.248446

The results in table 5 show the results with naringenin as the independent variable: the variables selected by the Lasso method are shown in column 1, namely, the variables 231, 65 and 214 are selected, and the p values (column 4) of the variables are all less than the significance level 0.05, and have significance differences. The meanings of the above independent variables: and aligning the fingerprint peaks of n batches of samples (n is not less than 10) of dendrobium huoshanense according to the retention time.

Column 2 gives the beta parameter values for each variable in particular. The beta value is positive, which indicates that the variable has positive influence on the dendrobe naringenin micromolecules; the beta value is negative, which indicates that the variable has negative influence on the dendrobium naringenin micromolecule. The absolute value of the beta value shows the degree of influence of the variable on the dendrobium naringenin small molecules. Specifically, in table 5, the effect of variables 231 and 65 on dendrobii naringenin small molecules is positive, with the positive effect of variable 65 being the greatest; the effect of variable 214 on the dendrobii naringenin small molecule is negative.

According to the results of the first generation molecular data, the technology can clearly distinguish dendrobium huoshanense from other dendrobium, has the function of identifying species, and can use a first generation sequencing molecular sequence as the identification index of the species. Meanwhile, the contents of two chemical micromolecules of the characteristic fingerprint of dendrobium huoshanense are stable in each sample in the species, and the sample can be clearly distinguished from other dendrobium and also can be used as one of identification indexes of the species. Therefore, a generation of data and a fingerprint map can be used as indexes for identifying the species, and the generation of data and the fingerprint map have the following association relationship: (1) when the first generation data identification sample is dendrobium huoshanense, the fingerprint of the sample has specific characteristics, namely, the first generation data can identify the species and the fingerprint characteristics of the species can be known; (2) when the fingerprint identification sample is dendrobium huoshanense, the generation data can be deduced. Therefore, the two are sufficient requirements, and the fingerprint characteristics of the species are stable in each representative sample, so that the content of the drug effect components of the species can be determined. Therefore, the first generation sequencing data and the fingerprint spectrum can be used as the identification indexes of the species and can be used as the evaluation indexes of the medicinal materials. If first-generation sequencing and fingerprint spectrum are respectively adopted for determination in variety identification and quality evaluation, the identification results of the two methods can be combined together, so that the identification result of the dendrobium medicinal material is more accurate and reliable.

Example 7 first generation sequencing of Dendrobium huoshanense and application of fingerprint of related characteristics of small molecular components in quality evaluation of Dendrobium nobile

1. And determining the sequence of the dendrobium medicinal material to be detected by using the next generation sequencing primer sequence and the amplification sequencing parameter, and identifying the dendrobium medicinal material to be detected as the dendrobium huoshanense medicinal material.

First generation sequencing primer sequence:

ITS-26SE：5’GAATTCCCCGGTTCGCTCGCCGTTAC 3’；

ITS-17SE：5’ACGAATTCATGGTCCGGTGAAGTGTTCG 3’；

the amplification sequencing parameters were: performing PCR circulation after denaturation at 98 ℃ for 2min, wherein the PCR circulation parameter is 98 ℃ for 20 s; 30s at 52 ℃; 1min at 68 ℃, 38 cycles, 7min at 68 ℃, and setting the temperature preservation at 4 ℃ after the amplification is finished.

Through the sequencing, the medicinal material to be detected is determined to be the dendrobium huoshanense variety.

2. Preparation of test sample

Precisely weighing dendrobium huoshanense powder, placing the dendrobium huoshanense powder into a 100ml volumetric flask, precisely adding 50ml of methanol-water with the volume ratio of 75:25 into each 1g of sample, carrying out ultrasonic treatment for 30 minutes at the power of 250W and the frequency of 40kHz, cooling, filtering, carrying out rotary evaporation and concentration on the filtrate until the filtrate is dry, correspondingly dissolving each 1g of dendrobium huoshanense powder with 5ml of methanol-water with the volume ratio of 75:25, passing the supernatant through a 0.45 mu m microporous filter membrane, and taking the subsequent filtrate to obtain the micromolecular component content determination sample of dendrobium huoshanense.

3. Chromatographic detection

Chromatographic conditions are as follows:

a chromatographic column: GraceAllitima C18 chromatography column (250 mm. times.4.6 mm, 5 μm); mobile phase: phase A: 0.4% acetic acid +20mmol/L ammonium acetate in water, phase B: acetonitrile; gradient elution: 0-12 min, 2-15% of B, 12-35 min, 15-24% of B, 35-45 min and 24-36% of B; 45-60 min, 36-75% B; 60-80 min, 75-95% B, and the flow rate is 1.0 mL/min; the column temperature is 30 ℃; the sample volume is 20 mu L; the detection wavelength is 280 nm.

The sample preparation method comprises the following steps:

During detection, measuring a full-component fingerprint chromatogram with the wavelength of 280nm, and comparing the obtained full-component fingerprint chromatogram with the fingerprint chromatogram which is the comparison in figure 1 in a similarity manner; the similarity is more than 0.85, and the quality is qualified.

The traditional Chinese medicine fingerprint is characterized by associating pharmacodynamic activity control with various quantitative characteristics which are obtained by detection of an analytical instrument and reflect the distribution of complex chemical substance components contained in the traditional Chinese medicinal materials, semi-finished products and traditional Chinese medicines (or botanicals), integrally reflecting the types, the quantities and the content characteristics of the chemical substance components contained in the traditional Chinese medicinal materials, the semi-finished products and the traditional Chinese medicines (or botanicals) macroscopically, and energetically revealing the spectrum of potential complex biological activity information characteristics.

According to the method, firstly, the types of the dendrobium are determined through first-generation sequencing, then the relevance research is carried out by measuring the content data of the small molecular components of the dendrobium and the whole peak area of the fingerprint of the dendrobium, and the intrinsic relevance of the dendrobium is found out through relevant data modeling, so that the quality of the dendrobium can be comprehensively evaluated, the quality of dendrobium huoshanense medicinal materials can be effectively and accurately identified and controlled, the analysis result is more reliable, and the interference of other types of dendrobium medicinal materials is avoided.

The above description is only for the preferred embodiment of the present invention, but the scope of the present invention is not limited thereto, and any changes or substitutions that can be easily conceived by those skilled in the art within the technical scope of the present invention are also included in the scope of the present invention. Therefore, the protection scope of the present invention shall be subject to the protection scope of the claims.

Claims

1. A quality detection method of dendrobium huoshanense is characterized by comprising the following steps:

1) with ITS-26 SE: 5 'GAATTCCCCGGTTCGCTCGCCGTTAC 3' and

ITS-17 SE: 5 'ACGAATTCATGGTCCGGTGAAGTGTTCG 3' is used as a primer to carry out PCR amplification sequencing so as to identify the variety of the dendrobium to be detected as dendrobium huoshanense;

2) performing chromatographic detection on a sample with the sample amount of n to obtain detection data with chemical small molecular components of schaftoside and/or naringenin as reference components;

4) taking the content value of the schaftoside and/or naringenin in chromatographic data as a response variable, taking the peak area value of other components in a fingerprint as an independent variable to establish an analysis model, when the chromatographic data shows an ultrahigh-dimensional situation, firstly adopting an SIS (sure Independence screening) method to reduce the ultrahigh dimension to d dimension, wherein d is n or d is [ n/log n ], screening variables by a Lasso (the Least Absolute library and Selection operator) method to establish a fingerprint model of related characteristics of chemical small molecular components, and the basic model is as follows:

y＝X^Tβ+ε

wherein y is a response variable, and y ═ y₁，y₂，…，y_n)^T(ii) a X is a matrix, X ═ X₁，x₂，…，x_n)^T；E(ε)＝0；Var(ε)＝σ²I_n(ii) a Random with epsilon as modelAn error term; σ is the standard deviation of the random error term; n is the sample size; i is_nIs an n × n unit array.