CN103018382A - Detection method of fingerprint spectrum similarity - Google Patents

Abstract

The invention relates to a detection method of fingerprint spectrum similarity. The method is characterized by: adding a proper amount of an internal standard component that is not contained in a standard sample into the standard sample; measuring the peak areas of n components and the internal standard component in the standard sample, respectively taking the ratios of the peak areas of the n components to the peak area of the internal standard component as coordinates, and determining a coordinate point of a n-dimensional space; adding an equal amount of the same internal standard component into a to-be-detected sample, and acquiring a coordinate point determined by the to-be-detected sample in the n-dimensional space in the same way; and taking the distance between the coordinate point of the standard sample and the coordinate point of the to-be-detected sample in the n-dimensional space as the similarity, which can be greater when the distance is smaller. The greater the similarity is, the better the quality of the to-be-detected sample is. The method provided in the invention can accurately and efficiently distinguish the quality of a to-be-detected sample.

Description

A kind of detection method of fingerprint similarity

Technical field

The present invention relates to the detection method of fingerprint similarity, the field of quality control that belongs to the complicated ingredient system, be specifically related to the method for quality control of middle medicines natural products and the preparation take it as raw material, can determine accurately and efficiently the quality good or not of testing sample.

Background technology

Fingerprint pattern technology is because the stability that forms of Analysis of Complex chemical substance on the whole, become the Quality Evaluation Model that contains the complicated chemical composition of accepting extensively both at home and abroad, it can be than kind and the quantity of the contained chemical constitution of more comprehensive reflection, thereby estimate better the quality of Chinese medicine and preparation thereof, tobacco, food etc.In recent years, many countries have all accepted finger-print and have been used for estimating traditional Chinese medicine quality.U.S. FDA proposes to set up and to declare the chromatographic fingerprinting of medicinal material and preparation thereof in relevant herbal medicine quality control, criticize the consistance of a quality to investigate herbal products.British Herbal Pharmacopoeia, India herbal medicine allusion quotation, German medicinal plant association, Canada is medicinal and fragrant plant association etc. all finger-print as one of standard of traditional Chinese medicine quality control.

Form the finger-print of a practicality, at first to collect a large amount of qualified samples, set up finger-print by chromatography or spectroscopic methodology etc., select on this basis suitable method, form standard model, follow-up sample to be detected then is to detect similarity with standard model by certain detection method, estimates the situations such as the true and false, quality quality or steady quality of testing sample by the similarity extent index.

Similarity detection method commonly used, mainly be in standard model and testing sample, to choose a composition that jointly contains as the reference peak, determine the coordinate points of hyperspace with the ratio at all total peaks and reference peak, then utilize the similarity of correlation coefficient process or Cosin method examination criteria sample and testing sample.But the ratio of each contained composition is relevant in the existing detected similarity of method and the sample, and it is irrelevant with its content, contain all the components in the standard model such as testing sample, and each component content be in the standard model corresponding component content 1/2nd, with both similar situations of traditional technique in measuring, it is 100% similar to draw both.

Therefore, one both relevant with the ratio of contained each composition in the sample, and the simple and easy to do similarity detection method relevant with each component content, and accuracy and the range of application that promotes finger-print had very big help.

Summary of the invention

The objective of the invention is to overcome the defective of prior art, a kind of detection method of new fingerprint similarity is provided, the detected similarity of the method, not only relevant with the ratio of each composition in the sample, and, thereby accuracy and the range of application of raising finger-print relevant with the content of each composition.

The technical scheme that realizes the foregoing invention purpose is:

A kind of detection method of fingerprint similarity is characterized in that step is as follows:

A kind of detection method of fingerprint similarity is characterized in that step is as follows:

(1) in standard model, adds the composition that standard model does not contain, as interior mark;

(2) peak area of n composition and interior mark composition in the bioassay standard sample is respectively with the peak area (Y of n composition ₁, Y ₂... Y _n) and interior mark composition peak area (S ₂) ratio (Y ₁/ S ₂, Y ₂/ S ₂... Y _n/ S ₂) be coordinate, the coordinate B of sample at n-dimensional space settles the standard _Y=(Y ₁/ S ₂, Y ₂/ S ₂... Y _n/ S ₂) ^T

(3) in testing sample, add same interior mark composition, and make in its content and the standard model identical, with the peak area (X of the composition of n in the testing sample ₁, X ₂... X _n) with the peak area (S of interior mark composition ₁) ratio (X ₁/ S ₁, X ₂/ S ₁... X _n/ S ₁), the coordinate B of sample at n-dimensional space settles the standard _X=(X ₁/ S ₁, X ₂/ S ₁... X _n/ S ₁) T;

(4) with the coordinate points (B of standard model _Y) and the coordinate points (B of testing sample _X) the distance at n-dimensional space (d ( _{X, Y})) as the similarity detected value, distance is less, both are more similar, and the quality of testing sample is better.Described (d ( _{X, Y)}) specifically function is as follows:

$d_{\mathrm{ij}}=\sqrt{[\underset{k=1}{\overset{n}{\mathrm{\Σ}}}{(\frac{X_k}{S_1}-\frac{Y_k}{S_2})}^2}]$

Wherein: the number at the total peak of n-, X _kThe peak area at k peak of-testing sample; Y _kThe peak area S at k peak the in-standard model ₁The peak area S of the interior mark composition the in-testing sample ₂The peak area of mark composition in the-standard model.

The detection method of aforementioned fingerprint similarity is characterized in that interior mark composition random time before assay adds.

The detection method of aforementioned fingerprint similarity is characterized in that B _XWith B _YBetween the detection method that adopts of distance be Euclidean distance, mahalanobis distance or Pasteur's distance.

The detection method of aforementioned fingerprint similarity is characterized in that B _XWith B _YBetween the e of distance ^-xOr cos (atan (x)) function is as similarity, make its numerical value interval by (0 ,+∞) change (1,0).

Compared with prior art, the inventive method has the following advantages:

1, utilize the detected fingerprint similarity of this method, can be related with ratio and the content of each composition in the sample, thus enlarged accuracy and the range of application of finger-print.

2, because the interior index composition content that adds in standard model and testing sample is consistent, can reduce to a certain extent because the error that the chemical constitution checkout equipment causes.

3, because the coordinate points of the peak area that utilizes ingredient in the sample and the ratio formation hyperspace of the peak area of exogenous interior mark composition, standard model and testing sample are detected respectively on the detecting instrument of different brands becomes possibility, thereby improves the applicability of finger-print.

Description of drawings

Fig. 1 is the high-efficient liquid phase chromatogram that has added the stir-baked FLOS GENKWA with vinegar standard model of baicalein;

Mark 1-6,8-11 is 10 total peaks, and wherein 6 is cyanidenon, and 7 is baicalein, and 8 is apiolin, and 9 is Hydroxygenkwanin, and 10 is Genkwanin, and 11 is Yuanhuacine;

Fig. 2 is No. 1 stir-baked FLOS GENKWA with vinegar testing sample high-efficient liquid phase chromatogram that has added baicalein.

Embodiment

In order further to set forth technological means of the present invention, below in conjunction with drawings and Examples, embodiment and principle of work that a kind of fingerprint similarity detection method that foundation the present invention is proposed is applied to the stir-baked FLOS GENKWA with vinegar finger-print further specify as follows:

Lilac daphne is the dry flower of Isolated From Thymelaeaceae Species, is used for the oedema turgor, chest abdomen ponding, and phlegm and retained fluid is gathered, and the circulation of vital energy in the wrong direction is coughed and is breathed heavily, difficulty in urination and defecation.Because lilac daphne is poisonous, always the careful usefulness of doctor family commonly use clinically the stir-baked FLOS GENKWA with vinegar made from stir-baking with vinegar, but traditional fingerprint similarity detection method can't be distinguished living lilac daphne and stir-baked FLOS GENKWA with vinegar.So select detection method of the present invention, concrete operation step is as follows:

1. determine the standard model of stir-baked FLOS GENKWA with vinegar by early-stage Study, after the stir-baked FLOS GENKWA with vinegar standard model is pulverized, cross 60 mesh sieves, put dry 1h in 60 ℃ of baking ovens, accurately weighed 0.5g puts in the tool plug triangular flask, accurate 25mL 50% ethanol that adds, close plug, weighed weight, ultrasonic 30min, taking-up lets cool, supply the weight that subtracts mistake with 50% ethanol, shake up, add baicalein reference substance (interior mark composition), making baicalein concentration is 0.013mg/ml, and extract is crossed 0.45 μ m filter membrane.

2. filtrate is utilized high performance liquid chromatograph analysis, and chromatographic condition is chromatographic column: Dalian Yi Lite SinoChrom ODS-BP C ₁₈(250 * 4.6mm, 5 μ m); Mobile phase: 0.05% formic acid water (A)-0.05% formic acid acetonitrile (B); Flow velocity: 1.0mL/min; Detect wavelength: 332nm; Column temperature: 30 ℃; Sample size: 20 μ L; The gradient elution program sees Table 1.The final liquid chromatogram that forms is seen accompanying drawing 1, determines that wherein the 1-11 peak is the total peak of detection similarity, and read corresponding peak area from high performance liquid chromatograph.

Table 1 gradient elution program

Time	Gradient (A)	Gradient (B)
			0min	95％	5％
10min	85％	15％
			25min	80％	20％
35min	80％	20％
			50min	70％	30％
60min	70％	30％
			75min	25％	75％

80min	0％	100％
			90min	0％	100％

3. with 10 batches of living lilac daphnes to be measured (being called for short in the following form: " giving birth to product ") and 10 batches of stir-baked FLOS GENKWA with vinegar (being called for short in the following form: " vinegar product "), after pulverizing respectively, cross 60 mesh sieves, put dry 1h in 60 ℃ of baking ovens, accurately weighed 0.5g puts in the tool plug triangular flask, accurate 25mL 50% ethanol, the close plug of adding, weighed weight, ultrasonic 30min, taking-up lets cool, and supplies the weight that subtracts mistake with 50% ethanol, shake up, add the baicalein reference substance, make baicalein concentration also be 0.013mg/ml, extract is crossed 0.45 μ m filter membrane, filtrate is utilized high performance liquid chromatograph analysis, and chromatographic condition is chromatographic column: Dalian Yi Lite SinoChrom ODS-BP C ₁₈(250 * 4.6mm, 5 μ m); Mobile phase: 0.05% formic acid water (A)-0.05% formic acid acetonitrile (B); Flow velocity: 1.0mL/min; Detect wavelength: 332nm; Column temperature: 30 ℃; Sample size: 20 μ L; The gradient elution program sees Table 1.A final wherein width of cloth liquid chromatogram that forms is seen Fig. 2, reads the peak area with the corresponding total peak of standard model from high performance liquid chromatograph.

4. see the similarity of the living lilac daphne of lower each batch of detection and stir-baked FLOS GENKWA with vinegar and standard model according to following formula.Concrete outcome sees Table 2.More traditional correlation coefficient process, included angle cosine and the composition apiolin that contains take lilac daphne itself detect the similarity degree that each batch given birth to lilac daphne and stir-baked FLOS GENKWA with vinegar and standard model as interior target Furthest Neighbor simultaneously, and the result sees Table respectively 3, table 4, table 5.

$d_{\mathrm{ij}}=\sqrt{[\underset{k=1}{\overset{n}{\mathrm{\Σ}}}{(\frac{X_k}{S_1}-\frac{Y_k}{S_2})}^2}]$

(number at the total peak of n-, n=10 in this example, X _kThe peak area at k peak of-testing sample, Y _kThe peak area at k peak the in-standard model, S ₁The peak area of the baicalein the in-testing sample, S ₂The peak area of-standard model baicalein).

Table 2 is given birth to the similarity (the inventive method is take baicalein as interior mark) of lilac daphne and stir-baked FLOS GENKWA with vinegar and standard model

The product of giving birth to	Similarity	The vinegar product	Similarity
				R1	1.0579	C1	0.4279
R2	0.7051	C2	0.4591
				R3	0.6605	C3	0.3079
R4	0.7538	C4	0.4722
				R5	1.0320	C5	0.2327
R6	0.6653	C6	0.3131
				R7	0.8760	C7	0.3271
R8	0.6583	C8	0.3352
				R9	0.5830	C9	0.1073
R10	0.5906	C10	0.2755

Table 3 is given birth to the similarity (traditional correlation coefficient process) of lilac daphne and stir-baked FLOS GENKWA with vinegar and standard model

The product of giving birth to	Similarity	The vinegar product	Similarity
				R1	0.7610	C1	0.9648
R2	0.9571	C2	0.9599
				R3	0.9862	C3	0.9845
R4	0.8819	C4	0.9781
				R5	0.9776	C5	0.9900
R6	0.9904	C6	0.9928
				R7	0.9532	C7	0.9855
R8	0.9218	C8	0.9863
				R9	0.9693	C9	0.9976
R10	0.9743	C10	0.9982

Table 4 is given birth to the similarity (Cosin method) of lilac daphne and stir-baked FLOS GENKWA with vinegar and standard model

The product of giving birth to	Similarity	The vinegar product	Similarity
				R1	0.8904	C1	0.9837
R2	0.9802	C2	0.9815

R3	0.9931	C3	0.9927
				R4	0.9418	C4	0.9893
R5	0.9896	C5	0.9954
				R6	0.9955	C6	0.9966
R7	0.9782	C7	0.9930
				R8	0.9643	C8	0.9928
R9	0.9857	C9	0.9988
				R10	0.9881	C10	0.9992

Table 5 is given birth to the similarity (take apiolin as interior target Furthest Neighbor) of lilac daphne and stir-baked FLOS GENKWA with vinegar and standard model

The product of giving birth to	Similarity	The vinegar product	Similarity
				R1	1.3745	C1	0.5476
R2	0.5792	C2	0.4750
				R3	0.2710	C3	0.3563
R4	0.9124	C4	0.5753
				R5	0.3947	C5	0.2644
R6	0.1830	C6	0.1716
				R7	0.4883	C7	0.2651
R8	0.6401	C8	0.3001
				R9	0.5493	C9	0.1464
R10	0.5040	C10	0.1525

The result can find out the similarity of the living lilac daphne that detects with the present invention and standard model all more than 0.5, and the similarity of stir-baked FLOS GENKWA with vinegar and standard model can be distinguished living lilac daphne and stir-baked FLOS GENKWA with vinegar preferably all below 0.5.Correlation coefficient process, the similarity value that Cosin method detects is all at (except indivedual batches) more than 0.9, but the definite boundary value of neither one can be distinguished all and give birth to lilac daphne and stir-baked FLOS GENKWA with vinegar, the composition apiolin that contains take lilac daphne itself is as interior mark, with the detected similarity of distance coefficient method, giving birth to lilac daphne and stir-baked FLOS GENKWA with vinegar does not have definite boundary value to distinguish yet.Therefore make a living lilac daphne and stir-baked FLOS GENKWA with vinegar of the present invention provides an effective differentiating method.

Claims (5)

1. the detection method of a fingerprint similarity is characterized in that step is as follows:

(1) in standard model, adds the composition that standard model does not contain, as interior mark;

(2) peak area of n composition and interior mark composition in the bioassay standard sample is respectively with the peak area (Y of n composition ₁, Y ₂... Y _n) and interior mark composition peak area (S ₂) ratio (Y ₁/ S ₂, Y ₂/ S ₂... Y _n/ S ₂) be coordinate, the coordinate B of sample at n-dimensional space settles the standard _Y=(Y ₁/ S ₂, Y ₂/ S ₂... Y _n/ S ₂) ^T

(3) in testing sample, add same interior mark composition, and make in its content and the standard model identical, with the peak area (X of the composition of n in the testing sample ₁, X ₂... X _n) with the peak area (S of interior mark composition ₁) ratio (X ₁/ S ₁, X ₂/ S ₁... X _n/ S ₁), the coordinate B of sample at n-dimensional space settles the standard _X=(X ₁/ S ₁, X ₂/ S ₁... X _n/ S ₁) ^T

(4) with the coordinate points (B of standard model _Y) and the coordinate points (B of testing sample _X) the distance at n-dimensional space (d ( _{X, Y})) as the similarity detected value, distance is less, both are more similar, and the quality of testing sample is better.

2. the detection method of described fingerprint similarity according to claim 1 is characterized in that interior mark composition random time before assay adds.

3. the detection method of described fingerprint similarity according to claim 1 is characterized in that B _XWith B _YBetween the detection method that adopts of distance be Euclidean distance, mahalanobis distance or Pasteur's distance.

4. the detection method of described fingerprint similarity according to claim 1 is characterized in that B _XWith B _YBetween the e of distance ^-xOr cos (atan (x)) function is as similarity, make its numerical value interval by (0 ,+∞) change (1,0).

5. the detection method of described fingerprint similarity according to claim 1, it is characterized in that described (d ( _{X, Y})) specifically function is as follows:

$d_{\mathrm{ij}}=\sqrt{[\underset{k=1}{\overset{n}{\mathrm{\Σ}}}{(\frac{X_k}{S_1}-\frac{Y_k}{S_2})}^2}]$

Wherein: the number at the total peak of n-, X _kThe peak area at k peak of-testing sample; Y _kThe peak area S at k peak the in-standard model ₁The peak area S of the interior mark composition the in-testing sample ₂The peak area of mark composition in the-standard model.

Images