CN103364389A - Analysis method for quickly detecting and judging albumin products - Google Patents
Analysis method for quickly detecting and judging albumin products Download PDFInfo
- Publication number
- CN103364389A CN103364389A CN2012100973670A CN201210097367A CN103364389A CN 103364389 A CN103364389 A CN 103364389A CN 2012100973670 A CN2012100973670 A CN 2012100973670A CN 201210097367 A CN201210097367 A CN 201210097367A CN 103364389 A CN103364389 A CN 103364389A
- Authority
- CN
- China
- Prior art keywords
- sample
- concentration
- prediction
- albumin
- matrix
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Images
Landscapes
- Investigating, Analyzing Materials By Fluorescence Or Luminescence (AREA)
Abstract
The invention relates to a quick nondestructive testing method of albumin products and particularly relates to an analysis method used for predicting albumin concentration through the characteristic spectrum segment of albumin and further verifying the identity of albumin products. The invention further relates to an analysis device executing the analysis method and an analysis system comprising the analysis device. The method, the device and the system provided by the invention combine a laser Raman spectrum with a partial least squares method to predict albumin contents in the albumin products, and compared with other quantitative techniques, the method, the device and the system have the advantages of simplicity in operation, rapid analysis, and the like.
Description
Technical field
The present invention relates to a kind of harmless method for quick of albumin products, relate to especially the analytical approach true and false of poor quality that the albuminous feature spectral coverage of a kind of employing is predicted albumin concentration and further differentiated albumin products, particularly human blood albumin products.
Background technology
Human serum albumin solution is for clarification, slightly sticky liquid, and is almost colourless, or yellow, brown color or green, comes from people's venous plasma, is prepared from through steps such as a series of precipitations, centrifugal, filtration, sterilizations.This medicine has the title of " life goods ", " help medicine " always, be mainly used in treating blood and hold the not enough acute and intensive care patient that causes of expansion, comprise malnutrition, burn, major injury, infection, major operation, hemorrhage, renal excretion excessively, pancreatitis, gestosis, hyperbilirubinemia of newborn, cirrhosis and the ephrosis oedema or the ascites that cause, the treatment of the aspects such as cancer operative results is a kind of special medicine of clinical emergency treatment.Its application is comparatively extensive, belongs to the well sold and in short supply medicine on the market.The situation in short supply of human serum albumin solution has continued to have a surplus in 2 years at home, nowadays more becomes worse.Thereby have the lawless person to make profit as one's only aim, and make and sell fake and inferior human serum albumin solution, lead extensive patients into a trap, upset medical market.
The false making of human serum albumin solution mainly contains two kinds of situations: one, do not contain human serum albumin; Two, the real content of human serum albumin is lower than the content that indicates on the sample packaging.
Demand for basic unit cracks down on counterfeit goods the sample of examination on a large scale regular need to have some accurate, quick, easy methods.In recent years, sample not being caused analysis means chemistry, machinery, photochemistry and thermal decomposition in analytic process, is one of the study hotspot in analysis science field.Along with the development of Chemical Measurement, spectroscopic methodology is subject to paying attention to widely both at home and abroad: near infrared spectrum is overlapping serious, and is not ideal enough to the mensuration precision of a lot of medicines; Infra-red sepectrometry requires quite high to the test environment humiture, sample making also bothers, and sense cycle is longer, can not satisfy the requirement of quick test; Raman spectroscopy have the identification of fingerprint of chemical molecular, fast, need not numerous and diverse sample pre-treatments, Non-Destructive Testing, be not subjected to the advantage such as hydrone interference, be applicable to the detection of counterfeit drug, can be used for the discriminating true and false of poor quality of human serum albumin.Present various human serum albumin solution quality standard is all used Chinese Pharmacopoeia appendix protein determination first method, and namely traditional Kjeldahls method carries out protein content determination, operates more loaded down with trivial details consuming timely, and part comes with some shortcomings when being applied to fast detecting.Reported in literature is arranged, adopt Raman spectroscopy that human serum albumin solution is carried out qualitative analysis and differentiate, still human serum albumin is not carried out so far the relevant report of assay.Adopt Raman spectrum to carry out quantitative test, need to eliminate the instability in measuring, the factor such as Removing Random No, sample background are disturbed on the one hand, the test sample device causes SPECTRAL DIVERSITY is on proofreading and correct resultant impact; Be the optimization of spectrogram information on the other hand, the SPECTRAL REGION that sample message is outstanding is selected, filter out the most effective SPECTRAL REGION, improve operation efficiency.At present, also not to the research report of the Raman spectrum quantitative test of human serum albumin solution.
Therefore, how at manpower, test site, test apparatus etc. all under the limited condition, adopt Raman spectrometer, develop accurate, easy-to-use Raman spectral information disposal route, thereby effectively and exactly human serum albumin is realized harmless, low-cost, easy, quick, as to be easy to penetration and promotion human serum albumin fast detecting, thereby reach timely discovery and monitoring management to the fake and inferior problem of the human serum albumin true and false, become one of numerous scientific researches, medicine inspection personnel outline.
Summary of the invention
In order to solve the problems referred to above of this area, the invention provides a kind of prediction albumin concentration and the further method true and false of poor quality of differentiating albumin products of novelty, the Raman spectrum that the method records sample carries out smothing filtering and eliminates noise, second derivative processing, quantitatively regretional analysis, then sets up model by the cross-validation method.
Therefore, aspect first, the invention provides a kind of method of predicting the albumin concentration of testing sample, the method comprises the steps:
1) obtains the Raman spectrogram of albumin standard items: the albumin standard solution of n series concentration as calibration sample, is obtained the Raman spectrogram of a described n calibration sample by Raman spectrometer;
Choose the Raman peaks scope corresponding with C=C symmetrical stretching vibration on the aromatic amino acid phenyl ring as the feature spectral coverage the Raman spectrogram of n the calibration sample that 2) selected characteristic spectral coverage: from step 1) obtains, described feature spectral coverage is preferably 960-1100cm
-1
3) Savitsky-Golay (following abbreviation S-G) smoothing processing: will from the Raman spectrogram of the standard solution of each concentration, the data of selected above-mentioned feature spectral coverage carry out multiple spot (preferred 9～21 points, more preferably 9～17 points, 13～17 points most preferably, can specifically select 9 points, 13 points, and 17 points at 15, preferred 13 and 17 points, most preferably 17 points) Savitsky-Golay smothing filtering elimination noise;
4) data that second derivative processing: to step 3) obtain are carried out second derivative and are processed;
5) spectrum matrix and the concentration matrix of n calibration sample carry out the regretional analysis of offset minimum binary standard measure in the spectroscopic data that offset minimum binary standard measure (PLS) regretional analysis: with step 4) obtains, and calculate the prediction concentrations of each calibration sample;
The prediction concentrations of each calibration sample that obtains 6) cross validation: to step 5) is carried out cross validation by the F-check, when acquisition reaches optimum main cause subnumber in described cross validation take actual concentration as horizontal ordinate, prediction concentrations is the typical curve of the described feature spectral coverage of ordinate, sets up thus albuminous partial least square method model;
7) obtain the Raman spectrogram of testing sample: by Raman spectrometer, identical Parameter Conditions when selecting with mensuration albumin standard solution obtains the Raman spectrogram of described testing sample;
8) concentration of prediction testing sample: choose the feature spectral coverage in the Raman spectrogram of testing sample, with as above face step 3 of this feature spectral coverage) and 4) describedly carry out multiple spot Savitsky-Golay smoothing processing and carry out second derivative and process, obtain the spectrum matrix of albumin solution to be measured, then use step 6) the partial least square method model that obtains carries out content prediction to described testing sample, obtains the prediction concentrations of described testing sample.
In the present invention, described albumin products can be the commercial articles such as bovine albumin, human serum albumin.
In the method for the albumin concentration of prediction testing sample of the present invention, second derivative is processed following carrying out:
First order derivative: f ' (x)=dy/dx
Second derivative: f " (x)=d^2y/dx^2=d (dy/dx)/dx
F ' (x) and f " (x) be respectively the mode of first order derivative, second derivative function representation, y is the absorbance log of Raman spectrum, and x is the wave number on the Raman spectrum.
In the method for the albumin concentration of prediction testing sample of the present invention, the quantitative regretional analysis of PLS is following to be carried out:
If A is that n calibration sample is at the spectrum matrix at m wavelength place, C is the concentration matrix of albumin in n calibration sample, E, F are respectively residual matrix, described spectrum matrix A Orthogonal Decomposition is the product of absorbance hidden variable matrix T and loading matrix P, and described concentration Matrix C Orthogonal Decomposition is the product of concentration hidden variable matrix U and loading matrix Q:
A(n×m)＝T(n×h)P(h×m)+E(n×m)
C(n×l)＝U(n×h)Q(h×l)+F(n×l)
Then, hidden variable matrix T, U are done linear regression, in parallel with diagonal matrix B:
U(n×h)＝T(n×h)B(h×h)，
To the sample that will predict in the forecast set, the spectrum matrix of establishing the described concentrated sample that will predict is A
Pre, then by:
A
pre＝T
preP
Can obtain T
Pre, then:
C
pre＝T
preBQ
T
PreBe the matrix that produces in the prediction concentrations computation process, C
PreConcentration for prediction.
In albumin concentration Forecasting Methodology of the present invention, cross validation can carry out according to known method is following:
(a) reject k sample from n calibration sample, described k is the common divisor of sample number, is n/4 to the maximum, and minimum is 1;
(b) come the calculating parameter matrix with n-k remaining sample, use the parameter matrix of trying to achieve to predict the concentration of a disallowable k sample, with the prediction concentrations C of a described k sample
I, preWith its concentration known C
iRelatively, can get its residual sum of squares (RSS) PRESS:
(c) deleted k sample recovered, reject k the sample of not yet rejecting again, calculate and go back to (b), the concentration of each calibration sample occurs once in PRESS, and only occurs once;
(d) last, the prediction residual quadratic sum when calculating the main cause subnumber and being 1 to h, when the prediction residual quadratic sum reach hour or the main cause subnumber of residual sum of squares (RSS) when no longer reducing as optimum main cause subnumber.
Aspect second, the invention provides a kind of genuine/counterfeit discriminating method of albumin products, may further comprise the steps:
1) obtains albumin sample to be measured, obtain the albumin sign concentration of testing sample by the packaging label of described testing sample;
2) above-mentioned Forecasting Methodology according to the present invention is predicted the albumin concentration of testing sample;
3) with the albumin concentration of the testing sample of above-mentioned prediction and the difference between the sign concentration separately and predetermined discrimination standard relatively, thus judge that described testing sample is genuine piece, suspicious specimen or adulterant.
In one embodiment of the invention, described predetermined discrimination standard is: if the relative deviation of the albumin concentration of the testing sample of prediction and its sign concentration is in ± 10%, can judge that then this testing sample is genuine piece, if the albumin concentration of prediction and its sign relative concentration deviation are ± 10%～± 20%, then be judged to be suspicious specimen, if prediction concentrations and sign relative concentration deviation then are judged to be adulterant greater than ± 20%.
Aspect the 3rd, the invention provides a kind of analytical equipment of the albumin concentration for predicting testing sample, this device comprises processor and controller, described controller comprises following modules:
1) data reception module: described data reception module receives by the testing sample of Raman spectrometer acquisition or the Raman spectrogram of calibration sample, and described calibration sample is the albumin standard solution of n series concentration;
2) the feature spectral coverage is chosen module: described feature spectral coverage is chosen module and is configured to choose the Raman peaks scope corresponding with C=C symmetrical stretching vibration on the aromatic amino acid phenyl ring as the feature spectral coverage from the Raman spectrum diagram data that data reception module receives, and described feature spectral coverage is preferably 960-1100cm
-1
3) S-G smoothing processing module: described S-G smoothing processing module is configured to the data of above-mentioned feature spectral coverage selected from above-mentioned Raman spectrogram are carried out multiple spot, preferred 9～21 points, more preferably 9～17 points, most preferably 13～17 S-G smothing filterings are eliminated noise;
4) second derivative processing module: described second derivative processing module is configured to process carry out second derivative through the data of S-G smoothing processing;
5) offset minimum binary standard measure regretional analysis module: the spectroscopic data that described offset minimum binary standard measure regretional analysis module is configured to process through second derivative carries out the regretional analysis of offset minimum binary standard measure, and calculates the prediction concentrations of each calibration sample;
6) partial least square method makes up module: the prediction concentrations to described each calibration sample is carried out cross validation by the F-check, when acquisition reaches optimum main cause subnumber in being chosen in cross validation take actual concentration as horizontal ordinate, prediction concentrations is the typical curve of the described feature spectral coverage of ordinate, sets up albuminous partial least square method model;
7) concentration prediction module: described concentration prediction module is configured to Raman spectrum diagram data to testing sample through above-mentioned module 2)-4) data after processing, the albuminous partial least square method model that makes up the module acquisition with described partial least square method carries out content prediction to obtain the prediction concentrations of described testing sample to testing sample.
Analytical equipment of the present invention can also comprise discrimination module as required, described discrimination module is configured to the albumin concentration of the testing sample of prediction is compared in conjunction with the sign concentration of described testing sample and predetermined discrimination standard, thereby judges that described testing sample is genuine piece, suspicious specimen or adulterant.
Comprising in the situation of discrimination module that analytical equipment of the present invention has in fact become a kind of discrimination and analysis device that can differentiate the medicine true and false.It will be understood by those skilled in the art that, described discrimination module can integrate with analytical equipment of the present invention as mentioned above, only be present in the analytical equipment of the present invention as a kind of function, perhaps also can be used as the discriminating gear that is independent of analytical equipment of the present invention and exist.Under latter event, this discriminating gear can be connected with any suitable data communication or electric connection mode with analytical equipment of the present invention, be combined into an analytic system of differentiating the medicine true and false, it perhaps only is the autonomous device of mancarried device or similar type, it can receive from the sample data of analytical equipment of the present invention or other analytical equipments (comprising concentration data, spectroscopic data etc.), according to its discrimination standard default or that be set by the user described sample data is differentiated.
In a preferred embodiment, described predetermined discrimination standard is: if the albumin concentration of the testing sample of prediction and its sign relative concentration deviation are in ± 10%, can judge that then this testing sample is genuine piece, if the albumin concentration of prediction testing sample and its sign relative concentration deviation are ± 10%～± 20%, then be judged to be suspicious specimen, if prediction concentrations and sign relative concentration deviation then are judged to be adulterant greater than ± 20%.
In the various embodiments of analytical equipment of the present invention, modules can arrange with reference to the various parameters, algorithm and the data processing method that adopt in the analytical approach of first and second aspects of the invention described above and the method for discrimination.
Aspect the 4th, the invention still further relates to a kind of albumin products analytic system, it comprises Raman spectrometer and analytical equipment of the present invention, and described analytical equipment can be connected communicatedly with described Raman spectrometer, and described communication connection can be via wireless or wired connection.
The invention has the advantages that does not need each albumin products is carried out loaded down with trivial details conventional chemical analytical control, greatly reduced the analytical work amount, detection efficiency and speed have been improved, can carry out fast examination to the human serum albumin solution on the market, avoid examination at random blindly, greatly reduced the cost of examination at random.Beneficial effect of the present invention is comprehensively as follows:
1, satisfying the requirement that the quick nondestructive to albumin solution detects, have that detection speed is fast, efficient is high, easy to operate, characteristics of not adding any reagent, is a kind of analysis test method of environmental protection.
2, the Raman spectrum that sample is recorded carries out smothing filtering and eliminates noise, second derivative processing, quantitatively regretional analysis, then set up quantitative model by cross-validation method, can realize the rapid screening to commercially available human serum albumin solution, harmless, easy, accuracy rate is high.
3, the present invention adopts laser Raman spectroscopy to combine with partial least square method, measures the content of human serum albumin solution, has the characteristics such as easy and simple to handle, that analysis is quick with respect to other quantitative techniques.
Description of drawings
Fig. 1 is the original Raman spectrogram of 9 groups of standard solution.
Fig. 2 is the linear graph of PLS.
Fig. 3 is the schematic flow sheet according to analytical approach of the present invention and method of discrimination.
Fig. 4 is the configuration schematic diagram according to an embodiment of analytical equipment of the present invention.
Embodiment
Term definition
The below provides the definition of more employed terms among the present invention.Not it should be noted that such as not definition that at this implication should be with the in the art employed common definition that those of ordinary skills were known accurately for it, or the definition in the instrument reference book of this area is as the criterion; As different from implication in this area, then be as the criterion with the definition among the present invention.
Feature spectral coverage: in the method for the albumin concentration of prediction testing sample of the present invention, because the packing of commercially available albumin solution is different, consider the variable thickness of vial, for guaranteeing that laser all focuses on solution inside, and make glass reduce to minimum to the interference of Raman signal, focal position is contrasted.Measure after will being respectively charged in the vial of standard jar and different manufacturers commercially available back with a albumin solution, found that focal position more the number of deeply convinceing substantially be reducing tendency, but when the part position in the standard jar and in the commercially available vial characteristic peak strength ratio of albumin solution more approaching.The analysis of the original Raman spectrum by the standard white protein solution finds that these curves of spectrum are regularity distribution, the increase of the intensity concentration of each raman characteristic peak in the human serum albumin solution of variable concentrations and increasing, comprehensive relatively under with C=C symmetrical stretching vibration (1004cm on the aromatic amino acid phenyl ring
-1± 4cm
-1) Raman peaks the most outstanding, so the selective light spectral limit is 960-1100cm
-1As feature spectral coverage of the present invention, the spectral coverage scope that namely various albumin solutions all is suitable for.
S-G is level and smooth: level and smooth the measured spectroscopic data of Raman spectrometer is accompanied by stochastic error and noise inevitably owing to be subjected to the impact of various factors, and the often rough curve that obtains smoothly can reduce high-frequency random noises.The Savitsky-Golay smoothing algorithm is most widely used spectrum smoothing method, with the spectroscopic data of polynomial expression under least square method, spectrum is carried out smoothing processing, also can be used for the differentiate of spectroscopic data.The essence of this algorithm is a kind of weighted mean, can avoid the distortion of spectrum peak shape.In the present invention, the signal to noise ratio (S/N ratio) of the Raman spectrogram by having weighed albumin standard solution and albumin solution to be measured selects 13 points, and 17 Savitsky-Golay smothing filterings comparatively suitable at 15, most preferably and 17 points at 15.
Second derivative: derivative method is during to Sample Scan, because sampling and instrument parameter etc. are on the impact of spectrum, other distortion of translation, rotation (along with diminishing of wave number, poor spectrum value presents rule and rises) and collection of illustrative plates may occur spectrogram.But low frequency background and constant term in the derivative method erasure signal reduce the power of high-order term, thereby high-frequency signal are displayed, and spectrum is carried out the rotation that the second order differentiate can be eliminated spectrum.The differentiate of spectrum can have more high-order, but Spectra Derivative can lose some low-frequency information when processing background correction and disturb, and can amplify stochastic error when eliminating systematic error.
Partial least square method (PLS): establish A and be n calibration sample at the absorbance matrix at m wavelength place, C is the concentration matrix of a kind of component (being specially in the present invention albumin) in n calibration sample, and E, F are respectively residual matrix.PLS is spectrum matrix A Orthogonal Decomposition the product of absorbance hidden variable matrix T and loading matrix P not only, is concentration Matrix C Orthogonal Decomposition the product of concentration hidden variable matrix U and loading matrix Q also:
A(n×m)＝T(n×h)P(h×m)+E(n×m)
C(n×l)＝U(n×h)Q(h×l)+F(n×l)
Then, hidden variable matrix T, U are done linear regression, in parallel with diagonal matrix B:
U(n×h)＝T(n×h)B(h×h)，
To the sample that will predict in the forecast set, establish its spectrum matrix A
Pre, then by:
A
pre＝T
preP
Can obtain T
Pre, then:
C
pre＝T
preBQ
T
PreBe the matrix that produces in the prediction concentrations computation process, C
PreBe the concentration of prediction, B, P and Q are such as top definition.
Partial least square method (PLS) utilizes principal component analysis (PCA) that absorbance matrix and concentration matrix are decomposed into respectively first proper vector and load vectors, then between these steady variablees, set up mutual relationship with partial least square method, thereby obtain the mathematic correction model (being the partial least square method model) between spectrum matrix and the concentration matrix.These be partial least square method in the principle of running background work, obtaining at last actual concentrations is horizontal ordinate, prediction concentrations is the linear diagram of ordinate.
Cross validation: cross validation (Cross-validation) is mainly used in the modelling application, when application PLS method is done quantitative test, dimension definite extremely important adopts the cross check method to determine dimension usually, and it can adopt following step to carry out:
Be provided with n calibration sample:
(a) from n calibration sample, reject k sample (k is the common divisor of sample number, is n/4 to the maximum, and minimum is 1);
(b) use n-k remaining sample to come the parameter matrix of computation model, use the model parameter matrix of trying to achieve to predict the concentration of a disallowable k sample, with the prediction concentrations C of a described k sample
I, preWith its concentration known C
iRelatively, can get its residual sum of squares (RSS):
(c) deleted k sample recovered, reject k the sample of not yet rejecting again, calculate and transfer back to (b), the concentration of each calibration sample occurs once in PRESS, and only occurs once.
(d) last, obtain a certain factor and count the factor number that the corresponding total PRESS minimum of h or PRESS no longer reduce.
In Forecasting Methodology of the present invention, in setting up the PLS regression modeling, from given modeling sample, take out most of sample and carry out established model, stay the fraction sample to forecast with the model of just setting up, and ask the prediction error of this fraction sample, record they square add and.This process is carried out always, until all samples have all been forecast once and only forecast once.The prediction error of each sample square added and, be called prediction residual quadratic sum (PRESS), use the validation-cross method, determine the main cause subnumber that reaches hour as PRESS, i.e. optimum main cause subnumber by the F-check.This Optimized model is a detection singular point, strong response point, determines all collaborative processes of square residual error RMSECV of optimum main gene and corresponding validation-cross.
The selection of main cause subnumber is directly connected to the actual prediction ability of model.If the main cause subnumber that uses when setting up model is very few, can not fully reflect the spectral information that the tested component of sample produces, cause insufficient match; If use the main cause subnumber too much, some have comprised the information of noise and have also mixed calculating, will cause over-fitting, thereby reduce the predictive ability of model.In the present invention, through screening and preliminary experiment, the main cause subnumber of selection is generally (being h≤10) below 10.
Analytical equipment of the present invention and the Pharmaceutical Analysis system that comprises this analytical equipment
In Fig. 4, illustrate the configuration schematic diagram according to an embodiment of analytical equipment of the present invention.In Fig. 4, analytical equipment 1 comprises processor, controller, described controller links to each other with data reception module, and the control data reception module receives the spectroscopic data (including but not limited to Raman spectrum data, ir data etc.) of standard items or testing sample from Raman spectrometer or other measuring equipments or database (local data base or be positioned at remote data base on the internet) under the instruction of processor.Be noted that if only be used for the concentration of prediction testing sample at this, then as shown in Figure 4, do not comprise discrimination module according to analytical equipment of the present invention.If need to differentiate the medicine true and false, then in another embodiment, can comprise that according to analytical equipment of the present invention discrimination module (discrimination module shown in dotted line frame among Fig. 4) is to be used for judging the true and false of poor quality of testing sample according to predetermined discrimination standard.
In Fig. 4, analytical equipment 1 according to the present invention comprises that also feature spectral coverage as hereinbefore defined chooses module, S-G smoothing processing module, and the second derivative processing module, the quantitative regretional analysis module of PLS, PLS makes up module, and the concentration prediction module.
Although do not illustrate, in analytical equipment 1 of the present invention, except the equipment of necessity such as processor, input-output unit, memory storage (such as internal storage, hard disk etc.), can comprise as required other hardware devices such as display device, printing device etc., can comprise that also network communication equipment is used for being connected with remote data base or carrying out exchanges data with other equipment by LAN (Local Area Network), internet, WiMAX or mobile communications network etc.It being understood that in the present invention processor can be understood as the function of independent executing data processing capacity or executing data processor and controller.In addition, the processor of controller of the present invention or implementation controller function can be carried out analytical approach of the present invention and required various algorithms, the data processing function of method of discrimination.
In the present invention, Raman spectrometer is used for obtaining the Raman spectrogram of testing sample and pharmaceutical standards product, as long as therefore can obtain the Raman spectrogram of sample, in analytic system of the present invention, can use the Raman spectrometer of Multiple Type, such as the DXR of ThermoScientific company and AlmegaXR, the LabRam of Horiba company, the inVia of Renishaw company etc.
In analytic system of the present invention, analytical equipment of the present invention is connected with Raman spectrometer and is connected with the transmission of data by various connected modes, include but not limited to cable directly connect, by wired or wireless LAN (Local Area Network) connection, the long-range connection in wired or wireless internet, bluetooth, infrared ray or by data transfer modes such as mobile networks.
Should be understood that analytical equipment of the present invention or system can realize in other mode, for example, as mentioned above, Raman spectrometer only uses as the instrument that obtains spectroscopic data, and adopts separately analytical equipment to analyze or process spectroscopic data.In addition, certainly, analytical equipment of the present invention can also be attached in the Raman spectrometer by any mode, at least can realize that purpose of the present invention gets final product, for example will be for the function of each step that realizes method of the present invention or Module-embedding hardware or the software to Raman spectrometer, such as the form of computer program or single-chip microcomputer in the mode of software or hardware or software and hardware combining.
Concentration prediction of the present invention and method of discrimination
Method of the present invention (comprising concentration prediction method or method of discrimination) can realize by the computer program of programming and by means of software for calculation, also can realize by hardware.The example that is used for realizing the software for calculation of method of the present invention is TQ Analysis, and they can be used singly or in combination.In addition, method of the present invention can also be write with language such as Matlab, VB, VC, C++.
Analysis of the present invention and method of discrimination can carry out according to schematic flow sheet shown in Figure 3:
The below illustrates the flow process of concentration prediction method of the present invention and method of discrimination according to the flow process of Fig. 3:
At step S1, the serial albumin standard solution that preparation concentration increases progressively.
At step S2, receive the spectroscopic data of albumin standard solution from Raman spectrometer.
At step S3, it is level and smooth that the spectrum that receives is carried out Savitsky-Golay.
At step S4, will carry out second derivative operator through the level and smooth spectrogram of Savitsky-Golay among the step S3, obtain spectrum after the new second derivative.
At step S5, will be to establishment of spectrum PLS quantitative model after the second derivative of step S4 acquisition.
At step S6, the Raman spectrum of the testing sample albumin solution of acquisition.
At step S7, the Raman spectrum of testing sample is repeated the data processing method of above-mentioned steps S3-S4.
At step S8, with the concentration of PLS quantitative model prediction testing sample.
At step S9, judge predicting the outcome, if the albumin concentration of the testing sample of prediction and its sign relative deviation are in ± 10%, can judge that then this testing sample is genuine piece, if albumin concentration and its sign relative concentration deviation ± 10%～± 20% of prediction, then be judged to be suspicious specimen, if prediction concentrations and sign relative concentration deviation then are judged to be adulterant greater than ± 20%.Suspicious specimen and adulterant are all inspected by random samples to relevant departments and are carried out Legal Inspection.
Below in conjunction with specific embodiment the true and false analytical approach of poor quality of fast detecting human blood albumin products of the present invention is described.
The foundation of embodiment 1 human serum albumin typical curve
1. the preparation of sample
Owing to can't obtain the standard items of human serum albumin, select commercially available human serum albumin solution (measuring its protein content through Chinese Pharmacopoeia version in 2010 three appendix VIB first method (Kjeldahls method)) to prepare the standard solution of a series of concentration as mother liquor.Get genuine piece human serum albumin concentration and be respectively 24.30% (A) and 4.88% (B), mutually dilution, the standard jar of packing into behind the mixing.Concrete configuration and CONCENTRATION DISTRIBUTION are as shown in table 1.
The configuration of table 1, human serum albumin standard solution
| Numbering | The volume of A/(ml) | The volume of B/(ml) | Concentration/(%) |
| 1 | 0 | 10 | 4.88 |
| 2 | 1 | 9 | 6.82 |
| 3 | 2 | 8 | 8.76 |
| 4 | 4 | 6 | 12.65 |
| 5 | 5 | 5 | 14.59 |
| 6 | 6 | 4 | 16.53 |
| 7 | 8 | 2 | 20.42 |
| 8 | 9 | 1 | 22.36 |
| 9 | 10 | 0 | 24.30 |
The used human serum albumin solution of this method is the commercially available sample of different brands, variable concentrations, different batches, amounts to 17, and the details of sample see Table 2.
Table 2, sample message table
2, spectra collection
Detect with Raman spectrometer.The LASER Light Source emission wavelength is 780nm, laser intensity 150mW, spectral scan scope 50～3411cm
-1, flat general specimen holder.With the albumin standard solution unified standard jar of packing into, each commercially available albumin solution to be measured is in original packing, avoid label, lie low on general specimen holder, suitably adjust the position bottle and groove are fitted fully, time shutter 10s, scanning times 10 times, gather Raman spectrum, gatherer process is 2 minutes.
3, data are processed
Obtain the Raman spectrogram of the more smooth human serum albumin solution of baseline except fluorescence background with data processing software (this software is Raman spectrometer standard configuration software) bales catch.Fig. 1 is the original Raman spectrum of 9 groups of standard solution, initial analysis finds that the curve of spectrum is regularity distribution, the increase of the intensity concentration of each raman characteristic peak in the human serum albumin solution of variable concentrations and increasing, comprehensive relatively under with C=C symmetrical stretching vibration (1004cm on the aromatic amino acid phenyl ring
-1± 4cm
-1) Raman peaks the most outstanding, so the selective light spectral limit is 960-1100cm
-1
The Raman spectrogram of the standard model that collects is carried out after 17 S-G smothing filterings eliminate noise, second derivatives and process, adopt PLS to carry out quantitative regretional analysis, by internal verification-cross-validation method (cross-validation), set up the PLS model of albumin standard solution.
In the structure of model, can calculate related coefficient (r) and cross validation mean square deviation (RMSECV).R illustrates that more near 1 the linearity of PLS match is better; RMSECV is less, and the linearity that match is described is less to the error of the prediction concentrations of modeling sample and actual concentrations, thereby explanation PLS model is better.
The concentration prediction of embodiment 2 testing samples
The PLS model that employing is set up is above measured the sample of commercially available human serum albumin solution, the results detailed in Table 3.
Table 3
17 human serum albumin samples to the special enamel agate of commercially available Losec, Shanghai Vaccine and Serum Institute, Shanghai Lay scholar, the emerging medicine in Shanghai, the learned pharmacy in Jiangxi, hundred special 6 producers are predicted, have contained three kinds of variable concentrations of commercially available human serum albumin: 5%, 20%, 25%.The demonstration that predicts the outcome, maximum deviation has proved that the method has higher accuracy in 3.5%.
Although by citation in the actual analysis specific detail and parameter declaration the present invention, should be appreciated that, this just by example mode rather than the mode of restriction describe, these are set and numerical value can change, these change still within the scope of the present invention.For the those skilled in the art that enlightened by the present invention, many changes and variation all are apparent.Therefore, the present invention is intended to comprise change and the variation of these scope of the invention.
Claims (10)
1. method of predicting the albumin concentration of testing sample, the method comprises the steps:
1) obtains the Raman spectrogram of albumin standard items: the albumin standard solution of n series concentration as calibration sample, is obtained the Raman spectrogram of a described n calibration sample by Raman spectrometer;
Choose the Raman peaks scope corresponding with C=C symmetrical stretching vibration on the aromatic amino acid phenyl ring as the feature spectral coverage the Raman spectrogram of n the calibration sample that 2) selected characteristic spectral coverage: from step 1) obtains;
3) Savitsky-Golay smoothing processing: will from the Raman spectrogram of n calibration sample, the data of selected above-mentioned feature spectral coverage carry out multiple spot, preferred 9～21 points, more preferably 9～17 points, most preferably 13～17 Savitsky-Golay smothing filterings are eliminated noise;
4) data that second derivative processing: to step 3) obtain are carried out second derivative and are processed;
5) spectrum matrix and the concentration matrix of n calibration sample carry out the regretional analysis of offset minimum binary standard measure in the spectroscopic data that offset minimum binary standard measure regretional analysis: with step 4) obtains, and calculate the prediction concentrations of each calibration sample;
The prediction concentrations of each calibration sample that obtains 6) cross validation: to step 5) is carried out cross validation by the F-check, when acquisition reaches optimum main cause subnumber in described cross validation take actual concentration as horizontal ordinate, prediction concentrations is the typical curve of the described feature spectral coverage of ordinate, sets up thus albuminous partial least square method model;
7) obtain the Raman spectrogram of testing sample: by Raman spectrometer, identical Parameter Conditions when selecting with mensuration albumin standard solution obtains the Raman spectrogram of described testing sample;
8) concentration of prediction testing sample: choose the feature spectral coverage in the Raman spectrogram of testing sample, with as above face step 3 of this feature spectral coverage) and 4) describedly carry out multiple spot Savitsky-Golay smoothing processing and carry out second derivative and process, obtain the spectrum matrix of albumin solution to be measured, then use step 6) the partial least square method model that obtains carries out content prediction to described testing sample, obtains the prediction concentrations of described testing sample.
2. the genuine/counterfeit discriminating method of an albumin products may further comprise the steps:
1) obtains albumin sample to be measured, obtain the albumin sign concentration of testing sample by the packaging label of described testing sample;
2) albumin concentration of method prediction testing sample according to claim 1;
3) with the albumin concentration of the testing sample of above-mentioned prediction and the difference between the sign concentration separately and predetermined discrimination standard relatively, thus judge that described testing sample is genuine piece, suspicious specimen or adulterant.
3. analytical equipment that is used for the albumin concentration of prediction testing sample, this device comprises processor and controller, described controller comprises following modules:
1) data reception module: described data reception module receives by the testing sample of Raman spectrometer acquisition or the Raman spectrogram of calibration sample, and described calibration sample is the albumin standard solution of n series concentration;
2) the feature spectral coverage is chosen module: described feature spectral coverage is chosen module and is configured to choose the Raman peaks scope corresponding with C=C symmetrical stretching vibration on the aromatic amino acid phenyl ring as the feature spectral coverage from the Raman spectrum diagram data that data reception module receives;
3) Savitsky-Golay smoothing processing module: described Savitsky-Golay smoothing processing module is configured to the data of above-mentioned feature spectral coverage selected from above-mentioned Raman spectrogram are carried out multiple spot, preferred 9～21 points, more preferably 9～17 points, most preferably 13～17 Savitsky-Golay smothing filterings are eliminated noise;
4) second derivative processing module: described second derivative processing module is configured to that the spectroscopic data through the Savitsky-Golay smoothing processing is carried out second derivative and processes;
5) offset minimum binary standard measure regretional analysis module: the spectroscopic data that described offset minimum binary standard measure regretional analysis module is configured to process through second derivative carries out the regretional analysis of offset minimum binary standard measure, and calculates the prediction concentrations of each calibration sample;
6) partial least square method makes up module: the prediction concentrations to described each calibration sample is carried out cross validation by the F-check, when acquisition reaches optimum main cause subnumber in being chosen in cross validation take actual concentration as horizontal ordinate, prediction concentrations is the typical curve of the described feature spectral coverage of ordinate, sets up albuminous partial least square method model;
7) concentration prediction module: described concentration prediction module is configured to Raman spectrum diagram data to testing sample through above-mentioned module 2)-4) data after processing, the albuminous partial least square method model that makes up the module acquisition with described partial least square method carries out content prediction to obtain the prediction concentrations of described testing sample to testing sample.
4. each described method or analytical equipment according to claim 1-3 is characterized in that described feature spectral coverage is 960-1100cm
-1
5. each described method or analytical equipment according to claim 1-4 is characterized in that described second derivative treatment step or module are configured to carry out the second order differentiate according to following formula:
First order derivative: f ' (x)=dy/dx
Second derivative: f " (x)=d^2y/dx^2=d (dy/dx)/dx
F ' (x) and f " (x) be respectively the mode of first order derivative, second derivative function representation, y is the absorbance log of Raman spectrum, and x is the wave number on the Raman spectrum.
6. each described method or analytical equipment according to claim 1-5 is characterized in that described offset minimum binary standard measure regretional analysis step or module are configured to:
If A is that n calibration sample is at the absorbance matrix at m wavelength place, C is the concentration matrix of albumin in n calibration sample, E, F are respectively residual matrix, described spectrum matrix A Orthogonal Decomposition is the product of absorbance hidden variable matrix T and loading matrix P, and described concentration Matrix C Orthogonal Decomposition is the product of concentration hidden variable matrix U and loading matrix Q:
A(n×m)＝T(n×h)P(h×m)+E(n×m)
C(n×l)＝U(n×h)Q(h×l)+F(n×l)
Then, hidden variable matrix T, U are done linear regression, in parallel with diagonal matrix B:
U(n×h)＝T(n×h)B(h×h)，
To the sample that will predict in the forecast set, the spectrum matrix of establishing the described concentrated sample that will predict is A
Pre, then by:
A
pre＝T
preP
Can obtain T
Pre, then:
C
pre＝T
preBQ
T
PreBe the matrix that produces in the prediction concentrations computation process, C
PreConcentration for prediction.
7. method according to claim 6 or analytical equipment is characterized in that cross validation step or partial least square method make up module and be configured to the following described cross validation that carries out:
(a) reject k sample from n calibration sample, described k is the common divisor of sample number, is n/4 to the maximum, and minimum is 1;
(b) come the calculating parameter matrix with n-k remaining sample, use the parameter matrix of trying to achieve to predict the concentration of a disallowable k sample, with the prediction concentrations C of a described k sample
I, preWith its concentration known C
iRelatively, can get its residual sum of squares (RSS):
(c) deleted k sample recovered, reject k the sample of not yet rejecting again, calculate and go back to (b), the concentration of each calibration sample occurs once in PRESS, and only occurs once;
(d) last, the prediction residual quadratic sum when calculating the main cause subnumber and being 1 to h, when the prediction residual quadratic sum reach hour or the main cause subnumber of residual sum of squares (RSS) when no longer reducing as optimum main cause subnumber.
8. each described analytical equipment according to claim 3-7, also comprise discrimination module, described discrimination module is configured to the albumin concentration of the testing sample of prediction is compared in conjunction with the sign concentration of described testing sample and predetermined discrimination standard, thereby judges that described testing sample is genuine piece, suspicious specimen or adulterant.
9. according to claim 2 or 8 described method or analytical equipments, it is characterized in that described predetermined discrimination standard is: if the albumin concentration of the testing sample of prediction and its sign relative concentration deviation are in ± 10%, can judge that then this testing sample is genuine piece, if the albumin concentration of prediction testing sample and its sign relative concentration deviation are ± 10%～± 20%, then be judged to be suspicious specimen, if prediction concentrations and sign relative concentration deviation then are judged to be adulterant greater than ± 20%.
10. Pharmaceutical Analysis system, comprise Raman spectrometer and according to claim 3-9 in each described analytical equipment, described analytical equipment can be connected communicatedly with described Raman spectrometer.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN2012100973670A CN103364389A (en) | 2012-04-05 | 2012-04-05 | Analysis method for quickly detecting and judging albumin products |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN2012100973670A CN103364389A (en) | 2012-04-05 | 2012-04-05 | Analysis method for quickly detecting and judging albumin products |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN103364389A true CN103364389A (en) | 2013-10-23 |
Family
ID=49366223
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN2012100973670A Pending CN103364389A (en) | 2012-04-05 | 2012-04-05 | Analysis method for quickly detecting and judging albumin products |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN103364389A (en) |
Cited By (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105092482A (en) * | 2014-05-08 | 2015-11-25 | 上海市食品药品检验所 | Method for rapidly measuring concentration of trichloro-tert-butyl alcohol in posterior pituitary injection |
| CN106770150A (en) * | 2016-05-19 | 2017-05-31 | 王桂文 | A kind of method of the single microsporidian spore trehalose concentration of fast quantification |
| CN107110776A (en) * | 2014-11-11 | 2017-08-29 | 光谱传感器公司 | Target analysis analyte detection and quantization in sample gas with complex background composition |
| CN109799224A (en) * | 2019-03-25 | 2019-05-24 | 贵州拜特制药有限公司 | Quickly detect the method and application of protein concentration in Chinese medicine extract |
| EP3971909A1 (en) * | 2020-09-21 | 2022-03-23 | Thorsten Kaiser | Method for predicting markers which are characteristic for at least one medical sample and/or for a patient |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN1157919A (en) * | 1995-06-28 | 1997-08-27 | 株式会社京都第一科学 | Method of optically measuring component in solution |
| US5712167A (en) * | 1995-07-19 | 1998-01-27 | Kyoto Dai-Ichi Kagaku Co., Ltd. | Method of measuring Amadori compound by light scattering |
| US6560478B1 (en) * | 1998-03-16 | 2003-05-06 | The Research Foundation Of City University Of New York | Method and system for examining biological materials using low power CW excitation Raman spectroscopy |
| WO2013096856A1 (en) * | 2011-12-22 | 2013-06-27 | Massachusetts Institute Of Technology | Raman spectroscopy for detection of glycated analytes |
-
2012
- 2012-04-05 CN CN2012100973670A patent/CN103364389A/en active Pending
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN1157919A (en) * | 1995-06-28 | 1997-08-27 | 株式会社京都第一科学 | Method of optically measuring component in solution |
| US5712167A (en) * | 1995-07-19 | 1998-01-27 | Kyoto Dai-Ichi Kagaku Co., Ltd. | Method of measuring Amadori compound by light scattering |
| US6560478B1 (en) * | 1998-03-16 | 2003-05-06 | The Research Foundation Of City University Of New York | Method and system for examining biological materials using low power CW excitation Raman spectroscopy |
| WO2013096856A1 (en) * | 2011-12-22 | 2013-06-27 | Massachusetts Institute Of Technology | Raman spectroscopy for detection of glycated analytes |
Non-Patent Citations (2)
| Title |
|---|
| D. M. HAALAND ，E. V. THOMAS: "Partial least-squares methods for spectral analysis. 1. Relation to other quantitative calibration methods and the extraction of quantitative information", 《ANALYTICAL CHEMISTRY》 * |
| DAHU QI，ANDREW J. BERGER: "Chemical concentration measurement in blood serum and urine samples using liquid-core optical fiber Raman spectroscopy", 《APPLIED OPTICS》 * |
Cited By (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105092482A (en) * | 2014-05-08 | 2015-11-25 | 上海市食品药品检验所 | Method for rapidly measuring concentration of trichloro-tert-butyl alcohol in posterior pituitary injection |
| CN107110776A (en) * | 2014-11-11 | 2017-08-29 | 光谱传感器公司 | Target analysis analyte detection and quantization in sample gas with complex background composition |
| CN106770150A (en) * | 2016-05-19 | 2017-05-31 | 王桂文 | A kind of method of the single microsporidian spore trehalose concentration of fast quantification |
| CN109799224A (en) * | 2019-03-25 | 2019-05-24 | 贵州拜特制药有限公司 | Quickly detect the method and application of protein concentration in Chinese medicine extract |
| EP3971909A1 (en) * | 2020-09-21 | 2022-03-23 | Thorsten Kaiser | Method for predicting markers which are characteristic for at least one medical sample and/or for a patient |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN104062257B (en) | A kind of based on the method for general flavone content near infrared ray solution | |
| CN101303294B (en) | Application method of near-infrared on-line test technology in Chinese medicine Yiqing granule production | |
| CN101852734B (en) | Fake medicine discrimination and analysis device, system and method | |
| CN103364389A (en) | Analysis method for quickly detecting and judging albumin products | |
| CN101275912B (en) | Fluid food browning testing method | |
| CN104931470A (en) | Fluorescence hyperspectral technology-based pesticide residue detection device and method | |
| CN103645155B (en) | The fast non-destructive detection method of fresh mutton tenderness | |
| CN108802000A (en) | A kind of lossless quick cholecalciferol-cholesterol content quantitative method based on the full spectrum analysis of Raman | |
| CN104048939A (en) | Near infrared rapid detection method for blood sugar content in live pig blood | |
| CN103411906B (en) | The near infrared spectrum qualitative identification method of pearl powder and oyster shell whiting | |
| CN106950192A (en) | A kind of method of Contents of Main Components quick detection in vegetable protein beverage based on near-infrared spectral analysis technology | |
| CN104777143A (en) | Method for similarity identification of expired drugs based on Raman spectroscopy | |
| CN103411955A (en) | Concentration detection method for mixed solution of cephalosporin antibiotics based on Raman spectrum | |
| CN103969211A (en) | A method for detecting moisture content of compound salvia tablets using near infrared spectroscopy | |
| CN106198447A (en) | Chemical Mixed Fertilizer main component harmless quantitative detection method based on near-infrared spectrum technique | |
| CN101620178B (en) | Method for quickly detecting additive chemical component in Chinese patent medicine, health-care food or food based on near-infrared spectrum technique | |
| CN103760130B (en) | The method of Tween-80 content near infrared ray compound Moschus injection | |
| CN104596976A (en) | Method for determining protein of paper-making reconstituted tobacco through ear infrared reflectance spectroscopy technique | |
| CN108195794B (en) | A kind of method that near infrared ray pheretima extracts Lumbrokinase potency in intermediate | |
| CN106706555A (en) | Milk powder determination method and system based on near infrared spectroscopy technology | |
| CN104792686A (en) | Method for detecting microbe quantity and drug content of semisolid preparation through near infrared spectroscopy | |
| Flåten et al. | PAT within the QbD framework: real-time end point detection for powder blends in a compliant environment | |
| CN109324018B (en) | Method for improving accuracy of protein content modeling basic data of near infrared spectrum analysis technology | |
| CN102879368B (en) | Fluorescence method for simultaneously measuring danofloxacin and flumequine in milk | |
| CN104819955B (en) | Ligusticum wallichii method and application are detected based on population least square method supporting vector machine algorithm |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20131023 |
|
| RJ01 | Rejection of invention patent application after publication |