JP3025145B2 - Multichannel chromatogram analysis method and apparatus - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to chromatography techniques such as high performance liquid chromatography and gas chromatography, and more particularly to a method and apparatus for analyzing a multi-channel chromatogram obtained by a diode array detector or the like.

[0002]

2. Description of the Related Art Many methods have been proposed for separating and resolving overlapping peaks on a multi-channel chromatogram (a chromatogram having three-dimensional information of an absorbance component, a time component, and a wavelength component). Typical examples are Factor analysis and Principal component analysis as one method of multivariate analysis (for example, (USP 4807148, JKS Strasters et.al.),
(Journal of Liguid Chromatography, 12 (1988) 3-22,
HR Keller et.al.), (Chemometrics and Intelligent
Laboratory Systems, 12 (1992) 209-224, J. Craig Ham
ilton et. al.), (Journal of Chemometrics, 4 (1990)
1-13) etc.).

[0003] The above factor analysis has an unsolved problem, and several improvements have been attempted. As another method, non-linear least-square analysis has also been proposed (for example, (Ito, et al., Proceedings of the 7th Symposium on Liquid Chromatography (1986), p.5). ),(Ito,
Et al., Proceedings of the 22nd Tokyo Symposium on Applied Spectrometry (1987), p.141) (Ito et al., Proceedings of the 19th HPLC Research Symposium (1988), pp.30). This method is called FJ Knorr
Method of fitting overlapping peaks on a multi-channel chromatogram obtained from GS / MS using Gaussian (normal distribution function) proposed by FJKnor
al., Aralytical Chemistry, 53 (1981) 821-825)
From EM as a more realistic model function
It is further developed by replacing with G (Exponentially modified Gaussian). This method will be described later in detail, for reference.

The data of the multi-channel chromatogram obtained from the diode array detector can be represented by a matrix Dij of the following equation (1). D = R · A · S = R · X (1) where i is the time sequence number, j is the wavelength sequence number, k is the component sequence number, Dij is the absorbance, and Rik is standardized. The retained waveform (chromatogram intensity), Akk is a quantitative factor relating to the concentration of the k-component, Skj is a normalized spectral intensity, and Xkj is a quantitative spectral intensity multiplied by a factor relating to the concentration of the k-component. In the case of a mass spectrometer, the wavelength number is the number of the m / z value.

Here, a trial matrix R′ik is introduced instead of the standardized chromatogram intensity matrix Rik. Using the trial matrix R'ik and the measured data matrix Dij from the above equation (1), a trial spectrum intensity matrix X'kj can be calculated first, and eventually a trial data matrix such as the following equation (2) D'ij can be obtained. D '= R' · (R 'T · R') -1 · R 'T · D --- (2) provided that, R' ^T of the T is the transposed matrix, -1 (R ^'T · R') Represents an inverse matrix.

In the least square method, parameters are determined so that Dij-D'ij is minimized, and R'ik of each component is best-fit. In this way, the matrices R'ik and X'ik of each component can be separated and decomposed from the matrix Dij containing the data of the overlapping peaks.

In this method, an EMG that can also represent an asymmetric tailing peak is used as the R'ik matrix (formula (3)).

[0008]

(Equation 1)

Where t _RK is the retention time of the _K -component, σ _K is the standard deviation of the k-component, and τ _K is the time constant of the k-component. JP-A-60-2447 discloses a chromatographic quantitative analysis method and apparatus for detecting a temporal change in absorbance at multiple wavelengths and performing quantitative analysis as three-dimensional information. In this chromatographic quantitative analysis method and apparatus, functions f1 (λ) and f2 (λ) indicating two standard two-dimensional spectra measured in advance are used. Then, the two-dimensional composite composition function fs (λ) of the measurement sample having the overlapping peak is calculated by using the above functions f1 (λ) and f2 (λ).
Calculate an equation that can be expressed using The overlapping peak is decomposed by the calculated equation.

As a method of analyzing a general chromatogram, that is, a chromatogram having two-dimensional information (absorbance component and time component), a method using deconvolution processing has been proposed (for example, see US Pat.
aul Benjamine Crilly), (IEEE Transaction on Instrum
entation and Measurement, 40 (1991) 558-562, PaulBen
jamine Crilly), (Journal of Chemometrics, 5 (1991)
85-95)).

Here, the deconvolution will be defined. The original data is detected by the detection device and is a detection characteristic of the detection device, and is expanded by a unique device function h (t) (spread function). The process of removing the data spread by the device function h (t) from the spread data is defined as deconvolution. The deconvolution definition formula is shown below.

[0012]

(Equation 2)

However, the detected waveform is D (t), the original waveform is d (t), and the spread function is h (t).

Several techniques have been proposed to speed up the convergence of deconvolution applied to the analysis of chromatograms having two-dimensional information. That is, Jacobi method, Gauss-Seidel method, Fourier transform method,
Van Cittert method, Constrained Iterative method, Jansson
Method, Gold's ratio method, etc. (in Japanese, edited by Shigeo Minami, "Waveform Data Processing for Scientific Measurement",
CQ Publishing Company (1986), pp.122-139, PAJansson, Decon
volution With Applications in Spectroscopy, New Yor
k, Academic (1984)).

[0015]

However, the above-described conventional method of separating and decomposing peaks by the least squares method is strong against noise and is practical, but on the other hand, some model function must be introduced, and the Is not always accurately reproduced. In other words, a model function is indispensable for the nonlinear least squares method, but the validity of this model function is questionable, and the best fit with the actual peak shape is not always achieved. This is because the synthesized composition function fs (λ) does not always accurately reproduce the actual peak shape, as described above, even in the chromatographic quantitative analysis method and apparatus described in JP-A-60-24447. I can't say that.

The deconvolution typified by the Jansson method is susceptible to noise, is greatly amplified after processing, and a pseudo peak often appears. Therefore, even if the above-mentioned conventional deconvolution is applied to the analysis of a three-dimensional multi-channel chromatogram, noise components cannot be removed, and high-precision peak separation and decomposition cannot be performed.

An object of the present invention is to realize a multi-channel chromatogram analysis method and apparatus capable of analyzing an overlapping peak of a multi-channel chromatogram with high precision and analyzing the same.

[0018]

The present invention is configured as follows to achieve the above object. In a multi-channel chromatogram analysis method for analyzing chromatogram data having a three-dimensional component including a characteristic component, a wavelength component, and a time component of a measurement object detected by a detection device, chromatogram data having a three-dimensional component From the compressed two-dimensional chromatogram data, and from the compressed two-dimensional chromatogram data, the detection characteristic of the detection device, and the spread of the data due to a unique device function. The method includes a deconvolution processing step of performing a deconvolution processing for removal, and a step of identifying a characteristic component of the measurement object from the two-dimensional chromatogram data from which the data spread has been removed.

Also, in the multi-channel chromatogram analysis method, a compression step of compressing chromatogram data having a three-dimensional component with respect to a wavelength component and converting it into two-dimensional chromatogram data, and a compressed two-dimensional chromatogram data From the detection characteristics of the detection device,
A deconvolution process for removing a spread of data by a normal distribution function that is an intrinsic device function, changing a standard deviation of the normal distribution function, and isolating overlapping peaks of the two-dimensional chromatogram data; and Performing a convolution process of adding each of the components of the two-dimensional chromatogram from which the data spread has been removed to the data before the deconvolution process by adding the data spread using an apparatus function; and Calculating the spectral information of each component based on the obtained data; identifying and quantifying the characteristic component of the object to be measured from the calculated spectral information; A process for displaying on the display means the restored data on which the volume processing has been executed. Provided with a door.

In the multi-channel chromatogram analysis method, three-dimensional chromatogram data
A deconvolution processing step that performs deconvolution processing to remove the spread of data due to a unique device function, which is the detection characteristic of the detection device, and compresses the deconvoluted chromatogram data with respect to time components to obtain the spectrum of each component. The method includes a step of calculating information and a step of identifying a characteristic component of the measurement target based on the calculated spectrum information.

Further, in a multi-channel chromatogram analyzing apparatus for analyzing chromatogram data having three-dimensional components of a measurement object, a wavelength component and a time component detected by the detecting means, And a data compression unit for compressing the chromatogram data having the wavelength component with respect to the wavelength component and converting the data into two-dimensional chromatogram data, and a detection characteristic of the detecting means based on the compressed two-dimensional chromatogram data. A deconvolution unit that performs a deconvolution process that removes the data spread due to the above, and a component identification unit that identifies the characteristic component of the measurement object from the two-dimensional chromatogram data from which the data spread is removed.

Also, the measuring object detected by the detecting means
The three-dimensional composition of the characteristic component of the object, the wavelength component, and the time component
Multi-chamber for analyzing chromatogram data
Three-dimensional chromatogram
The start point and end point of the time component to be analyzed
The data setting section to be set and the data setting section
Waveform chromatogram data with defined three-dimensional components
Two-dimensional chromatogram data compressed for long components
A data compression unit for converting the data into
From the chromatogram data, the detection characteristics of the detection means
Of the data using the normal distribution function
Performs deconvolution processing to remove the spread.
Change the standard deviation of the normal distribution function,
Isolate overlapping peaks in the original chromatogram data
Deconvolution part and deconvolution
Separate the processed chromatogram data into each component,
Separated standard part and the above components are separated and standardized
Using the above device function, the
Deconvolution by adding data spread every time
Perform convolution processing to restore data before processing
Re-convolution part and the restored data
Calculate the spectrum information of each component based on the data
A spectrum calculator and the calculated spectrum information
Component identification to identify and quantify the characteristic components of the measurement object
Quantity calculator and at least the identified and quantified characteristic components
And the restored data on which the convolution process was executed
And a display unit for displaying the data.

Also, the measuring object detected by the detecting means
The three-dimensional composition of the characteristic component of the object, the wavelength component, and the time component
Multi-chamber for analyzing chromatogram data
Three-dimensional chromatogram
Start of time and wavelength components to be analyzed
A data setting part in which a point and an end point are set;
From the three-dimensional chromatogram data with points
It is a detection characteristic of the above-mentioned detection means, and is based on a unique device function
Deconvolution processing to remove the spread of data
Deconvolution unit that performs
Standardized chromatogram data,
Compresses the time components and calculates the spectral information of each component
The standardized spectrum calculation unit to be calculated and the calculated spectrum
Based on the torque information, the characteristic component of the measurement object is identified,
And a component identification and quantification calculation unit for quantification.

In addition, the multi-channel chromatogram
In the analysis method, the object to be measured is determined based on the output of the detector.
Including information on characteristic components, wavelength components, and time components
To reduce the noise component of the data,
Compensate for spread with flow from reduced data
And outputs the result of the arithmetic processing.

Preferably, the multi-channel chromate
In the method of analyzing a togram, the noise component is reduced.
The step includes a smoothing process. Also, preferably,
In the analysis method of the multi-channel chromatogram ,
The step of performing the arithmetic processing includes a numerical analysis processing.

Also, preferably, the multi-channel
In the analysis method of the chromatogram, mosquito separating the sample
The sample separated by the ram is detected by the detector. Ma
Preferably, the above multi-channel chromatogram
In the analysis method, the detector may include a plurality of diodes.
Are arranged side by side.

In addition, the multi-channel chromatogram
In the analysis method, based on the output of the detector,
Data containing information on the characteristic components, wavelength components, and time components of the object.
Data, reduce the noise component of the data, and
Estimate the upstream state from the reduced data and calculate
And outputs the result of the arithmetic processing.

Further , the characteristic component, the wavelength component, and the
Multi-channel analysis of data containing information on time and time components
The above data is obtained by the analyzer of the channel chromatogram.
Noise reducer that reduces the noise component of
Compensate for spread with flow from reduced data
An operation processing unit that performs the operation and outputs the result of the operation processing
And an output device.

[0029]

In a multi-channel chromatogram having a three-dimensional component of a characteristic component, a wavelength component, and a time component of a measurement object, a noise component included in the chromatogram increases or decreases in the wavelength component direction. I have.
If the multi-channel chromatogram is compressed in the direction of the wavelength component, the noise components cancel each other. For this reason, the multi-channel chromatogram is compressed with respect to wavelength components and converted into two-dimensional chromatogram data.
This conversion process removes noise components. In the two-dimensional chromatogram data from which the noise component has been removed, the spread of the data due to the device function is removed by deconvolution processing, and the overlapping peaks are decomposed with high accuracy. Then, the components of the measurement object are identified based on the data decomposed with high accuracy.

The noise component included in the multi-channel chromatogram having a three-dimensional component increases and decreases in the time component direction. Therefore, the deconvolution processing is first performed on the multi-chromatogram data. Then, the data on which the deconvolution processing has been executed is compressed with respect to a time component for each component, and spectrum information of each component is calculated. At this time, the noise components cancel each other and are removed from the data. Then, the component of the measurement target is identified based on the spectrum from which the noise component has been removed.

[0031]

DESCRIPTION OF THE PREFERRED EMBODIMENTS An embodiment according to the present invention will be described with reference to FIGS. FIG. 1 is an operation flowchart of a multi-channel chromatogram analysis method according to one embodiment of the present invention, and FIG. 2 is a schematic configuration diagram of a multi-channel chromatogram analyzer according to one embodiment of the present invention.
The example of FIG. 2 shows that the present invention is applied to drug monitoring HPLC.
(High Performance Liquid Chromatography) system. The multi-channel chromatogram is composed of an absorbance (characteristic component) component, a wavelength component,
It has a three-dimensional component with a time component.

In FIG. 2, the pump 40 sends an eluent 49 according to a command from the control unit 44. A sampler 43 having a movable needle is used for a standard sample 52 contained in a sample container.
Alternatively, an unknown sample (measurement target) 51 is injected and sent to the flow path to the column 41. The sample 51 or 52 is sent to the column 41 together with the eluent 49, and the contained components are separated and developed, and detected by the diode array detector 42.
The multi-channel chromatogram, which is the detection data, is supplied to the data analyzer 45. In addition, components and the like to be described later are input to the data analysis unit 45 from the input unit 54.

The data analysis unit 45 includes a data matrix determination unit (data setting unit) 451, a data compression unit 452, a deconvolution unit 453, and a separation normalization unit 454.
And The data analysis unit 45 includes a reconvolution unit 455, a spectrum calculation unit 456, and a component identification quantitative calculation unit 457. The result analyzed by the data analysis unit 45 is displayed on the CRT 46 and the printer 47.

FIGS. 3 to 6 are multi-channel chromatograms of the unknown sample 51 obtained according to one embodiment of the present invention. In FIG. 3A, not only quantification but also component identification is difficult because component 2 and component 3 overlap. For this reason, it is necessary to separate and decompose the overlapping peaks, accurately determine the retention time, and accurately determine the spectrum, and identify the components. In addition, by accurately determining the size of the peak by separation and decomposition, quantitative analysis can be performed by comparison with the standard sample 52.

Next, the operation of the analyzing method and apparatus of the first embodiment according to the present invention will be described with reference to FIGS. 2, the data of the multi-channel chromatogram is supplied from the diode array detector 42 to the control unit 44 via the data matrix determination unit 451, and the CRT 46
3A, a contour map of a three-dimensional chromatogram as shown in FIG. 3A is displayed. The operation flow of FIG. 1 is started from a state where the contour map is displayed on the CRT 46.

The operation flow is started in step 11 of FIG. Next, in step 12, the components of the overlapping peak to be separated and resolved are specified. That is, if the operator
Referring to the contour map displayed at 46, keys, a mouse,
The overlapping peak components 2 and 3 shown in FIG. 3A are input from the input unit 54 to the data matrix determination unit 451 using a pen or the like. Alternatively, the coordinates of the peak components 2 and 3 are input by numerical values or by moving a line on the display.
Since the input of the peak component is performed twice, the fact that there are two overlapping peaks is recognized by the data matrix determination unit 45.
1 is determined. Further, it can be expected that a valley is formed between the two points of the overlapping peak component after deconvolution by this input.

Subsequently, in step 13, the data matrix determination unit 451 searches for the start point and the end point of the entire designated overlapping peak, and determines the start point and the end point of the time of the data matrix Dij shown in the above equation (1). Then, the data matrix Dij is determined. The number of measurement points from the start point to the end point is 20 to 5
0 is suitable, but if the number of points is more than this, it is desirable to reduce the number of points by an integrated average or the like. Regarding the number of points in the wave number direction, the number of measurement channels may be simply adopted.

Next, in step 14, the data compression unit 4
52 is the wavelength direction (λ direction) according to the following equation (4).
, Ie, summation, to obtain a compressed data sequence vector Di · (FIG. 3B).

[0039]

(Equation 3)

Here, the meaning of compressing the data matrix Dij in the wavelength direction as in the above equation (4) will be described. Although not shown, the multi-channel chromatogram shown in FIG. 3A usually includes many noise components (height direction noise components). This noise component is considered to oscillate in the wavelength direction plus or minus with respect to the predetermined reference level. Therefore, if the three-dimensional data is summed in the wavelength direction, the noises cancel each other out, and two-dimensional data Di · from which noise has been removed is obtained.

Next, a deconvolution process is performed on the two-dimensional data Di. That is, it is considered that the two-dimensional data Di · has been convoluted (expanded) by the following equation (5-1). Here, σ is a standard deviation, hi (σ) is a Gaussian-type device function represented by the following equation (5-2), and di · (σ) is a two-dimensional original data function.

[0042]

(Equation 4)

In step 15, the deconvolution unit 453 sets the standard deviation σ of the device function hi (σ). The initial value of the standard deviation σ is, for example,
1/10 of the time difference between two overlapping peaks specified in 2
Value. In addition, 1 of the time width of Di matrix
/ 30, a fixed value, an input value, etc. can also be adopted. Further, the initial value can be set to 0.

Subsequently, in step 16, the deconvolution unit 453 outputs the data Di.multidot.
And the function hi (σ) is used to calculate the function di · (σ). That is, deconvolution processing is performed on the data Di.
In this case, the deconvolution process can use a method such as the Jansson method.

Next, in step 17, the deconvolution unit 453 determines whether the data row vector after the deconvolution processing is sufficiently separated and the overlapping peak is isolated. As a criterion, it is used that the height of the valley between the peaks is 3% or less of the maximum value.
If this criterion is met, go to step 18. If so,
Proceeding to step 26, increase the standard deviation σ, return to step 15, and try deconvolution again. As the increment of the standard deviation σ, for example, the initial value of the standard deviation described above is used (however, when the initial value is set to 0, an appropriate value other than 0 is used).

In step 17, when an abnormally large pseudo peak or the like occurs, it is determined that the processing is defective, and the deconvolution unit 453 transmits an error signal to the control unit 44, and causes the CRT 46 to display the occurrence of the error. , Step 2
Proceed to 5 to end the flow. In step 17, when the deconvolution unit 453 determines that the peak is isolated (the deconvolution waveform in FIG.
(shown in (c)), the process proceeds to step 18, and the finally determined standard deviation is registered as the final standard deviation σ0.

Next, in step 19, the separation normalizing section 4
54 is an isolated data d using the following equation (6-1).
From i · (σ0), a normalization matrix rik (σ0) for each component is calculated. The separation / decomposition of the present method is performed in this step, and the processing is traced back from this point as a turning point. Next, in step 20, the re-convolution unit 455 calculates the following equation (6-
According to 2), each of the normalized matrices rik (σ0) is re-convoluted into the holding strength matrix R′ik using the same device function hi (σ0) as that at the time of deconvolution (FIGS. 4A to 4A). C
c)). Thereby, it is possible to express each waveform of the overlapping peak component of the data obtained by the detector 42 and containing no noise.

[0048]

(Equation 5)

Here, fk is a factor.

Subsequently, in step 21, the spectrum calculation unit 456 performs the processing shown in FIG.
As shown in (C), a quantitative spectral intensity matrix X'kj of each component is calculated. X ′ = (R ′ ^T R ′) ⁻¹ R ′ ^T D --- (7) Next, in step 22, the component identification quantitative calculation unit 457
Performs component identification based on the calculated spectral waveform of the quantitative spectral intensity matrix X′kj and the retention time of R′ik. The spectrum of the drug is stored in advance, and the library is searched to identify the component. Subsequently, in step 23, the component identification quantitative calculation unit 457
Using the magnitude of the matrix X'kj, a quantitative calculation is performed with reference to the quantitative spectrum intensity matrix of the standard sample 52.

In step 24, the results obtained by the above-described steps are displayed on the CRT 46 and printed out on the printer 47 by the selection of the operator. That is,
An image before deconvolution shown in FIG. 3B and an image of each reconvoluted component ((A in FIG. 4)
a), (Bb), and (Cc)) and the spectrum image of each component (FIG. 5)
(A), (B), (C)). In this case, FIG.
In the image before deconvolution shown in
The colors for displaying the components 2 and 3, which are the overlapping peaks, are different from each other, so that the overlapping can be clearly displayed. Then, the flow ends in step 25.

As described above, according to the embodiment of the present invention, the noise component is removed by compressing the data of the multi-channel chromatogram in the wavelength direction. Then, deconvolution processing is performed on the two-dimensional data from which the noise component has been removed. Therefore, it is possible to realize a multi-channel chromatogram analysis method and apparatus capable of analyzing the overlapping peaks of the multi-channel chromatogram with high accuracy and analyzing. Further, according to an embodiment of the present invention, each component subjected to deconvolution processing and separated is subjected to reconvolution processing by a device function. Then, each component subjected to the reconvolution processing is displayed, and based on each component subjected to the reconvolution processing, an image of unseparated data is displayed in different colors for each component. Therefore, it is possible to clearly indicate to the operator at which position the components overlap.

According to the embodiment of the present invention, the matrix R′ik and the matrix S′kj are obtained by deconvolution processing by applying a certain deformation to the peak shape. Can be released from

FIG. 7 is an operation flowchart of another embodiment of the multi-channel chromatogram analyzing method of the present invention, and FIG. 8 is a schematic configuration diagram of another embodiment of the multi-channel chromatogram analyzing apparatus. In FIG. 8, a data analysis unit 45 includes a data matrix determination unit (data setting unit).
451, a deconvolution unit 458, and a normalized intensity matrix calculation unit (data compression unit) 459. Further, the data analysis unit 45 includes a quantitative intensity matrix calculation unit 460.
And a component identification quantitative calculation unit 461. Other configurations are the same as those in the example of FIG.

The flow starts in step 61 of FIG. Next, in step 62, a region in the time direction and the wavelength direction of the multi-channel chromatogram for separating and decomposing the overlapping peak is input from the input unit 54 to the data matrix determination unit 451 and specified. In this case, the starting point and the ending point are generally designated by numerical values.
6 can be designated by moving the line on the contour map displayed. In this step 61, a numerical value is also input as to what component peak overlaps the designated area.

In step 63, the data matrix determination section 45
1 determines the boundary of the data matrix Dij from the start point and end point of each of time and wavelength. In this embodiment, since the deconvolution process is performed prior to data compression, the number of points for both time and wavelength is preferably about 10.

Next, in step 64, the deconvolution unit 458 sets a device function hi (σ). In this setting, a Gaussian-type device function is generally adopted in the same manner as in step 15 in FIG. 1, but the actually measured spread function may be used as the device function hi (σ). However, when the deconvolution process is performed more strongly by the following steps, it is simpler to adopt the Gaussian type.

Then, in step 65, the deconvolution unit 458 executes the same deconvolution processing as in step 16 shown in FIG. However, since the data matrix Dij is not compressed, processing of about 100 points is performed.

Next, in step 66, the deconvolution unit 458 determines whether or not there is a region having no overlap by the number of components, that is, whether or not the deconvolution processing is sufficient. This is a pass if the spectrum is isolated, for example, as shown by sections K = 2 and 3 in FIG. As a condition, it is desirable that the number of points in the section be three or more.

As a special case, as shown in a deconvolution-processed data r'i3 summation section, only one component peak appears in a certain wavelength section. One of the convolution-processed data R'i3 is obtained from the deconvolution-processed data r'i3, and the other data R'i2 can be determined. Of course, if each peak is isolated by the deconvolution processing, the convolution-processed data R'ik is obtained, and a pass is determined.

In the steps 65 and 66, the deconvolution process can be performed until the overlapping peaks are isolated. However, the deconvolution process that can obtain the spectral intensity matrix S'kj described later with high accuracy is sufficient ( Note: If the overlapping peaks can be resolved at some wavelength to such an extent that the peaks are isolated, a normalized holding intensity matrix R'ik can be obtained, so that the method of FIG.

If it is determined in step 66 that the test is rejected, the process proceeds to step 72, in which the width of the standard deviation σ is slightly widened and the apparatus function hi (σ) is updated, as in step 26 in FIG. Proceed to. If it is determined in step 66 that the test passes, the process proceeds to step 67.

In step 67, the normalized intensity matrix calculator 459 calculates a normalized spectrum intensity matrix S'kj according to the following equations (7-1) and (7-2), and calculates S'2j shown in FIG.
And a spectrum like S'3j.

[0064]

(Equation 6)

Based on the above equation (7-1), dij is calculated, and the calculated dij is substituted into the above equation (7-2) to calculate a normalized spectrum intensity matrix S'kj. However, i ₁ and i ₂ are the start and end points of time to specify the section of the chromatogram spectrum of k- ingredient.

When the peak is isolated in the time axis direction, a normalized holding strength matrix R'ik is calculated according to the following equation (8).

[0067]

(Equation 7)

Here, j ₁ and j ₂ are the starting point and the ending point of the wavelength section in which the holding function of the k-component is obtained.

Here, as described above, a multi-channel chromatogram usually contains many noise components. This noise component is, for a predetermined reference level,
It is considered that the vibration oscillates positively and negatively in the time axis direction. Therefore, if the three-dimensional data is summed in the time axis direction, the noises cancel each other out, and a normalized spectrum intensity matrix S'kj and a normalized holding intensity matrix R'ik from which noise has been removed are obtained.

Next, in step 68, the quantitative intensity matrix calculator 460 calculates a quantitative retained intensity matrix R'ikA'kk based on the following equation (9). R′A ′ = DS ′ ^T (S ′S ′ ^T ) ^{−1 −1} (9) If the normalized holding strength matrix Rik has been calculated in step 67, the following equation (10) is used. , A quantitative spectral intensity matrix X′kj is calculated. X ′ = (R ′ ^T R ′) ⁻¹ R ′ ^T D --- (10) Then, in step 69, the component identification quantitative calculation unit 461
Using the calculated spectrum intensity matrix and the holding intensity matrix, component identification,
Perform a quantitative calculation.

Subsequently, in step 70, step 2 in FIG.
As in the case of No. 4, the result obtained by the above steps was
And prints out to the printer 47 by the selection of the operator. Then, the process ends in step 71.

As described above, according to another embodiment of the present invention, the noise component is removed by compressing the data of the multi-channel chromatogram in the time axis direction.
Then, the normalized spectrum intensity matrix S′kj and the normalized holding intensity matrix R′ik from which the noise component has been removed can be calculated. Therefore, as in the examples of FIGS. 1 and 2, it is possible to realize a multi-channel chromatogram analysis method and apparatus capable of analyzing the overlapping peaks of the multi-channel chromatogram with high accuracy and analyzing.

Also, as in the examples of FIGS. 1 and 2, the function can be released from the model function. In another embodiment of the present invention, a method of obtaining a spectrum from the hem of a conventionally well-known overlapping peak (a portion where the influence of overlap is small) and decomposing the peak is supported by a decovolution process, and expanded. It is thought that it was done. When three or more components overlap, it was difficult to decompose in the prior art, but can be decomposed in the above embodiment.

FIG. 9 is a diagram showing an example of a three-dimensional chromatogram obtained by performing deconvolution processing without executing data compression processing, and is a comparative example between the present invention and the prior art. As shown in FIG. 9, when the three-dimensional chromatogram is simply deconvolved,
Noise and spurious peaks appear. (B) of FIG.
(A) shows a cross section along the line AA ′. Thus, simply deconvolution processing of a three-dimensional chromatogram is equivalent to the existence of a two-dimensional chromatogram having several different wavelengths. In this case, the information of the entire three-dimensional chromatogram is not used and processed, and the disadvantage that the noise and the pseudo peaks that were present in the case of the two-dimensional chromatogram are amplified remains. .

On the other hand, the present invention, as described above,
In order to overcome the disadvantages of the prior art, data is compressed and noise is canceled out.

In the above embodiment, general smoothing and baseline processing are appropriately performed in the deconvolution processing.

In addition to the initial value of the standard deviation of the device function, the input value of the plurality of estimated retention times and the initial value of the fitting parameter of the time constant for specifying the tailing peak are input from the input unit 54 to the data analysis unit 45. Can be entered. The above-described example is an example in which the present invention is applied to a drug monitoring HPLC system. However, the present invention is not limited to the above-described system, and is applicable to other chromatogram analysis systems.

In the above example, the characteristic component of the object to be measured is the absorbance. However, the present invention can be applied to the case of a fluorescence intensity.

[0079]

Since the present invention is configured as described above, it has the following effects. The multi-channel chromatogram is compressed with respect to wavelength components, converted into two-dimensional chromatogram data, and noise components are removed. In the two-dimensional chromatogram data from which the noise component has been removed, the spread of the data due to the device function is removed by deconvolution processing, and the overlapping peaks are decomposed with high accuracy. Then, the components of the measurement object are identified based on the data decomposed with high accuracy. Therefore, it is possible to realize a multi-channel chromatogram analysis method and apparatus capable of analyzing the overlapping peaks of the multi-channel chromatogram with high accuracy and analyzing.

The multi-chromatogram data is first subjected to the deconvolution processing. Then, the data on which the deconvolution processing has been executed is compressed with respect to a time component for each component, and spectrum information of each component is calculated. At this time, the noise components cancel each other and are removed from the data. Then, the component of the measurement target is identified based on the spectrum from which the noise component has been removed. Therefore, as described above, it is possible to realize a multi-channel chromatogram analysis method and apparatus capable of analyzing an overlapping peak of a multi-channel chromatogram with high accuracy and analyzing the same.

[Brief description of the drawings]

FIG. 1 is an operation flowchart of a multi-channel chromatogram analysis method according to an embodiment of the present invention.

FIG. 2 is a schematic configuration diagram of a multi-channel chromatogram analyzer according to one embodiment of the present invention.

FIG. 3 is a diagram showing a data display example of a multi-channel chromatogram.

FIG. 4 is a waveform diagram showing a holding strength matrix of a component k subjected to reconvolution processing.

FIG. 5 is a waveform chart showing a spectrum intensity matrix of a component k.

FIG. 6 is a waveform diagram of a data matrix subjected to deconvolution processing.

FIG. 7 is an operation flowchart of a multi-channel chromatogram analysis method according to another embodiment of the present invention.

FIG. 8 is a schematic configuration diagram of a multi-channel chromatogram analyzer according to another embodiment of the present invention.

FIG. 9 is a waveform chart for comparing the present invention with a conventional technique.

[Explanation of symbols]

 Reference Signs List 40 pump 41 column 42 diode array detector 43 sampler 44 control unit 45 data analysis unit 451 data matrix determination unit 452 data compression unit 453, 458 deconvolution unit 454 separation normalization unit 455 reconvolution unit 456 spectrum calculation unit 457, 461 Component identification quantitative calculator 459 Normalized intensity matrix calculator 460 Quantitative intensity matrix calculator

──────────────────────────────────────────────────続 き Continuation of the front page (56) References JP-A-51-43953 (JP, A) JP-A-60-24447 (JP, A) JP-A-61-83963 (JP, A) 274761 (JP, A) JP-A-6-88814 (JP, A) JP-A-7-103959 (JP, A) JP-A-60-22659 (JP, A) JP-A-4-130270 (JP, A) JP-A-1-191055 (JP, A) JP-A-2-253156 (JP, A) (58) Fields investigated (Int. Cl. ⁷ , DB name) G01N 30/86 G01N 30/74 JICST file (JOIS )

Claims (13)

(57) [Claims] 1. A multi-channel chromatogram analyzing method for analyzing chromatogram data having three-dimensional components of a measurement object detected by a detection device, a wavelength component, and a time component. A compression step of compressing the chromatogram data having the component with respect to the wavelength component and converting the data into two-dimensional chromatogram data; and detecting, based on the compressed two-dimensional chromatogram data, the detection characteristics of the detection device. A deconvolution processing step of performing a deconvolution process for removing a data spread by the device function of (i), and a step of identifying a characteristic component of the measurement object from the two-dimensional chromatogram data from which the data spread is removed. A method for analyzing a multi-channel chromatogram, characterized in that: 2. The characteristic of a measuring object detected by a detecting device.
It has a three-dimensional component of the
Multi-channel for analyzing chromatogram data
In the chromatogram analysis method, the chromatogram data having a three-dimensional component is converted to a wavelength component.
And convert it to two-dimensional chromatogram data.
From the compressed two-dimensional chromatogram data
This is the detection characteristic of the
A decon to remove the spread of data by the normal distribution function
Performs a volume process to calculate the standard deviation of the normal distribution function.
By changing the difference, the two-dimensional chromatogram data
Deconvolution treatment to isolate peaks
And the two-dimensional chromatogram from which the data spread was removed
Each component of the data is expanded using the above device function.
The data before deconvolution processing is added
Of executing the convolution process to restore data
And the specs of the respective components based on the restored data.
And calculating the characteristic information of the object to be measured from the calculated spectrum information.
Identifying and quantifying the component, and at least the identified and quantified characteristic component
Displays the restored data that has been subjected to the
Multichannel chromatography grayed characterized in that it comprises a step of displaying on the stage, the
Ram analysis method. 3. The characteristic of a measurement object detected by a detection device.
It has a three-dimensional component of the
Multi-channel for analyzing chromatogram data
In the chromatogram analysis method, the three-dimensional chromatogram data is used to detect
Detection characteristics, and data expansion due to unique device functions
Deconvolution processing to remove the
Convolution processing steps and deconvoluted chromatogram data
Is compressed with respect to the time components, and the spectrum of each component is
Calculating the information and the characteristics of the measurement object based on the calculated spectrum information.
Multichannel chromatography grayed characterized in that it comprises a step of identifying the gender component, the
Ram analysis method. 4. The characteristic of the object to be measured detected by the detecting means.
It has a three-dimensional component of the
Multi-channel for analyzing chromatogram data
In the chromatogram analyzer, the chromatogram data having the three-dimensional component is
Two-dimensional chromatogram data compressed for wavelength components
From the data compression unit that converts the data into data and the compressed two-dimensional chromatogram data,
This is the detection characteristic of the detection means.
Deconvolution processing to remove the spread of data
And a two-dimensional chromatogram with the data spread removed
Component identification unit that identifies characteristic components of the measurement object from data
And a multi-channel chromatograph comprising:
Ram analysis equipment. 5. The characteristic of the object to be measured detected by the detecting means.
It has a three-dimensional component of the
Multi-channel for analyzing chromatogram data
In the chromatogram analyzer, the start of the time component to be analyzed of the three-dimensional chromatogram data
It has a data setting unit in which a point and an end point are set, and a three-dimensional component set by the data setting unit
Compress chromatogram data with respect to wavelength components,
From the data compression unit that converts it to the original chromatogram data, and from the compressed two-dimensional chromatogram data,
This is the detection characteristic of the
A decon to remove the spread of data by the normal distribution function
Performs a volume process to calculate the standard deviation of the normal distribution function.
By changing the difference, the two-dimensional chromatogram data
Deconvolution section to isolate peaks, and chromatogram data after deconvolution processing
Separation standardization part that separates the data into each component and normalizes it, and data that is separated and standardized
Using the position function, add the data spread for each component data.
To restore the data before deconvolution processing.
Reconvolution that performs convolution processing
And the specifications of each component based on the restored data.
A spectrum calculator for calculating the torque information, and a characteristic component of the object to be measured from the calculated spectrum information.
Component identification and quantification calculation unit for identifying and quantifying the component, and at least the identified and quantified characteristic component and
Displays the restored data that has been subjected to the volume processing.
Multichannel chromatography grayed, characterized in that it comprises a that display unit, the
Ram analysis equipment. 6. The characteristic of the object to be measured detected by the detecting means.
It has a three-dimensional component of the
Multi-channel for analyzing chromatogram data
In chromatogram analysis apparatus, three-dimensional chromatogram time component to be analyzed data and
A data setting unit in which the start point and the end point of the wavelength component are set, and a three-dimensional chromatogram in which the start point and the end point are set
From the data, the detection characteristics of the detection means
Deconvolution which removes the spread of data by the permutation function
The deconvolution section that performs the processing and the deconvoluted chromatogram data
Normalize, compress the time components, and
A normalized spectrum calculation unit for calculating the spectrum information, and a characteristic of the measurement object based on the calculated spectrum information.
Identify gender component, multichannel chromatography grayed, characterized in that it comprises a component identifying quantitative calculation unit for quantifying
Ram analysis equipment. 7. A characteristic of an object to be measured based on an output of a detector.
Obtain data including information on component, wavelength component and time component
To reduce the noise component of the data,
To correct the spread accompanying the flow from the
Arithmetic processing and outputting the result of the arithmetic processing.
Analysis method of multi-channel chromatogram. 8. The multi-channel chromatograph according to claim 7,
In the gram analysis method, the noise component is reduced.
The process includes a smoothing process.
Analysis method of channel chromatogram. 9. The multi-channel chromatography according to claim 8, wherein
In the gram analysis method, the step of performing the arithmetic processing includes:
Multi-channel including numerical analysis processing
Chromatogram analysis method. 10. The multi-channel chroma according to claim 9,
In the analysis method of
Detecting the separated sample with the detector.
To analyze multichannel chromatograms. 11. The multi as claimed in claim 7, 8, 9 or 10.
In the channel chromatogram analysis method,
The output device is configured by arranging multiple diodes.
Analysis method of multi-channel chromatogram. 12. The measurement object based on the output of the detector.
Data containing information on characteristic components, wavelength components, and time components
To reduce the noise component of the data,
Estimate the upstream state from the reduced data and perform arithmetic processing
And outputting the result of the arithmetic processing.
Multi-channel chromatogram analysis method. 13. The characteristic component, the wavelength component and the time of the object to be measured.
Multi-channel for analyzing data including information on inter-components
In a chromatogram analyzer, a noise reducer for reducing the noise component of the data and a spread accompanying the flow from the noise-reduced data are supplemented.
A multi-channel chromatograph comprising: an arithmetic processing unit that performs an operation to correct the error; and an output unit that outputs a result of the arithmetic processing.
Ram analysis equipment.