Data processing system

Abstract

The data processing system of the present application can effectively exert the component number automatic estimation function, comprising: a source data holding section that holds source data of a three-dimensional chromatogram including chromatographic data obtained by chromatographic analysis and a frequency spectrum; an arithmetic processing section configured to execute a peak estimation process that repeatedly performs a component estimation step of estimating a three-dimensional chromatogram of one peak component included in one peak waveform portion of the source data held in the source data holding section until three-dimensional chromatograms of all estimated peak components estimated in the component estimation step are synthesized to obtain synthetic data that approximates the source data, thereby estimating the peak included in the peak waveform portion; and a maximum number storage section that stores a maximum number of estimated peak components, the arithmetic processing section being configured to end the peak estimation process regardless of the approximation condition of the synthetic data with respect to the source data when the number of estimated peak components reaches the maximum number.

Description

Data processing system Technical Field

[0001] This invention relates to a data processing system for processing three-dimensional chromatographic data. Background Technology

[0002] In liquid chromatography (LC) using multi-channel detectors such as photodiode array (PDA) detectors, three-dimensional chromatographic data with three dimensions—time, wavelength, and signal intensity (absorbance)—can be obtained by continuously acquiring the absorption spectra of the sample dissolved from the analytical column.

[0003] When using liquid chromatography to quantify a target component in a sample, the wavelength with the highest absorbance of the target component is typically used to prepare the chromatogram. The area of ​​the target component's peak on the chromatogram is then determined for quantification. However, the sample may sometimes contain impurities other than the target component. The peaks of these impurities may overlap with the peak of the target component, forming a single peak waveform. In such cases, the peak area of ​​either the target component or the impurity cannot be determined solely from this single overlapping peak waveform. Therefore, it is necessary to deduce the specific peak overlap patterns that formed the waveform.

[0004] As an algorithm for automatically inferring multiple peaks contained in a peak waveform portion, there is a known algorithm that applies a peak model function, such as an Exponential Modified Gaussian (EMG) function, to the chromatographic waveform while adjusting the parameters of the peak model function (see Patent Document 1). The algorithm disclosed in Patent Document 1 has an automatic component number inference function as follows: for each peak contained in the peak waveform portion, it infers its own three-dimensional chromatogram. If the loss of the inference result (a value representing the degree of approximation of the three-dimensional chromatogram of the synthesized data obtained by synthesizing the chromatogram and spectrum of the inferred peak components relative to the source data; the smaller this value, the more approximate it is to the source data) is above a predetermined value, then the number of peaks is added by one. By repeating this process, the number of peak components contained in the peak waveform portion is automatically inferred.

[0005] [Existing Technical Documents]

[0006] [Patent Literature]

[0007] Patent Document 1: International Publication No. 2016 / 035167 Summary of the Invention

[0008] [The problem the invention aims to solve]

[0009] Analysis using algorithms with automatic component estimation capabilities can be performed on any range of analysis objects (wavelength range and hold time range). However, there is a situation where the number of estimated peak components changes even with slight variations in the analysis object range. This phenomenon is mainly due to the magnitude of noise components contained in the data within the analysis object range, which is difficult to eliminate through algorithmic modifications.

[0010] Moreover, if the algorithm described above is used, in addition to the peaks of the main components and secondary components contained in the sample, the peaks of impurities with concentrations much lower than those of the main components / secondary components can also be quantified. However, it cannot handle situations where impurities are to be ignored and only the main components / secondary components are to be quantified.

[0011] The present invention was made in view of the aforementioned problems, and its purpose is to enable the automatic component prediction function to function effectively while preventing the prediction of unnecessary peak components.

[0012] [Technical means to solve the problem]

[0013] The data processing system of the present invention includes: a source data holding unit that holds source data comprising chromatographic data and a spectrum of three-dimensional chromatography obtained by chromatographic analysis; a computational processing unit configured to perform peak estimation processing, wherein the peak estimation processing involves repeatedly performing a component estimation step to estimate a peak component contained in a peak waveform portion of the source data held in the source data holding unit, until synthetic data obtained by synthesizing all the three-dimensional chromatograms of the three-dimensional chromatograms in which the peak components have been estimated in the component estimation step approximates the source data, thereby estimating the peak contained in the peak waveform portion; and a maximum number storage unit that stores a maximum number of the estimated peak components, wherein the computational processing unit is configured to terminate the peak estimation processing regardless of the approximation of the synthetic data relative to the source data when the number of the estimated peak components reaches the maximum number.

[0014] That is, the data processing system of the present invention is a system that performs peak prediction processing to predict the number of peak components contained in a peak waveform portion and the three-dimensional chromatogram of each peak component. In the peak prediction processing, in principle, the component prediction step of predicting the three-dimensional chromatogram of a peak component contained in the peak waveform portion is repeatedly executed until the synthetic data obtained by synthesizing the three-dimensional chromatograms of all the predicted peak components in the component prediction step approximates the source data. On the other hand, a maximum number of predicted peak components is set. When the number of predicted peak components reaches the set maximum number, the peak prediction processing ends regardless of the approximation state of the synthetic data of the predicted peak components relative to the source data.

[0015] For example, when analyzing a data range (peak waveform portion) where three peak components are predicted using an existing algorithm with automatic component number prediction, if the maximum number of predicted peak components is set to 2, the peak prediction process will terminate without executing the subsequent component prediction steps once the number of predicted peak components reaches two. In this case, the synthetic data from the three-dimensional chromatography of the two predicted peak components may contain a loss of more than one peak component compared to the source data of the same range, but this loss will be ignored. This is particularly effective when it is known in advance that the sample contains a main component, secondary components, and impurities, but the presence of impurities is ignored, and only the main and secondary components are quantified. Conversely, when analyzing a peak waveform portion where two peak components are predicted using an existing algorithm, even if the maximum number of predicted peak components is set to 3, the presence of three peak components will not be predicted. Instead, the automatic component number prediction function will effectively perform its function and predict the presence of two peak components, just like the existing algorithm. This is clearly different from the case where the prediction is forced to include a user-specified number of peak components in a specified peak waveform portion.

[0016] Here, the so-called peak waveform refers to the portion where one or more peaks combine to form a peak waveform. Furthermore, the term "synthetic data approximates source data" means that the difference between the synthetic data obtained through methods such as least squares and the source data meets a specified condition, thus allowing the synthetic data to be evaluated as approximating the source data. An example of this specified condition is that the difference between the synthetic data and the source data is below a specified threshold.

[0017] [The effects of the invention]

[0018] As described above, the data processing system of the present invention has an automatic component estimation function that automatically estimates the number of peak components contained in a specified peak waveform portion by repeatedly performing the component estimation step. It is configured such that when the number of times the component estimation step is executed reaches a set maximum number, the peak estimation process ends regardless of the approximate state of the synthesized data of the estimated peak components relative to the source data. Therefore, the automatic component estimation function can be effectively implemented, and the existence of unnecessary peak components can be estimated from the data of the analysis object range. Attached Figure Description

[0019] Figure 1 is a block diagram schematically illustrating one embodiment of a data processing system.

[0020] Figure 2 is a flowchart illustrating a series of processes related to peak prediction.

[0021] Figure 3 is a flowchart illustrating an example of the actions in the peak prediction processing of this embodiment.

[0022] Figures 4A and 4B are used to compare the prediction results obtained through peak prediction processing. Figure 4A shows the case where the maximum number of predicted peak components is not set (comparative example), and Figure 4B shows the case where the maximum number of predicted peak components is set (executive).

[0023] [Explanation of Symbols]

[0024] 1: Data Processing System

[0025] 2: Source Data Storage Department

[0026] 4: Computation and Processing Unit

[0027] 6: Maximum number storage unit

[0028] 100: Analytical apparatus Detailed Implementation

[0029] Hereinafter, an embodiment of the data processing system of the present invention will be described with reference to the accompanying drawings.

[0030] Figure 1 shows an embodiment of the data processing system.

[0031] The data processing system 1 includes a source data storage unit 2, a computational processing unit 4, and a maximum data storage unit 6. Analytical data acquired using the analytical apparatus 100 is imported into the data processing system 1. The analytical apparatus 100 is configured to perform liquid chromatography analysis on a sample to acquire absorbance spectra at fixed time intervals. That is, the analytical data imported from the analytical apparatus 100 into the data processing system 1 is three-dimensional chromatographic data containing both chromatogram and spectrum.

[0032] The source data storage unit 2 is a storage area that pre-stores three-dimensional chromatographic data (hereinafter referred to as source data) imported from the analysis device 100. The source data storage unit 2 can be implemented using non-volatile flash memory or hard disk drive, etc.

[0033] The processing unit 4 is configured to perform analytical processing of the source data of the three-dimensional chromatography stored in the source data storage unit 2. In addition to quantitative processing to determine the concentration of components contained in the sample based on the area values ​​of the peaks on the chromatogram, the processing unit 4 also performs peak estimation processing to estimate the number of peak components contained in the waveform portion of the peaks within the specified analytical range and the three-dimensional chromatogram of each peak component. The processing unit 4 achieves these functions by executing programs in a computer circuit including a central processing unit (CPU).

[0034] The maximum number storage unit 6 is a storage area that stores a set value for the maximum number of peak components (presumed peak components) predicted to be included in a specified peak waveform portion during peak prediction processing. The maximum number of predicted peak components can be arbitrarily set by the user.

[0035] The series of processes related to peak prediction are illustrated in Figure 2 along with Figure 1.

[0036] First, when the user specifies source data as the analysis target, the processing unit 4 reads the specified source data (step 101). The processing unit 4 displays the read three-dimensional chromatogram source data on a display (illustration omitted) communicatively connected to the data processing system, allowing the user to specify the analysis target range (the retention time range and wavelength range of the analysis target) (step 102), and then allows the user to set the maximum number of predicted peak components (step 103). The maximum number of predicted peak components set by the user is stored in the maximum number storage unit 6. The setting of the maximum number of predicted peak components can also be performed before specifying the analysis target range. Subsequently, when the user inputs an execution instruction for peak prediction processing, the processing unit 4 performs peak prediction processing (step 104).

[0037] For an example of the actions in the peak inference process, the flowchart in Figure 3 is used for illustration.

[0038] At the start of the peak estimation process, the number N of estimated peak components is 0 (step 201). When the peak estimation process begins, the computational processing unit 4 uses a pre-prepared peak model function to execute component estimation steps 202-204, which determine the position and size of a peak estimated to be included in the waveform portion of the chromatogram within the analyte range. In component estimation steps 202-204, firstly, while adjusting parameters such as the height or width of the peak model function, the peak model function is applied to the target peak waveform portion (step 202). The computational processing unit 4 estimates the position and size of the peak model function applied to the peak waveform portion as a peak contained in the peak waveform portion, and calculates to estimate the three-dimensional chromatogram containing the peak component and its spectrum (step 203). As a result, the number N of estimated peak components (estimated peak components) in the three-dimensional chromatogram increases by 1 (step 204).

[0039] After adding a predicted peak component through the component prediction steps 202-204, the processing unit 4 determines whether the total number N of predicted peak components has reached a preset maximum number (step 205). If the number N of predicted peak components has not yet reached the preset maximum number (step 205: No), the three-dimensional chromatograms of all predicted peak components are synthesized to produce synthetic data (step 206). The processing unit 4 uses the least squares method or the like to calculate the loss of the synthesized data relative to the source data (step 207), and determines whether the calculated loss is below a predetermined value (step 208). If the loss is below a predetermined value (step 208: Yes), it is determined that the synthesized data is similar to the source data and the peak prediction process ends.

[0040] On the other hand, if the loss of the synthesized data relative to the source data exceeds a predetermined value (step 208: No), the processing unit 4 executes the component estimation steps 202-204 again, and then adds a previously estimated peak component. Subsequently, the processing unit 4 determines whether the number of previously estimated peak components N has reached the set maximum number (step 205). If the number of previously estimated peak components N has reached the set maximum number (step 205: Yes), then steps 206 and 207 are not executed and the peak estimation process ends.

[0041] Figure 4A shows the prediction result when peak prediction processing was performed without setting the maximum number of predicted peak components (comparative example), and Figure 4B shows the prediction result when peak prediction processing was performed with the maximum number of predicted peak components set.

[0042] If peak estimation processing is performed on a peak waveform portion within a certain data range without setting a maximum number of predicted peak components, the component estimation steps are repeated until the loss of the synthesized data of the predicted peak components relative to the source data becomes below a specified value, as shown in Figure 4A. This results in the predicted peak waveform portion containing three peaks: principal component A, secondary component B, and impurity C. If the maximum number of predicted peak components is set to 2 for the peak waveform portion within the same data range, the number of predicted peak components will reach the set maximum of 2 before the loss of the synthesized data of the predicted peak components relative to the source data becomes below a specified value. Only the peaks of principal component A and secondary component B are predicted within the peak waveform portion. That is, the presence of the peak of impurity C is not predicted and is ignored. Furthermore, if the data range of the analysis object is slightly changed and peak inference processing is performed, the number of peaks inferred may be 2 or 3 if the maximum number of inferred peak components is not set. However, if the maximum number of inferred peak components is set to 2, the number of inferred peaks will remain unchanged at 2.

[0043] The embodiments described above are merely examples illustrating implementations of the data processing system of the present invention. Implementations of the data processing system of the present invention are as follows.

[0044] In one embodiment of the data processing system of the present invention, the data processing system includes: a source data holding unit that holds source data including chromatographic data and a spectrum of three-dimensional chromatography obtained by chromatographic analysis; a computational processing unit configured to perform peak estimation processing, wherein the peak estimation processing involves repeatedly performing a component estimation step to estimate a peak component contained in a peak waveform portion of the source data held in the source data holding unit, until synthetic data obtained by synthesizing all the three-dimensional chromatograms of the three-dimensional chromatograms in which the peak components have been estimated in the component estimation step approximates the source data, thereby estimating the peak contained in the peak waveform portion; and a maximum number storage unit that stores a maximum number of the estimated peak components, wherein the computational processing unit is configured to terminate the peak estimation processing regardless of the approximation of the synthetic data relative to the source data when the number of the estimated peak components reaches the maximum number.

[0045] In the first embodiment, the computational processing unit is configured to, before the number of predicted peak components reaches the maximum number, evaluate the loss of the synthetic data of the three-dimensional chromatogram of all predicted peak components relative to the source data whenever the component prediction step is performed, terminate the peak prediction process when the loss meets a predetermined condition, and terminate the peak prediction process regardless of the loss when the number of predicted peak components reaches the maximum number.

[0046] In the second embodiment of the first embodiment, the user is able to arbitrarily set the maximum number.

[0047] In the third embodiment, the peak waveform portion is included within the analysis object range specified by the user. According to this embodiment, the user can arbitrarily select the peak waveform portion for which they wish to perform peak estimation processing.

[0048] In the third form, the peak waveform portion can also be a portion of multiple peaks overlapping to form a single peak waveform. Therefore, the user can specify the portion that appears to be the shape of multiple overlapping peaks as the analysis object range, and perform peak inference processing on this peak waveform portion.

[0049] In the fourth embodiment, the computational processing unit is configured to, during the component estimation step, apply a pre-prepared peak model function to the peak waveform portion of the chromatogram while adjusting its parameters, thereby estimating the position and size of a peak contained in the peak waveform portion of the chromatogram.

Claims

1. A data processing system, characterized in that, include: The source data retention section retains source data including chromatographic data and spectra obtained through chromatographic analysis, as well as three-dimensional chromatographic data. The processing unit is configured to perform peak estimation processing, which involves repeatedly performing a component estimation step to estimate a peak component contained in a three-dimensional chromatogram of a peak waveform portion of the source data held in the source data holding unit, until the synthesized data obtained by synthesizing the three-dimensional chromatograms of all the estimated peak components in the component estimation step approximates the source data, thereby estimating the peak contained in the peak waveform portion. The system includes a maximum number storage unit that stores the maximum number of the predicted peak components. The arithmetic processing unit is configured to terminate the peak prediction processing when the number of the predicted peak components reaches the maximum number, regardless of how the synthesized data approximates the source data.

2. The data processing system according to claim 1, wherein the computational processing unit is configured to, before the number of predicted peak components reaches the maximum number, evaluate the loss of the synthetic data of the three-dimensional chromatograms of all predicted peak components relative to the source data whenever the component prediction step is performed, terminate the peak prediction processing when the loss meets a predetermined condition, and terminate the peak prediction processing regardless of the loss when the number of predicted peak components reaches the maximum number.

3. The data processing system according to claim 1, wherein the data processing system is configured such that the user can arbitrarily set the maximum number.

4. The data processing system according to claim 1, wherein the peak waveform portion is included within the scope of the analysis object specified by the user.

5. The data processing system according to claim 4, wherein the peak waveform portion is a portion in which multiple peaks overlap to form a peak waveform.

6. The data processing system according to any one of claims 1 to 5, wherein the computational processing unit is configured to, in the component estimation step, adjust the parameters of the peak model function in such a way that a pre-prepared peak model function is applicable to the peak waveform portion of the chromatogram, and based on the adjusted parameters of the peak model function, estimate the position and size of a peak contained in the peak waveform portion of the chromatogram.

Images