Peak Tracking Device and Peak Tracking Method

Abstract

The present invention relates to a peak tracking device, comprising: a chromatogram acquisition unit that acquires a plurality of chromatograms based on a plurality of measurement data obtained by giving a plurality of analysis condition data to an analysis device; a fraction calculation unit that calculates fraction data indicating the probability of which substance each peak included in each chromatogram belongs to among the substances included in the sample; and a fraction display unit that displays the fraction data calculated by the fraction calculation unit on a display.

Description

Technical Field

[0001] The present invention relates to a peak tracking device and a peak tracking method. Background Art

[0002] A chromatogram of a sample can be obtained from measurement data acquired from an analysis device. In order to improve the peak resolution or to shorten the analysis time, a method for optimizing the analysis condition data is explored.

[0003] When exploring a method, peak tracking is required to identify peaks derived from the same substance among different chromatograms obtained based on different analysis condition data. For peak tracking, for example, the area value of a peak, the optical spectrum of a peak, or the similarity of the MS spectrum of a peak can be used.

[0004] In addition, by performing a regression analysis between the analysis condition data and the analysis measurement data, useful information for method exploration can be provided. In Non-Patent Document 1 below, a regression analysis of the retention time is performed.

[0005] Non-Patent Document 1: "Ternary isocratic mobile phase optimization utilizing resolution Design space based on retention time and peak width modeling (Ternary isocratic mobile phase optimization using a resolution design space based on retention time and peak width modeling)", Journal of Chromatography A, Vol. 1273, 95P - 104P, January 18, 2013 Summary of the Invention

[0006] Technical Problem to be Solved by the Invention

[0007] As is well known, in the synthesis of pharmaceuticals with relatively large molecular weights such as nucleic acid pharmaceuticals and peptide pharmaceuticals, structural analogs are generated as by-products. Since the optical spectra of such analogs are also similar, peak tracking may not be sufficiently performed even using the similarity of the optical spectra. In addition, the MS spectra of such analogs are often similar in shape, and a huge amount of labor is required to explore the m / z values that output specific chromatograms for each peak. Even by using the method of peak area values, the combination cannot be locked in many cases, and peak tracking may not be sufficiently performed sometimes.

[0008] An object of the present invention is to provide useful information for a user to identify peaks included in a chromatogram.

[0009] Solution for Solving the Above Technical Problem

[0010] A peak tracking device according to an aspect of the present invention includes: a chromatogram acquisition unit that acquires a plurality of chromatograms based on a plurality of measurement data obtained by giving a plurality of analysis condition data to an analysis device; a fraction calculation unit that calculates fraction data indicating the probability of belonging to which substance contained in the sample for each peak included in each chromatogram; and a fraction display unit that displays the fraction data calculated by the fraction calculation unit on a display.

[0011] Advantages of the Invention

[0012] According to the present invention, it is possible to provide useful information for identifying peaks contained in a chromatogram to a user. BRIEF DESCRIPTION OF THE DRAWINGS

[0013] Figure 1 is an overall view of the analysis system of this embodiment.

[0014] Figure 2 is a configuration diagram of the computer of this embodiment.

[0015] Figure 3 is a functional block diagram of the computer of this embodiment.

[0016] Figure 4 is a diagram showing chromatograms obtained from different analysis condition data.

[0017] Figure 5 is a flowchart showing the peak tracking method of this embodiment.

[0018] Figure 6 is a diagram showing the relationship between solvent concentration and retention time.

[0019] Figure 7 is a diagram showing the predicted distribution of the retention time regression curve of substance A.

[0020] Figure 8 is a diagram showing the predicted distribution of the retention time regression curve of substance B.

[0021] Figure 9 is a diagram showing the predicted distribution of the retention time regression curve of substance C.

[0022] Figure 10 is a diagram showing the predicted distribution of the retention time regression curve of substance D.

[0023] Figure 11 is a diagram showing the predicted distribution of the retention time regression curve of substance E.

[0024] Figure 12 is a diagram showing a histogram of the classification distribution probability at a solvent concentration of "0.25".

[0025] Figure 13 It is a graph showing a histogram of the classification distribution probability at a solvent concentration of "0.5".

[0026] Figure 14 It is a graph showing a histogram of the classification distribution probability at a solvent concentration of "0.75".

[0027] Figure 15 It is a graph showing a histogram of the classification distribution probability at a solvent concentration of "1.0".

[0028] Figure 16 It is a graph showing the analysis support screen displayed on the display. DETAILED DESCRIPTION

[0029] Next, the configuration of the peak tracking device, method, and program according to the embodiments of the present invention will be described with reference to the drawings.

[0030] (1) Overall configuration of the analysis system

[0031] Figure 1 It is an overall view of the analysis system 5 of the present embodiment. The analysis system 5 includes a computer 1 and a liquid chromatograph 3. The computer 1 and the liquid chromatograph 3 are connected via a network 4. The network 4 can be, for example, a LAN (Local Area Network).

[0032] The computer 1 has functions such as setting analysis conditions for the liquid chromatograph 3 and acquiring and analyzing the measurement results in the liquid chromatograph 3. A program for controlling the liquid chromatograph 3 is installed in the computer 1.

[0033] The liquid chromatograph 3 includes a pump unit, an autosampler unit, a column oven unit, a detector unit, etc. The liquid chromatograph 3 also includes a system controller. The system controller controls the liquid chromatograph 3 according to the control instructions received from the computer 1 via the network 4. The system controller sends the data of the measurement results of the liquid chromatograph 3 to the computer 1 via the network 4.

[0034] (2) Configuration of the computer (peak tracking device)

[0035] Figure 2This is a configuration diagram of computer 1. In this embodiment, computer 1 is a personal computer. Computer 1 includes: CPU (Central Processing Unit) 101, RAM (Random Access Memory) 102, ROM (Read Only Memory) 103, display 104, operation unit 105, storage device 106, communication interface 107, and device interface 108.

[0036] CPU 101 controls computer 1. RAM 102 is used as a work area when CPU 101 executes programs. Control programs and the like are stored in ROM 103. Display 104 can be, for example, a liquid crystal display. Operation unit 105 is a device that accepts user operations and includes a keyboard, a mouse, etc. Display 104 is composed of a touch panel display, and display 104 can also have the function of operation unit 105. Storage device 106 is a device that stores various programs and data. Storage device 106 can be, for example, a hard disk. Communication interface 107 is an interface for communicating with other computers and devices. Communication interface 107 is connected to network 4. Device interface 108 is an interface for accessing various external devices. CPU 101 can access storage medium 109 via an external device connected to device interface 108.

[0037] In storage device 106, peak tracking program P1, analysis condition data AP, measurement data MD, chromatogram CG, retention time data RTD, and fraction data SD are stored. Peak tracking program P1 is a program for controlling liquid chromatograph 3. Peak tracking program P1 has: a function of setting analysis conditions for liquid chromatograph 3; a function of obtaining measurement results from liquid chromatograph 3 and generating chromatogram CG, etc., and analyzing the measurement results. Analysis condition data AP is data that describes the analysis conditions set in liquid chromatograph 3 and includes multiple analysis parameters. Measurement data MD is data of the measurement results obtained from liquid chromatograph 3. Retention time data RTD is data that shows the retention times of substances included in chromatogram CG. In this embodiment, the sample to be analyzed contains multiple substances, and in chromatogram CG, there are peaks of multiple substances. Thus, in retention time data RTD, there are data showing multiple retention times for each chromatogram CG. Fraction data SD is data calculated based on the belonging probabilities indicating which substances in the sample the peaks included in the chromatogram belong to.

[0038] Figure 3It is a functional block diagram of computer 1. Control unit 200 is a functional unit implemented by CPU 101 using RAM 102 as a work area and executing peak tracking program P1. Control unit 200 includes: analysis management unit 201, chromatogram acquisition unit 202, fraction calculation unit 203, and analysis support information display unit 204.

[0039] Analysis management unit 201 controls liquid chromatograph 3. Analysis management unit 201 receives the setting of analysis condition data AP by the user and the start instruction of the analysis process, and gives an instruction for the analysis process of liquid chromatograph 3. The user sets a combination of setting values of analysis parameters such as solvent concentration, solvent mixing ratio, gradient initial value, gradient slope, column temperature, etc. as analysis conditions. The user sets multiple combinations of the above analysis parameters. For example, a combination of analysis parameters that gradually changes the solvent concentration, or a combination of analysis parameters that gradually changes the column temperature, etc. are set as analysis conditions. The user generates multiple pieces of analysis condition data AP in this way and performs analysis processes based on multiple pieces of analysis condition data AP on the same sample.

[0040] Analysis management unit 201 additionally obtains measurement data MD from liquid chromatograph 3. As described above, the user performs analysis processes based on multiple pieces of analysis condition data AP. Analysis management unit 201 obtains multiple pieces of measurement data MD corresponding to multiple pieces of analysis condition data AP.

[0041] Chromatogram acquisition unit 202 obtains chromatogram CG based on measurement data MD, and chromatogram acquisition unit 202 stores the obtained chromatogram CG in storage device 106. As described above, analysis management unit 201 obtains multiple pieces of measurement data MD corresponding to multiple pieces of analysis condition data AP. Chromatogram acquisition unit 202 obtains multiple chromatograms CG corresponding to multiple pieces of measurement data MD.

[0042] Fraction calculation unit 203 calculates fraction data SD based on multiple chromatograms CG obtained by chromatogram acquisition unit 202. In the present embodiment, the sample analyzed by liquid chromatograph 3 contains multiple substances. Fraction calculation unit 203 calculates the belonging probability of each peak included in chromatogram CG to which substance included in the sample.

[0043] Figure 4 It is a diagram showing two chromatograms CG1 and CG2 obtained based on two different pieces of analysis condition data AP. Chromatograms CG1 and CG2 are obtained based on the measurement data MD obtained for the same sample. Refer to Figure 4 It can be seen that in chromatograms CG1 and CG2, due to the difference in analysis condition data AP, the retention time of each peak is different. The peaks connected by a dotted line in the figure are peaks derived from the same substance. Fraction calculation unit 203 calculates the belonging probability of each peak included in chromatograms CG1 and CG2 to which substance included in the sample.

[0044] Specifically, the fraction calculation unit 203 performs a regression analysis using the analysis parameters included in the analysis condition data AP, namely the solvent concentration, and the retention time data RTD of each peak obtained from the chromatogram CG. In the present embodiment, the fraction calculation unit 203 performs regression on the retention time data RTD using Bayesian inference. Then, the fraction calculation unit 203 calculates the belonging probability of each peak included in the chromatogram CG belonging to which substance included in the sample based on the regression analysis result of the retention time data RTD. Then, the fraction calculation unit 203 generates fraction data SD based on the calculated belonging probability.

[0045] The analysis support information display unit 204 displays information for analysis support on the display 104. The analysis support information display unit 204 includes a distribution display unit 205 and a fraction display unit 206. The distribution display unit 205 displays the result of the regression analysis in the fraction calculation unit 203 on the display 104. The fraction display unit 206 displays the fraction data SD calculated by the fraction calculation unit 203 on the display 104.

[0046] (3) Peak tracking method

[0047] Next, the peak tracking method executed in the computer 1 (peak tracking device) of the present embodiment will be described. Figure 5 is a flowchart showing the peak tracking method of the present embodiment. As is well known, the retention time of each substance in the chromatogram can be regressed against analysis parameters such as solvent concentration, PH, and column temperature. In the peak tracking method of the present embodiment, the belonging of each peak included in the chromatogram is determined simultaneously with the regression of the retention time. When solving the belonging problem simultaneously with the regression, since the solution is not necessarily unique, in the present embodiment, a distribution that can be used as a solution is obtained through calculation of the posterior distribution based on Bayesian inference.

[0048] At the start Figure 5 Before the processing shown, the user operates the operation unit 105 in advance to set multiple analysis conditions. After accepting such a setting operation by the user, the analysis management unit 201 stores multiple analysis condition data AP in the storage device 106.

[0049] Next, in Figure 5 In step S101 shown, the analysis management unit 201 sets multiple analysis condition data AP in the liquid chromatograph 3. Specifically, the analysis management unit 201 sets multiple analysis condition data AP for the system controller of the liquid chromatograph 3. Correspondingly, in the liquid chromatograph 3, multiple analysis processes are performed on the same sample based on the set multiple analysis condition data AP. In the liquid chromatograph 3, multiple measurement data MD are obtained corresponding to the multiple analysis condition data AP.

[0050] Next, in step S102, the analysis management unit 201 obtains a plurality of measurement data MD from the liquid chromatograph 3. The analysis management unit 201 stores the obtained plurality of measurement data MD in the storage device 106.

[0051] Next, in step S103, the chromatogram acquisition unit 202 obtains the plurality of measurement data MD stored in the storage device 106 in step S102, and obtains a plurality of chromatograms CG from the obtained plurality of measurement data MD.

[0052] Next, in step S104, the fraction calculation unit 203 obtains the plurality of chromatograms CG obtained in step S103, and obtains the retention time data RTD of each peak included in each chromatogram CG.

[0053] Figure 6 is a graph showing the relationship between the analysis condition data AP and the retention time data RTD obtained in step S104. In Figure 6 , the horizontal axis shows the solvent concentration among the analysis parameters included in the analysis condition data AP. In Figure 6 , the vertical axis shows the retention time. In Figure 6 , the retention times obtained for five solvent concentrations corresponding to five analysis condition data AP are charted. There are five solvent concentrations: "0.0", "0.25", "0.5", "0.75", "1.0", and here it is particularly assumed that the unit of the solvent concentration has no meaning.

[0054] For the five solvent concentrations, five retention times are plotted respectively. The above five retention times are the retention times of five peaks of substances A - E contained in the sample. In Figure 6 , the points of the five retention times are shown by five symbols. Specifically, starting from the order of the shortest retention time, they are shown by five symbols: quadrilateral, star, triangle, circle, and cross in sequence. However, in reality, due to the difference in solvent concentration, the order of the retention times may sometimes change, so Figure 6 the symbols in are just for convenience. That is, the same symbol does not necessarily represent the retention time of a peak from the same substance. Here, the five retention times at the concentration of "0.0" are corresponded to substances A - E in the order of the shortest retention time. That is, it is assumed that the quadrilateral at the concentration of "0.0" is substance A, the star is substance B, the triangle is substance C, the circle is substance D, and the cross is substance E. The main object of the present invention is to show the belonging probability of the peaks corresponding to the retention times represented by the five symbols at the concentrations of "0.25", "0.5", "0.75", and "1.0" to which substances A - E belong.

[0055] Next, in step S105, the score calculation unit 203 obtains the belonging probability (posterior distribution) indicating which substance each peak belongs to according to the following mathematical formula (1).

[0056] [Equation 1]

[0057] Π c Π i p(i|c)N(x ci |f(c), σ)…(1)

[0058] In the mathematical formula (1), c is the concentration. In the Figure 6 illustrated example, there are five kinds of concentrations c: "0.0", "0.25", "0.5", "0.75", "1.0". In the mathematical formula (1), i is the ID of the peak. In the Figure 6 illustrated example, starting from the order of increasing retention time, the IDs i = 0 - 4 are assigned in sequence. That is, in the Figure 6 , the ID i = 0 is assigned to the peak shown as a quadrilateral, the ID i = 1 is assigned to the peak shown as a star, the ID i = 2 is assigned to the peak shown as a triangle, the ID i = 3 is assigned to the peak shown as a circle, and the ID i = 4 is assigned to the peak shown as a cross.

[0059] p(i|c) indicates the probability that a peak is selected according to the classification distribution. p(i|c) is the probability that the peak with i = 0 - 4 is selected at the concentration c. That is, p(i|c) becomes the prior distribution, which is used to calculate the posterior distribution indicating which substance among substances A - E each peak belongs to. In the present embodiment, for simplicity of explanation, it is assumed that p(i|c) is constant and independent of i. However, by referring to the spectrum or area of the peak, an appropriate value can be set as the probability that the peak is selected at the concentration c. For example, by referring to the size relationship or spectrum similarity of the peaks of substances A - E, an appropriate value can be set for p(i|c).

[0060] In the mathematical formula (1), x is the retention time at the concentration c and ID = i. In the mathematical formula (1), f(c) is the regression function of the retention time with respect to the concentration c, and N(u, σ) indicates the probability of x under the conditions of the mean u and the standard deviation σ. That is, in the mathematical formula (1), the function N indicates the likelihood of x.

[0061] The fraction calculation unit 203 gives appropriate prior distributions to the assumed noise amount (standard deviation σ) or the regression coefficients inside the regression function f(c), and performs Bayesian estimation. The noise amount (standard deviation σ) is appropriately given a rough value according to the precision of the repeated processing of the liquid chromatograph 3 or the like. In addition, in the case where the regression function f(c) is the logarithm of a quadratic polynomial, it is well known that since the quadratic coefficient is very small in most substances, an appropriate normal distribution can be set empirically. Alternatively, an appropriate value can be set based on past cases where peak tracking has been completed and using information criteria such as WAIC. Regarding the first-order coefficient, the general tendency with respect to the change in the solvent concentration is known. In the case of Figure 6 , for all substances A - E, the retention time also generally shows a tendency to decrease to the right with respect to the solvent concentration. Therefore, more preferably, a flat prior distribution such as a t-distribution can be given empirically. As described above, the posterior distribution of each peak belonging to substances A - E can be obtained from the likelihood and the prior distribution. In addition, in the present embodiment, Bayesian inference using NUTS sampling is employed. Thus, the likelihood is calculated by marginalization for i(ID). Employing NUTS sampling is an example, and other methods of Bayesian theory can also be employed.

[0062] Figures 7 - 11 is a graph showing the predictive distribution (distribution of belonging probabilities) calculated by Bayesian inference for the retention time shown in Figure 6 . Figure 7 is the predictive distribution for substance A. That is, it is a graph showing the predictive distribution of substance A based on Bayesian inference indicated by the quadrilateral at the concentration "0.0". Figure 8 is the predictive distribution for substance B. That is, it is a graph showing the predictive distribution of substance B based on Bayesian inference indicated by the star at the concentration "0.0". Figure 9 is the predictive distribution for substance C. That is, it is a graph showing the predictive distribution of substance C based on Bayesian inference indicated by the triangle at the concentration "0.0". Figure 10 is the predictive distribution for substance D. That is, it is a graph showing the predictive distribution of substance D based on Bayesian inference indicated by the circle at the concentration "0.0". Figure 11 is the predictive distribution for substance E. That is, it is a graph showing the predictive distribution of substance E based on Bayesian inference indicated by the cross at the concentration "0.0".

[0063] In Figures 7 - 11 , the probabilities of each peak belonging to substances A - E are represented by shading. The darker the color, the higher the belonging probability; the lighter the color, the lower the belonging probability. As shown in Figure 11 , it can be seen that for the peak indicated by the cross, the probability of belonging to substance E is high at any concentration. In addition, as shown in Figure 8 andFigure 9 As shown, it can be seen that in solvents with a concentration above "0.25", the peak shown as a star not only has a probability of belonging to Substance B, but also has a certain probability of belonging to Substance C to some extent. In addition, as Figure 8 and Figure 9 shown, it can be seen that in solvents with a concentration above "0.25", the peak shown as a triangle not only has a probability of belonging to Substance C, but also has a certain probability of belonging to Substance B to some extent.

[0064] Since the classification distribution is peripherized during sampling, it is actually necessary to calculate the classification distribution separately. The score calculation unit 203 calculates the likelihood L(c) shown in the following mathematical formula (2) using sampling of the posterior distribution.

[0065] [Equation 2]

[0066] L(c) = Π i p(i|c)N(x ci |f(c), σ)…(2)

[0067] And, in step S106, the score calculation unit 203 calculates score data SD indicating which of Substances A - E each peak belongs to using the following mathematical formula (3). The score calculation unit 203 stores the score data SD in the storage device 106.

[0068] [Equation 3]

[0069]

[0070] Figures 12 - 15 is a graph showing a histogram of the classification distribution probability when performing Bayesian inference on Substance C. The score calculation unit 203 calculates the likelihood L(c) shown above using sampling of the posterior distribution and generates a histogram of the classification distribution probability.

[0071] Figure 12 is a graph showing a histogram of the classification distribution probability regarding Substance C at a concentration of "0.25". In Figure 12 , the horizontal axis is the probability of the classification distribution, and the vertical axis is the histogram (cumulative value) of the classification distribution probability. That is, it is a histogram obtained by calculating the classification distribution probability for a plurality of analysis condition data AP and statistically processing the results. As shown in the figure, at a probability of 1.0, the cumulative value of i = 2 reaches the maximum. In addition, at a probability of 1.0, the cumulative value of i = 3 follows closely. From this, it can be known that at a concentration of "0.25", the probability that the peak of i = 2 belongs to Substance C is very high. In addition, it can be known that at a concentration of "0.25", the peak of i = 3 also has a certain probability of belonging to Substance C to some extent.

[0072] In Figure 12Among them, the ratio shown in the upper part of the graph represents the fractional data SD of the belonging probability of each peak. The fractional data SD is a value showing the occupancy rate of the frequency with a belonging probability of 0.8 or more, and shows the fractional data SD when i = 0 - 4 starting from the right end. In the example in the figure, the fractional data SD of the peak with i = 0 is 0, the fractional data SD of the peak with i = 1 is 0.02, the fractional data SD of the peak with i = 2 is 0.69, the fractional data SD of the peak with i = 3 is 0.3, and the fractional data SD of the peak with i = 4 is 0.

[0073] Figure 13 It is a graph showing a histogram of the classification distribution probability of substance C at a concentration of "0.5". As shown in the figure, at a probability of 1.0, the cumulative value of i = 2 reaches the maximum. In addition, at a probability of 1.0, the cumulative value of i = 3 follows closely. From this, it can be known that at a concentration of "0.5", the probability that the peak of i = 2 belongs to substance C is also very high. In addition, it can be known that at a concentration of "0.5", the peak of i = 3 also has a certain probability of belonging to substance C to some extent. Moreover, it can be known that at a concentration of "0.5", the peak of i = 1 also has a very small probability of belonging to substance C.

[0074] In Figure 13 Among them, the ratio shown in the upper part of the graph is the same as Figure 12 and is the fractional data SD of the belonging probability of each peak, which is a value showing the occupancy rate of the frequency with a probability of 0.8 or more. In the example in the figure, the fractional data SD of the peak with i = 0 is 0.05, the fractional data SD of the peak with i = 1 is 0.09, the fractional data SD of the peak with i = 2 is 0.61, the fractional data SD of the peak with i = 3 is 0.24, and the fractional data SD of the peak with i = 4 is 0.

[0075] Figure 14 It is a graph showing a histogram of the classification distribution probability of substance C at a concentration of "0.75". Figure 15 It is a graph showing a histogram of the classification distribution probability of substance C at a concentration of "1.0". It can be known that at concentrations of "1.0" and "0.75", the probability that the peak of i = 2 belongs to substance C is also very high.

[0076] In Figure 14 the example, the fractional data SD of the peak with i = 0 is 0.14, the fractional data SD of the peak with i = 1 is 0.12, the fractional data SD of the peak with i = 2 is 0.59, the fractional data SD of the peak with i = 3 is 0.16, and the fractional data SD of the peak with i = 4 is 0. In Figure 15 the example, the fractional data SD of the peak with i = 0 is 0.14, the fractional data SD of the peak with i = 1 is 0.19, the fractional data SD of the peak with i = 2 is 0.51, the fractional data SD of the peak with i = 3 is 0.15, and the fractional data SD of the peak with i = 4 is 0. Additionally, inFigures 11 - 14 In the example shown, the occupancy rate of the frequency with a probability of the categorical distribution of 0.8 or more is used as the score data SD, but this is just one example. The score data SD may be any data generated based on the belonging probability indicating which substance each peak belongs to. In other words, the score data SD may be any data related to the belonging probability indicating which substance each peak belongs to.

[0077] Next, in step S107, the score display unit 206 displays the score data SD on the display 104. Then, in step S108, the distribution display unit 205 displays the distribution (predicted distribution) of the belonging probability calculated in step S105 on the display 104. Figure 16 FIG. is a diagram showing the analysis support screen displayed on the display 104. In Figure 16 the example shown, the score data SD of the belonging probability of substance C and the distribution of the belonging probability of substance C are displayed on the analysis support screen.

[0078] In Figure 16 , the "belonging score of substance C" is displayed based on the score data SD. The belonging score of substance C is the same as the score displayed at the upper part of the histogram shown in Figures 12 - 15 . In addition, the distribution of the belonging probability of substance C is the same as that shown in Figure 9 . In Figure 16 , an example of displaying the analysis support information regarding substance C is illustrated, and the same applies to substances A, B, D, and E.

[0079] In this way, the computer 1 of the present embodiment displays on the display 104 the score data SD based on the belonging probability indicating which substance each peak included in each chromatogram belongs to in the sample. Thus, the user can confirm the belonging probability of the peak. In addition, the computer 1 of the present embodiment displays the distribution of the belonging probability obtained by regression analysis on the display 104. Thus, the user can visually confirm the appropriateness of the belonging of the peak.

[0080] (4) Corresponding relationship between each component of the claims and each element of the embodiment

[0081] Hereinafter, examples of the corresponding relationship between each component of the claims and each element of the embodiment will be described, but the present invention is not limited to the following examples. In the above embodiment, the liquid chromatograph 3 is an example of the analysis device. In addition, in the above embodiment, the computer 1 is an example of the peak tracking device.

[0082] As each component of the claims, various elements having the configurations or functions described in the claims can also be used.

[0083] (5) Other embodiments

[0084] In the above-described embodiment, the computer 1 causes the score data SD calculated by the score calculation unit 203 to be displayed on the display 104. As another embodiment, the computer 1 may be configured to output the score data SD calculated by the score calculation unit 203 to another device, another program, process, etc. For example, it may be configured to output the score data SD to a device that performs processing for the purpose of AQBD (Analytical Quality by Design). In this case, the control unit 200 of the computer 1 has an output unit in addition to the functional block diagram shown in Figure 2 shown.

[0085] For example, there is a program or device that obtains a design space for retention time or resolution by performing regression analysis on the retention time or resolution of peaks. It may be configured to output the score data SD calculated in the present embodiment to the program or device that processes the design space. For example, in a device that inputs the score data SD, it is possible to associate the design space with the score data SD and present information.

[0086] In the above-described embodiment, the liquid chromatograph 3 is taken as an example of the analysis device of the present invention. The present invention can also be applied to a gas chromatograph. Further, in the above-described embodiment, the peak tracking device, i.e., the computer 1, is connected to the analysis device, i.e., the liquid chromatograph 3, via the network 4, and this case is taken as an example for description. As another embodiment, the computer 1 may also be a configuration built into the analysis device.

[0087] In the above-described embodiment, the case where the peak tracking program P1 is stored in the storage device 106 is taken as an example for description. As another embodiment, the peak tracking program P1 may also be stored in the storage medium 109 and provided. The CPU 101 may also access the storage medium 109 via the device interface 108 and store the peak tracking program P1 stored in the storage medium 109 in the storage device 106 or the ROM 103. Or the CPU 101 may also access the storage medium 109 via the device interface 108 and execute the peak tracking program P1 stored in the storage medium 109.

[0088] In addition, the specific configuration of the present invention is not limited to the above-described embodiment, and various changes and modifications can be made without departing from the gist of the invention.

[0089] (6) Solution

[0090] Those skilled in the art can understand that the above-described multiple exemplary embodiments are specific examples of the following solutions.

[0091] (Item 1)

[0092] One aspect of the peak tracking device of the present invention includes:

[0093] A chromatogram acquisition unit that acquires a plurality of chromatograms based on a plurality of measurement data obtained by giving a plurality of analysis condition data to an analysis device;

[0094] A score calculation unit that calculates score data based on the belonging probability indicating which substance each peak included in each chromatogram belongs to in the sample;

[0095] A score display unit that displays the score data calculated by the score calculation unit on a display.

[0096] It is possible to provide useful information for identifying the peaks included in the chromatogram to the user.

[0097] (Item 2)

[0098] Alternatively, in the peak tracking device according to Item 1,

[0099] The score calculation unit calculates the belonging probability by performing a regression analysis on the plurality of analysis condition data and the retention time of each substance obtained based on each analysis condition data.

[0100] Using the regressable retention time, it is possible to provide useful information for identifying the peaks included in the chromatogram to the user.

[0101] (Item 3)

[0102] Alternatively, in the peak tracking device according to Item 2,

[0103] The peak tracking device further includes a distribution display unit that displays the distribution of the belonging probability obtained by the regression analysis.

[0104] By referring to the distribution of the belonging probability, the user can confirm the appropriateness of the peak identification.

[0105] (Item 4)

[0106] Alternatively, in the peak tracking device according to Item 2 or Item 3,

[0107] The score calculation unit calculates the belonging probability based on the likelihood obtained from the error distribution when using Bayesian inference to assign an error to the regression function and the probability of each peak being selected.

[0108] According to Bayesian inference, it is possible to present the belonging probability indicating which substance the peak belongs to.

[0109] (Item 5)

[0110] Another aspect of the peak tracking device of the present invention includes:

[0111] A chromatogram acquisition unit that acquires a plurality of chromatograms based on a plurality of measurement data obtained by giving a plurality of analysis condition data to an analysis device;

[0112] A score calculation unit that calculates score data based on the belonging probabilities indicating which substances contained in the sample each peak contained in each chromatogram belongs to;

[0113] An output unit that outputs the score data calculated by the score calculation unit.

[0114] The score data can be used for other devices or programs.

[0115] (Item 6)

[0116] The peak tracking method of other aspects of the present invention includes the following steps:

[0117] Based on a plurality of measurement data obtained by giving a plurality of analysis condition data to an analysis device, acquire a plurality of chromatograms;

[0118] Calculate score data based on the belonging probabilities indicating which substances contained in the sample each peak contained in each chromatogram belongs to;

[0119] Display the calculated score data on a display.

[0120] (Item 7)

[0121] The peak tracking program of other aspects of the present invention causes a computer to execute the following processing:

[0122] Based on a plurality of measurement data obtained by giving a plurality of analysis condition data to an analysis device, acquire a plurality of chromatograms;

[0123] Calculate score data based on the belonging probabilities indicating which substances contained in the sample each peak contained in each chromatogram belongs to;

[0124] Display the calculated score data on a display.

Claims

1. A peak tracking device, characterized in that, Comprising: A chromatogram acquisition unit that acquires a plurality of chromatograms based on a plurality of measurement data obtained by giving a plurality of analysis condition data to an analysis device; A score calculation unit that calculates score data based on belonging probabilities indicating the probability that each peak included in each chromatogram belongs to a certain substance contained in a sample; A score display unit or an output unit, where the score display unit compares the score data calculated by the score calculation unit with the plurality of analysis condition data and displays them on a display, and the output unit compares the score data calculated by the score calculation unit with the plurality of analysis condition data and outputs them; The plurality of analysis condition data includes solvent concentration.

2. The peak tracking device according to claim 1, characterized in that: The score calculation unit calculates the belonging probability by performing a regression analysis on the plurality of analysis condition data and the retention times of each substance obtained based on each analysis condition data.

3. The peak tracking device according to claim 2, characterized in that: The peak tracking device further comprises a distribution display unit that displays the distribution of the belonging probabilities obtained by the regression analysis.

4. The peak tracking device according to claim 1, characterized in that: The score calculation unit calculates the belonging probability based on the likelihood obtained from the error distribution when using Bayesian inference to assign an error to the regression function and the probability of each peak being selected.

5. The peak tracking device according to claim 1, characterized in that: The score calculation unit calculates score data for each of the multiple substances contained in the sample.

6. A peak tracking method, characterized in that, Including the following steps: Based on a plurality of measurement data obtained by giving a plurality of analysis condition data to an analysis device, acquiring a plurality of chromatograms; Calculating score data based on belonging probabilities indicating the probability that each peak included in each chromatogram belongs to a certain substance contained in a sample; Comparing the calculated score data with the plurality of analysis condition data and displaying them on a display, or comparing the calculated score data with the plurality of analysis condition data and outputting them; The plurality of analysis condition data includes solvent concentration.

7. The peak tracking method according to claim 6, characterized in that: In the step of calculating the score data, score data is calculated for each of the multiple substances contained in the sample.

Images