Data processing method, information processing device, and program
The method estimates 'in silico' pooled QC values to correct measurement errors in analytical devices, addressing inefficiencies of physical pooled QC methods by enabling immediate analysis of samples and reducing user time and sample loss.
Patent Information
- Application Number
- JP2024026185
- Authority / Receiving Office
- JP · JP
- Patent Type
- Applications
- Current Assignee / Owner
- Filing Date
- 2024-02-26
- Publication Date
- 2025-09-05
AI Technical Summary
Existing methods for correcting measurement errors in analytical devices, such as mass spectrometers, rely on pooled QC, which can be difficult to create, especially for solid samples, and require additional sample handling that delays analysis and reduces sample volume, making them inefficient.
A data processing method that estimates and uses 'in silico' pooled QC values based on previous sample measurements to correct measurement values without physically creating a pooled QC, allowing for immediate analysis and reducing user time and sample loss.
Enables accurate correction of measurement values without the need for a physical pooled QC, allowing for immediate analysis of samples, including solids, reducing sample handling time, and increasing the number of samples analyzed per unit time.
Smart Images
Figure 2025129508000001_ABST
Abstract
Description
[Technical Field]
[0001] The present disclosure relates to a data processing method, an information processing device, and a program, and more particularly to data processing for correcting measurement values obtained by an analysis device. [Background technology]
[0002] Measurement values obtained by an analytical device may be affected by temperature, humidity, and the state of the analytical device at the time of measurement, as well as differences between the individual devices. For example, in a mass spectrometer, contamination of the mass spectrometer may affect measurement sensitivity. Therefore, even when the same sample is measured using the same mass spectrometer, the measurement values obtained before and after cleaning the mass spectrometer may differ. Therefore, the measurement values obtained by an analytical device may contain measurement errors. Measurement errors include, for example, errors caused by different timings at which samples are measured, errors caused by different environments in which samples are measured, and errors caused by different instruments used to analyze the samples.
[0003] As a method for correcting such measurement errors, Dunn, Warwick B., et al. "Procedures for large-scale metabolic profiling of serum and plasma using gas chromatography and liquid chromatography coupled to mass spectrometry," Nature protocols 6.7 (2011): 1060-1083 (Non-Patent Document 1) discloses a technique for correcting sample measurement values based on the measurement values of a pooled QC in which all samples are mixed in equal amounts. In Non-Patent Document 1, the pooled QC is measured between sample measurements. After the measurement, LOESS smoothing is performed using only the pooled QC measurement values to calculate an approximate curve. The measurement values of each sample are corrected based on the calculated approximate curve. This method is called QC-based robust LOESS signal correction (QC-RLSC) method.
[0004] A pooled QC is created by mixing all samples in equal amounts. By mixing all samples in equal amounts, the amount of each component contained in the pooled QC is the average of all samples for that component. In other words, the measured value of the pooled QC approaches the average of the measured values of all samples, preventing the measured value of the pooled QC from being extremely different from the measured values of the samples. Furthermore, by mixing all samples to create a pooled QC, components contained in at least one of the samples will be included in the pooled QC. Therefore, even if the component to be corrected is unknown at the start of the measurement, correction processing can be performed on the measured value of that component after the measurement. [Prior art documents] [Non-patent literature]
[0005] [Non-Patent Document 1] Dunn, Warwick B., et al. "Procedures for large-scale metabolic profiling of serum and plasma using gas chromatography and liquid chromatography coupled to mass spectrometry." Nature protocols 6.7 (2011): 1060-1083. Summary of the Invention [Problem to be solved by the invention]
[0006] In the method disclosed in Non-Patent Document 1, a pooled QC is created by mixing equal amounts of all samples. Therefore, for example, when the samples are solid, it may be difficult to create a pooled QC. Therefore, there is a need for a data processing method that can correct sample measurements using an equivalent correction method without creating a pooled QC.
[0007] The present disclosure has been made in view of such circumstances, and an object thereof is to provide a technique for correcting measurement values obtained by an analyzer without using pooled QC.
Means for Solving the Problems
[0008] A data processing method according to a first aspect of the present disclosure is executed by a computer to correct the N-th (N is an integer satisfying 3 ≦ N) measurement value among measurement values of a predetermined component obtained by measuring a plurality of samples with an analyzer. The computer includes a processor and an interface. The data processing method includes: 1) a step of receiving, by the interface, a plurality of measurement values obtained by the analyzer; 2) a step of estimating, by the processor, each reference value obtained when measuring a standard sample obtained by mixing each of the plurality of samples at two or more measurements including at least the M-th and Q-th (M and Q are natural numbers satisfying M < N < Q) measurements; and 3) a step of correcting, by the processor, the N-th measurement value using each reference value. The step of estimating each reference value includes: a) a step of estimating, by the processor, the Q-th reference value obtained when measuring the standard sample at the Q-th measurement using the measurement values from the T-th (T is a natural number satisfying T < Q) to the Q-th; and b) a step of estimating, by the processor, the M-th reference value obtained when measuring the standard sample at the M-th measurement using the measurement values from the L-th (L is a natural number satisfying L < M) to the M-th.
[0009] An information processing apparatus according to a second aspect of the present disclosure includes at least one or more processors and a memory accessible to the one or more processors. The memory stores one or more instructions to be executed by the processor. The processor receives measurement values of a predetermined component obtained by measuring a plurality of samples with an analyzer by executing the one or more instructions, and estimates each reference value obtained when measuring a standard sample obtained by mixing each of the plurality of samples at least at two measurements including the M-th and Q-th measurements (M and Q are natural numbers satisfying M < Q). Using each reference value, the N-th measurement value (N is a natural number satisfying M < N < Q) is corrected. When estimating each reference value, the processor estimates the Q-th reference value obtained when measuring the standard sample at the Q-th measurement using the measurement values from T (T is a natural number satisfying T < Q) to the Q-th measurement value, and uses the measurement values from L (L is a natural number satisfying L < M) to the M-th measurement value to estimate the M-th reference value obtained when measuring the standard sample at the M-th measurement.
[0010] A program according to a third aspect of the present disclosure is executed by a processor mounted on a computer. The program causes the computer to receive measurement values of a predetermined component obtained by measuring a plurality of samples with an analyzer, estimate each reference value obtained when measuring a standard sample obtained by mixing each of the plurality of samples at least at two measurements including the M-th and Q-th measurements (M and Q are natural numbers satisfying M < Q), and correct the N-th measurement value (N is a natural number satisfying M < N < Q) using each reference value. The program causes the computer to estimate the Q-th reference value obtained when measuring the standard sample at the Q-th measurement using the measurement values from T (T is a natural number satisfying T < Q) to the Q-th measurement value, and estimate the M-th reference value obtained when measuring the standard sample at the M-th measurement using the measurement values from L (L is a natural number satisfying L < M) to the M-th measurement value when estimating each reference value.
Advantages of the Invention
[0011] According to the data processing method of the present disclosure, it is possible to correct the measurement values obtained by an analytical device without using pooled QC. [Brief explanation of the drawings]
[0012] [Figure 1] FIG. 1 is a schematic diagram of an analysis system according to an embodiment. [Figure 2] FIG. 10 is a diagram for explaining the procedure for creating a pooled QC. [Figure 3] FIG. 1 is a diagram illustrating pooled QC and the measurement order of samples. [Figure 4] FIG. 10 is a diagram for explaining a method for correcting sample measurement values based on pooled QC measurement values in a comparative example. [Figure 5] FIG. 10 is a diagram for illustrating a data processing method according to an embodiment. [Figure 6] FIG. 10 is a diagram for explaining a data processing method when analysis data is added. [Figure 7] 10 is a flowchart showing data processing according to the embodiment. [Figure 8] 8 is a flowchart showing a subroutine of step S20 shown in FIG. 7. DETAILED DESCRIPTION OF THE INVENTION
[0013] Hereinafter, embodiments of the present disclosure will be described in detail with reference to the drawings. Hereinafter, a mass spectrometer will be illustrated as an example of an analytical device, but the present invention is not limited thereto and can be applied to analytical devices in general. Note that the same or corresponding parts in the drawings will be designated by the same reference numerals, and their description will not be repeated.
[0014] [Overall configuration of the analysis system] FIG. 1 is a block diagram showing the configuration of an analysis system 100 according to an embodiment. Referring to FIG. 1, analysis system 100 includes a processing device 10, an input device 20, a display device 30, and a mass spectrometer 40. Analysis system 100 corrects measurement errors in sample measurements acquired by mass spectrometer 40. Measurement errors include, for example, errors caused by different timings at which samples are measured, errors caused by different environments in which samples are measured, and errors caused by different instruments used to analyze the samples. Note that processing device 10 may be incorporated into mass spectrometer 40.
[0015] The processing device 10 includes a processor 11, a memory 12, and an input / output interface (I / F) 13. These components are connected to each other via a bus so as to be able to communicate with each other.
[0016] The processor 11 is an example of an electric circuit, and controls the operation of the processing device 10 by executing a given program. The program executed by the processor 11 may be stored in the memory 12, or may be stored in a storage device (not shown) external to the processing device 10. The processor 11 is, for example, a CPU (Central Processing Unit).
[0017] The memory 12 can store programs executed by the processor 11, mass spectrum data created by the mass spectrometer 40, and component measurement values. The component measurement values are calculated, for example, from the area value of a peak in the mass spectrum and the height of the peak in the mass spectrum. The programs stored in the memory 12 include a correction program 121. The memory 12 includes volatile memory (e.g., RAM (Random Access Memory)) and non-volatile memory (e.g., ROM (Read Only Memory), a hard disk drive, and a solid state drive). The above programs may also be stored in an external storage device accessible by the processor 11.
[0018] The input / output I / F 13 is an interface for exchanging various data between the processor 11 and devices connected to the input / output I / F 13. An input device 20, a display device 30, and a mass spectrometer 40 are connected to the input / output I / F 13. The input / output I / F 13 is realized by, for example, a terminal block, a connector, and a network adapter. Data can be exchanged via the input / output I / F 13 wirelessly, such as via Bluetooth (registered trademark) or a wireless LAN, or via a wired connection using a Universal Serial Bus (USB) or the like. The processing device 10 can receive measurement values acquired by devices other than the mass spectrometer 40 via the input / output I / F 13 and perform data processing on the measurement values.
[0019] The input device 20 accepts information input from a user to the processing device 10. The information includes, for example, the total number of samples, the type of sample, and the components to be measured. The input device 20 is configured, for example, by a touch panel, a mouse, and a keyboard.
[0020] The display device 30 displays information in accordance with instructions from the processing device 10. The information may be, for example, the mass spectrum of the sample, the measured values of the predetermined components before correction, and the measured values of the predetermined components after correction. The display device 30 may be, for example, a liquid crystal display capable of displaying images.
[0021] The mass spectrometer 40 analyzes the sample and generates mass spectrum data that indicates the mass distribution of ions derived from components contained in the sample. The generated mass spectrum data is sent to the processing device 10. The processing device 10 calculates the measured values of each component from the mass spectrum data. The mass spectrometer 40 is, for example, a liquid chromatography mass spectrometer, a gas chromatography mass spectrometer, a high-performance liquid chromatography tandem mass spectrometer, or a gas chromatography tandem mass spectrometer.
[0022] [Comparative Example] In measurements using a mass spectrometer, contamination can occur on the mass spectrometer during repeated measurements, which can affect the analytical results. For example, even for the same sample, measurements taken on different days may not match. In other words, measurements obtained by a mass spectrometer may contain measurement errors.
[0023] A known method for correcting measurement errors is a technique for correcting sample measurements based on the measurement values of a standard sample, known as a pooled QC, as described in Non-Patent Document 1. Figure 2 illustrates the procedure for creating a pooled QC. Assume that there are N samples to be measured. As shown in Figure 2, the pooled QC is created by mixing all samples in equal amounts. By mixing all samples in equal amounts, the amount of each component contained in the pooled QC is the average of all samples for that component. In other words, the measured value of the pooled QC approaches the average of the sample measurements, preventing the measured value of the pooled QC from being extremely different from the measured value of the sample. Furthermore, by mixing all samples to create a pooled QC, components contained in at least one of the samples are included in the pooled QC. Therefore, even if the component to be corrected is unknown at the start of the measurement, correction processing can be performed on the measured value of that component after the measurement. The created pooled QC is dispensed into an appropriate number of aliquots. In Figure 2, the created pooled QC is divided into M aliquots.
[0024] The mass spectrometer measures the pooled QC before and after sample measurement, sequentially obtaining sample and pooled QC measurements. Figure 3 illustrates the pooled QC and sample measurement sequence. In Figure 3, the white containers represent the samples to be measured, and the gray containers represent the pooled QC. The operator separates an equal amount of solution from each of the 36 samples to create one pooled QC. The pooled QC is then divided into appropriate portions and measured at appropriate times before or after the sample measurements. For example, as shown in Figure 3, the mass spectrometer measures the pooled QC twice consecutively before starting sample measurement, and then measures the pooled QC once every four samples. After the last sample is measured, the pooled QC is measured twice consecutively.
[0025] As a result of the series of measurements performed as described above, sample measurement values and pooled QC measurement values are obtained sequentially. The processing device 10 corrects the sample measurement values based on the pooled QC measurement values. FIG. 4 is a diagram illustrating a method for correcting the sample measurement values based on the pooled QC measurement values. The graph shown at the top of FIG. 4 shows the measurement values obtained in the order in which the samples and pooled QC were measured. White circles indicate sample measurement values, and gray circles indicate pooled QC measurement values. Here, for example, the measurement value of the 10th sample measured, indicated by point P1, is greater than the measurement value of the 21st sample measured, indicated by point P2.
[0026] Ideally, the measured values of the pooled QC should be equal regardless of the timing of measurement. Therefore, in order to correct the measured values of the samples, LOESS smoothing is performed only on the measured values of the pooled QC, and an approximate curve L1 is calculated. The measured values of each sample are corrected based on the calculated approximate curve L1. The graph shown at the bottom of Figure 4 shows the relative measured values of the samples after correction. Note that the relative measured values in this graph are normalized values where the value indicated by the approximate curve L1 is set to 1. In the graph shown at the bottom of Figure 4, the measured values of the pooled QC have been corrected so that they are approximately constant.
[0027] Point P3 indicates the corrected measurement value of the 10th sample measured, and point P4 indicates the corrected measurement value of the 21st sample measured. Before correction, point P1 was larger than point P2, but after correction, point P3 is smaller than point P4. In this way, by using the QC-RLSC method, measurement errors can be corrected and measurement accuracy can be improved.
[0028] However, the pooled QC correction method has several issues. For example, (1) it may be difficult to create a pooled QC if the samples are solid. Even if a pooled QC can be created, (2) because all samples are required to create the pooled QC, measurement cannot begin until the last sample is acquired. Delays in starting the measurement may alter the quality of the acquired samples. (3) Creating a pooled QC requires dispensing all samples in equal amounts, which can increase the user's time. (4) If the sample volumes are small, creating a pooled QC may result in insufficient sample volume for analysis. (5) Because the concentration of components contained in some samples is low in the pooled QC, measurement of the pooled QC may fail to obtain the measured values of those components. Finally, (6) the pooled QC must be measured between sample measurements, which reduces the number of samples analyzed per unit time.
[0029] [Data processing method according to the embodiment] Therefore, in the data processing method according to this embodiment, the processing device 10 creates a pooled QC based on the measured values of the sample and estimates the value that would be obtained if the sample were measured at the time of measurement. The processing device 10 regards the estimated value as the measured value of the pooled QC and corrects the measured value of the sample. By using the data processing method according to this embodiment, the processing device 10 can correct the measured value of the sample using a correction method equivalent to the QC-RLSC method without using the actual measured value of the pooled QC.
[0030] According to the data processing method of this embodiment, there is no need to actually create a pooled QC, so measurement values can be corrected even if the sample is solid. Also, there is no need to prepare a pooled QC before starting sample measurement, so measurement can start before all samples have been acquired. Furthermore, the user's work time required to create a pooled QC can be reduced.
[0031] FIG. 5 is a diagram for explaining a method for correcting the measurement values of a sample obtained by mass spectrometer 40 using the data processing method according to this embodiment.
[0032] Referring to Figure 5, first, the mass spectrometer 40 analyzes a sample. The processing device 10 receives the measurement values of a predetermined component from the mass spectrometer 40. The processing device 10 may call up the measurement values of the sample stored in the memory 12, or may receive measurement values measured by a device other than the mass spectrometer 40. The mass spectrometer 40 measures the samples sequentially, so the measurement values can be arranged in the order in which they were measured, as shown in "Actual Analysis" in Figure 5.
[0033] Next, the measurement value of the pooled QC is estimated based on the "measurement data of the samples." This estimation is performed on a computer without actual analytical processing, and is therefore referred to as "in silico analysis." The value estimated by "in silico analysis" is also referred to as the "in silico pooled QC" value. The in silico pooled QC value is also referred to as the reference value. The reference value is the value estimated to be obtained when the pooled QC, a standard sample obtained by mixing the measurement values of the samples, is measured at the same time as the samples are measured.
[0034] In the correction process shown in the comparative example, the mass spectrometer 40 must actually measure the pooled QC. Therefore, it is physically impossible to simultaneously measure the pooled QC and the sample. On the other hand, in in silico analysis, the mass spectrometer 40 does not actually measure the pooled QC. Therefore, the reference value can be calculated assuming that the pooled QC is measured simultaneously with the sample measurement. In FIG. 5 , for example, it is assumed that the in silico pooled QC is measured at the timing when the first sample is measured and the timing when the fourth sample is measured. In this case, a reference value corresponding to the first measurement and a reference value corresponding to the fourth measurement are calculated. Although the reference value corresponding to the fourth measurement is the second reference value, it will be referred to as the fourth reference value for convenience. The reference value may be calculated for all measurement values, or may be selectively calculated for some measurement values.
[0035] Specifically, the kth (k is a natural number) reference value I(k) is calculated by the following formula (1).
[0036]
number
[0037] In Equation (1), P(i) represents the i-th measured value. In Equation (1), when calculating the k-th reference value, the measured values from the r-th (r is a natural number satisfying r < k) to the k-th are used. This is because the reference value to be calculated is affected by the measurements before that reference value.
[0038] Also, in Equation (1), the measured value is multiplied by a coefficient corresponding to the measurement order. This is because the reference value is more greatly affected by the measurements near the measurement for which the calculation is targeted. Therefore, the coefficient becomes larger for the measured values closer to the k-th measured value. The weighting function used for calculating the reference value is not limited to the exp function, and may be an inverse square function, a logistic function, or a hinge function.
[0039] Also, the k-th reference value I(k) may be calculated by the following Equation (2).
[0040] <In measurements using a mass spectrometer, multiple components contained in a sample can be quantified in a single measurement. In such cases, the measured values are corrected for each component.
[0044] Furthermore, in chromatograms created by a mass spectrometer, retention times may vary stochastically between measurements, even for the same component. In such cases, the accuracy of the obtained measurement values may vary between measurements. Therefore, in the above-mentioned correction, a confidence interval may be calculated and the corrected measurement value may be obtained by adding the confidence interval.
[0045] In the method described in this embodiment, all samples do not need to be collected before the start of measurement. Therefore, this data processing method allows for correction of measured values of samples contained in different batches, so-called inter-batch correction. Figure 6 is a diagram for explaining a data processing method when experimental data is added. As shown in box 6A in Figure 6, assume that a user conducts Experiment 1 and obtains 12 samples. The user measures the samples and corrects the obtained measured values using in silico analysis. In this case, the measurement of the 12 samples is referred to as a batch. Then, the user conducts Experiment 2 and obtains 12 new samples in addition to the samples obtained in Experiment 1. The measurement of the 12 samples obtained in Experiment 2 is a different batch from Experiment 1. To correct the measured values of the samples obtained in Experiment 2 using the correction method described in the comparative example, the user needs to mix the samples obtained in Experiment 1 and Experiment 2 to create a pooled QC, and then measure the samples obtained in Experiment 1 again. On the other hand, when correcting the measurement values of samples obtained in Experiment 2 using the method described in this embodiment, it is possible to perform the correction process by combining the measurement values of samples obtained in Experiment 2 with the measurement values of samples already obtained in Experiment 1, as shown in box 6B. Therefore, according to the method described in this embodiment, even when correcting the measurement values of samples included in different batches, the measurement values can be corrected without re-measuring the already measured samples.
[0046] [Measurement value processing] Before the measurement values are corrected as described above, the measurement values obtained by mass spectrometer 40 may be processed. Examples of the processing include detecting and excluding abnormal values and complementing missing values.
[0047] The process of detecting and excluding outliers involves detecting and excluding measurements that are significantly different from other measurements due to errors during measurement. For example, cluster analysis using score plots from PCA of all samples and detection methods using box plots can be used to detect outliers.
[0048] The process of imputing missing values is a process of imputing missing values using statistical methods. Missing values are measured values corresponding to samples that have been excluded as outliers in measurement data, and measured values corresponding to samples for which no measurements were recorded. Examples of statistical methods for imputing missing values include single imputation and multiple imputation.
[0049] By processing the measured values as described above, it is possible to prevent abnormal values and missing values from affecting the calculation of the reference value, which in turn prevents the reference value from becoming abnormal and affecting the approximation curve, thereby improving the accuracy of the measurement value correction.
[0050] [Data processing flow] 7 is a flowchart illustrating an example of processing measurement data acquired by measuring a sample using the mass spectrometer 40. In one implementation example, the data processing subroutine in FIG. 7 is called from the main routine and executed when the processor of the processing device 10 executes the correction program 121. The processing device 10 can be considered an example of an information processing device.
[0051] In FIG. 7, it is assumed that the Nth (N is an integer of 3 or more) measurement value of a predetermined component obtained by measuring a plurality of samples is to be corrected.
[0052] In step S10, the processing device 10 receives the chromatogram of each of a plurality of samples acquired by the mass spectrometer 40.
[0053] In step S12, the processing device 10 receives X types of target components to be corrected from the user.
[0054] In step S14, the processing device 10 extracts the measured values of one type of target component among the X types of target components from the chromatogram received in step S10, and receives the plurality of extracted measured values.
[0055] In step S16, the processing device 10 detects an abnormal value from the plurality of measured values received in step S12, and excludes the detected abnormal value.
[0056] In step S18, the processing device 10 complements the missing values included in the plurality of measured values received in step S12.
[0057] In step S20, the processing device 10 estimates each reference value obtained when measuring a standard sample obtained by mixing each of the plurality of samples at two or more measurement times. FIG. 8 shows a reference value estimation processing subroutine for step S20.
[0058] Referring to FIG. 8, in step S32, the processing device 10 estimates the Q-th (Q is an integer satisfying N < Q) reference value. The Q-th reference value is calculated, for example, by formula (1) or formula (2).
[0059] In step S34, the processing device 10 estimates the M-th (M is a natural number satisfying M < Q) reference value. The M-th reference value is calculated, for example, by formula (1) or formula (2). In step S20, the processing device 10 estimates two or more reference values including the Q-th reference value and the M-th reference value. Then, the processing device 10 ends the reference value estimation processing subroutine and returns control to FIG. 7.
[0060] Referring to FIG. 7, in step S22, the processing device 10 calculates an approximation curve using two or more reference values including the Qth reference value and the Mth reference value, and corrects the Nth measurement value.
[0061] In step S24, the processing device 10 calculates the confidence interval of the measurement value corrected in step S22.
[0062] In step S26, the processing device 10 displays on the display device 30 the corrected measurement value calculated in step S22 and the confidence interval calculated in step S24.
[0063] In step S28, processing device 10 determines whether or not processing of the X types of target components accepted in step S12 has been completed. If processing of all target components has been completed (YES in step S28), processing device 10 ends the data processing subroutine and returns processing to the main routine; otherwise (NO in step S28), processing returns processing to step S14.
[0064] In the above-described data processing, in step S12, the processing device 10 receives the target component from the user, but the processing device 10 may perform the data processing on all peaks present in the chromatogram received in step S10 without receiving the target component from the user.
[0065] The two or more reference values used to calculate the approximate curve must include at least the reference value before and the reference value after the measurement value to be corrected. As described in Figure 8, the reference values used to calculate the approximate curve must include the Mth reference value and the Qth reference value. The approximate curve may be calculated using only these two reference values, or the approximate curve may be calculated using the reference values corresponding to all measurements.
[0066] In the above-described data processing, the user does not need to create a pooled QC. Therefore, this data processing can be applied to analyses of samples for which it is difficult to create a pooled QC. An example of a sample for which it is difficult to create a pooled QC is a solid sample. An example of a solid sample is silicone rubber.
[0067] Furthermore, by using this data processing, sample analysis can be started before all samples have been collected. Therefore, the time from sample collection to completion of analysis can be shortened. This data processing can also be applied, for example, to the measurement of substances that are likely to undergo denaturation and aggregation due to repeated freezing and thawing. When creating a pooled QC, the first sample collected must be stored without being measured until the last sample is collected. Generally, biological samples are stored frozen. However, samples may contain substances that are likely to undergo denaturation and aggregation due to repeated freezing and thawing, and it is preferable to avoid freezing and thawing these samples. Since this data processing allows sample measurement to be started before all samples have been collected, the number of times the samples need to be frozen and thawed can be reduced. Therefore, this data processing can also be applied to the analysis of substances that are likely to undergo denaturation and aggregation due to repeated freezing and thawing. Examples of substances that are likely to undergo denaturation and aggregation due to repeated freezing and thawing include insulin and mucin.
[0068] In the above-described data processing, since it is not necessary to prepare a pooled QC, the user's work time for preparing the pooled QC can be reduced. Also, in the above-described data processing, since it is not necessary to set aside a portion of the acquired sample for preparing the pooled QC, the entire amount of the acquired sample can be used for measurement.
[0069] Furthermore, in the above-described data processing, even for components contained in only some of the samples, each measurement value can be corrected based on the obtained measurement values.
[0070] Furthermore, in the above-described data processing, there is no need to measure pooled QC, so the number of samples that can be analyzed per unit time can be increased.
[0071] According to the data processing method of the present disclosure, a computer can perform calculations related to correction of sample measurement values using in silico pooled QC values calculated using the measurement values of the sample to be measured, without obtaining the measurement values of pooled QCs that are used in calculations to correct measurement errors in measurement values obtained by an analytical device. This allows the computer to perform calculations related to correction processing of sample measurement values even if it cannot obtain the measurement values of pooled QCs.
[0072] [Aspect] It will be appreciated by those skilled in the art that the exemplary embodiments described above are examples of the following aspects.
[0073] (Item 1) A data processing method according to one aspect is a data processing method executed by a computer to correct the N-th (N is an integer satisfying 3 ≦ N) measurement value among the measurement values of a predetermined component obtained by measuring a plurality of samples with an analyzer. The computer includes a processor and an interface. The interface receives a plurality of measurement values obtained by the analyzer, and the processor estimates each reference value obtained when measuring a standard sample obtained by mixing each of the plurality of samples at least at two measurements including the M-th and Q-th (M and Q are natural numbers satisfying M < N < Q) measurements. The processor corrects the N-th measurement value using each reference value. The step of estimating each reference value includes the step of estimating the Q-th reference value obtained when measuring the standard sample at the Q-th measurement using the measurement values from T (T is a natural number satisfying T < Q) to the Q-th measurement by the processor, and the step of estimating the M-th reference value obtained when measuring the standard sample at the M-th measurement using the measurement values from L (L is a natural number satisfying L < M) to the M-th measurement by the processor.
[0074] According to the data processing method described in Item 1, the measurement values obtained by the analyzer can be corrected without using pooled QC.
[0075] (Item 2) In the data processing method according to Item 1, the step of estimating the Q-th reference value may include a step of weighting by multiplying a large coefficient to the measurement values obtained at a measurement time close to the Q-th measurement in the order of measurement, and the step of estimating the M-th reference value may include a step of weighting by multiplying a large coefficient to the measurement values obtained at a measurement time close to the M-th measurement in the order of measurement.
[0076] The data processing method described in paragraph 2 allows for the correction of measurements obtained by an analytical instrument without using pooled QC. The reference value used in this case is estimated by weighting the measurement value obtained during a measurement close in order to the target reference value by multiplying it by a large coefficient.
[0077] (Clause 3) In the data processing method described in clause 1, the step of estimating each reference value may further include a step of calculating an average value which is an average of the multiple measurement values, and the average value may be used in the step of estimating the Qth reference value and the step of estimating the Mth reference value.
[0078] According to the data processing method described in paragraph 3, it is possible to correct the measured values obtained by the analyzer without using pooled QC. The reference value used in this case is the average value of multiple measured values.
[0079] (4) In the data processing method according to claim 1 or 2, the step of estimating the Qth reference value includes a step of calculating the Qth reference value using equation (3) where r is T and k is Q, where P(k) is the kth measurement value and I(k) is the kth reference value,
[0080]
number
[0081] The step of estimating the Mth reference value may include the step of calculating the Mth reference value using equation (3) where r is L and k is M.
[0082] According to the data processing method described in Section 4, it is possible to correct the measured values obtained by the analyzer without using pooled QC. The reference value used in this case is calculated by Equation (3).
[0083] (Item 5) In the data processing method according to item 1 or 3, the step of estimating the Qth reference value includes a step of calculating the Qth reference value using equation (4) where the kth measurement value is P(k) and the kth reference value is I(k), where the average value of the plurality of measurement values is Pmean, r is T, and k is Q,
[0084]
number
[0085] The step of estimating the Mth reference value may include the step of calculating the Mth reference value using equation (4) where r is L and k is M.
[0086] According to the data processing method described in Section 5, it is possible to correct the measured values obtained by the analyzer without using pooled QC. The reference value used in this case is calculated by Equation (4).
[0087] (Item 6) In the data processing method according to any one of Items 1 to 5, T and L may be 1.
[0088] According to the data processing method described in paragraph 6, measurement values in a predetermined order are used in calculating reference values in a predetermined order, starting from the first measurement value, thereby improving the accuracy of calculation of the reference values.
[0089] (Clause 7) In the data processing method described in any one of clauses 1 to 6, the step of correcting the Nth measurement value may correct the Nth measurement value based on an approximate curve calculated by performing LOESS smoothing using the Qth reference value and the Mth reference value.
[0090] According to the data processing method described in paragraph 7, the measurement values are corrected based on the approximation curve calculated by LOESS smoothing.
[0091] (Item 8) The data processing method according to any one of items 1 to 7 may further comprise the step of detecting and excluding abnormal values from the measurement values.
[0092] According to the data processing method described in paragraph 8, the accuracy of correcting the measurement values can be improved by excluding abnormal values contained in the measurement values.
[0093] (Item 9) In the data processing method described in Item 8, the abnormal values may be detected using one of a cluster analysis method using a score plot from PCA of the measured values, and a detection method using a box plot.
[0094] According to the data described in Section 9, the accuracy of correction of the measurement values can be improved by detecting and excluding abnormal values contained in the measurement values by a cluster analysis method using score plots from PCA of the measurement values and a detection method using box plots.
[0095] (Item 10) The data processing method according to any one of items 1 to 9 may further comprise a step of complementing missing values in the measurement values.
[0096] According to the data processing method described in paragraph 10, the accuracy of correcting the measurement values can be improved by complementing missing values contained in the measurement values.
[0097] (Item 11) In the data processing method according to item 10, the step of imputing the missing values may impute the missing values using any one of a single imputation method and a multiple imputation method.
[0098] According to the data processing method described in paragraph 11, missing values included in the measured values are imputed using either a single imputation method or a multiple imputation method, thereby improving the accuracy of correction of the measured values.
[0099] (Item 12) In the data processing method according to any one of Items 1 to 11, a step of deriving a confidence interval of a corrected value in the step of correcting the N-th measured value may be further provided.
[0100] According to the data processing method described in Item 12, the user can recognize the confidence interval of the corrected value and can easily determine whether to use the data for analysis.
[0101] (Item 13) In the data processing method according to any one of Items 1 to 12, the analysis device may be a mass spectrometer.
[0102] According to the data processing method described in Item 13, the measured values obtained by the mass spectrometer can be corrected without using pooled QC.
[0103] (Item 14) An information processing device according to one aspect includes at least one or more processors and a memory accessible to the one or more processors. The memory stores one or more instructions executed by the processor. By executing the one or more instructions, the processor receives measured values of a predetermined component obtained by measuring a plurality of samples with an analysis device, and at two or more measurement times including at least the M-th and Q-th (M and Q are natural numbers satisfying M < Q) measurement times, estimates each reference value obtained when measuring a standard sample obtained by mixing each of the plurality of samples, corrects the N-th (N is a natural number satisfying M < N < Q) measured value using each reference value, and when estimating each reference value, uses the Q-th measured value from T (T is a natural number satisfying T < Q) to estimate the Q-th reference value obtained when measuring the standard sample at the Q-th measurement time, and uses the M-th measured value from L (L is a natural number satisfying L < M) to estimate the M-th reference value obtained when measuring the standard sample at the M-th measurement time.
[0104] According to the information processing device described in Item 14, the measured values obtained by the analysis device can be corrected without using pooled QC.
[0105] (Item 15) When a program according to an aspect is executed by a processor mounted on a computer, the computer is caused to receive measurement values of a predetermined component obtained by measuring a plurality of samples with an analyzer, and when measuring at least two measurements including the M-th and Q-th measurements (M and Q are natural numbers satisfying M < Q), estimate each reference value obtained when measuring a standard sample obtained by mixing each of the plurality of samples, and use each of the reference values to correct the N-th measurement value (N is a natural number satisfying M < N < Q). When estimating each of the reference values, use the measurement value from the T-th (T is a natural number satisfying T < Q) to the Q-th to estimate the Q-th reference value obtained when measuring the standard sample at the Q-th measurement, and use the measurement value from the L-th (L is a natural number satisfying L < M) to the M-th to estimate the M-th reference value obtained when measuring the standard sample at the M-th measurement.
[0106] According to the program described in Item 15, the measurement values obtained by the analyzer can be corrected without using pooled QC.
[0107] The embodiments disclosed this time should be considered as illustrative in all respects and not restrictive. The scope of the present disclosure is indicated by the claims rather than the description of the above embodiments, and it is intended that all modifications within the meaning and scope equivalent to the claims are included. Also, each technology in the embodiments can be implemented alone or, if necessary, in combination with other technologies in the embodiments as much as possible.
Explanation of Reference Numerals
[0108] 10 Processing device, 11 Processor, 12 Memory, 13 Input / output I / F, 20 Input device, 30 Display device, 40 Mass spectrometer, 100 Analysis system, 12I Correction program.
Claims
1. 1. A data processing method executed by a computer to correct an Nth (N is an integer satisfying 3≦N) measurement value among measurement values of a predetermined component obtained by measuring a plurality of samples with an analytical device, comprising: the computer comprises a processor and an interface; receiving, by the interface, a plurality of measurements obtained by the analytical device; a step of estimating, by the processor, each of the reference values obtained when measuring a standard sample obtained by mixing each of the plurality of samples at two or more measurements including at least the Mth and Qth measurements (M and Q are natural numbers satisfying M<N<Q); correcting the Nth measurement value using each of the reference values by the processor; The step of estimating each reference value includes: A step of estimating, by the processor, a Qth reference value obtained when the standard sample is measured at the Qth measurement time using a Qth measurement value from T (T is a natural number satisfying T<Q); and estimating, by the processor, the Mth reference value that would be obtained if the standard sample were measured at the Mth measurement time using L to Mth measurement values (L is a natural number satisfying L<M).
2. the step of estimating the Qth reference value includes a step of weighting a measurement value obtained during a measurement that is closer to the Qth measurement in the measurement order by multiplying the measurement value by a larger coefficient; 2. The data processing method according to claim 1, wherein the step of estimating the Mth reference value comprises a step of weighting a measurement value obtained during a measurement that is closer to the Mth measurement in measurement order by multiplying the measurement value by a large coefficient.
3. the step of estimating each reference value further comprises the step of calculating an average value that is an average of the plurality of measurement values; The data processing method according to claim 1 , wherein the average value is used in the steps of estimating the Qth reference value and estimating the Mth reference value.
4. the step of estimating the Q-th reference value includes a step of calculating the Q-th reference value by equation (1) where r is T and k is Q, where P(k) is the k-th measurement value and I(k) is the k-th reference value; [Equation 1] 3. The data processing method according to claim 1, wherein the step of estimating the Mth reference value comprises the step of calculating the Mth reference value using equation (1) where r is L and k is M.
5. the step of estimating the Q-th reference value includes a step of calculating the Q-th reference value by equation (2) where the k-th measurement value is P(k), the k-th reference value is I(k), the average value of the plurality of measurement values is Pmean, r is T, and k is Q, [Equation 2] 4. The data processing method according to claim 1, wherein the step of estimating the Mth reference value comprises the step of calculating the Mth reference value using equation (2) where r is L and k is M.
6. 4. The data processing method according to claim 1, wherein said T and said L are 1.
7. 4. The data processing method according to claim 1, wherein the step of correcting the N-th measurement value corrects the N-th measurement value based on an approximate curve calculated by performing LOESS smoothing using the Q-th reference value and the M-th reference value.
8. 4. The data processing method according to claim 1, further comprising the step of detecting and excluding abnormal values from the measurement values.
9. The data processing method according to claim 8 , wherein the outliers are detected using one of a cluster analysis method using a score plot from PCA of the measured values and a detection method using a box plot.
10. 4. The data processing method according to claim 1, further comprising the step of imputing missing values in the measurement values.
11. The data processing method according to claim 10 , wherein the step of imputing missing values imputes the missing values using one of a single imputation method and a multiple imputation method.
12. 4. The data processing method according to claim 1, further comprising the step of deriving a confidence interval for the value corrected in the step of correcting the Nth measurement value.
13. 4. The data processing method according to claim 1, wherein the analysis device is a mass analysis device.
14. at least one processor; memory accessible to the one or more processors; the memory stores one or more instructions to be executed by the processor; The processor executes the one or more instructions to: Accepting measurement values of predetermined components obtained by measuring a plurality of samples with an analytical device; At least two or more measurements, including the Mth and Qth measurements (M and Q are natural numbers satisfying M<Q), estimate each reference value obtained when a standard sample obtained by mixing each of the plurality of samples is measured; correcting the Nth measurement value (N is a natural number satisfying M<N<Q) using each of the reference values; When estimating each of the reference values, Using a Qth measurement value from T (T is a natural number satisfying T<Q), a Qth reference value obtained when the standard sample is measured at the Qth measurement is estimated; An information processing device that estimates an Mth reference value obtained when the standard sample is measured at the Mth measurement time using Mth measurement values from L (L is a natural number satisfying L<M).
15. When executed by a processor installed in a computer, the computer Accepting measurement values of predetermined components obtained by measuring a plurality of samples with an analytical device; At least two or more measurements including the Mth and Qth measurements (M and Q are natural numbers satisfying M<Q), estimating each reference value obtained when a standard sample obtained by mixing each of the plurality of samples is measured; correcting the Nth measurement value (N is a natural number satisfying M<N<Q) using each of the reference values; When estimating each of the reference values, Using a Qth measurement value from T (T is a natural number satisfying T<Q), estimate a Qth reference value obtained when the standard sample is measured at the Qth measurement time; and estimating the Mth reference value that would be obtained when the standard sample is measured at the Mth measurement time, using the Mth measurement values from L (L is a natural number that satisfies L<M).