Signal processing method, signal processing device, and signal processing system

Abstract

A data processing device 12 is a signal processing device that processes the outputs of photomultiplier tubes 11a to 11d and that includes a processor which is configured to: acquire, in a time series, intensity signals output from the photomultiplier tubes 11a to 11d, which detect signal light generated by flow cytometry using a flow cytometer targeting a plurality of test samples; aggregate the frequency distribution of the intensities of the intensity signals on the basis of digital values DN of the time-series intensity signals, to generate a histogram; fit the histogram using an exponential function to calculate a ρ value representing the variation of each peak in the frequency distribution; and perform data analysis using the ρ value for the intensity signals output by the photomultiplier tubes 11a to 11d obtained by flow cytometry, targeting a plurality of target samples.

Description

Signal processing method, signal processing device, and signal processing system

[0001] One aspect of the embodiment relates to a signal processing method, a signal processing device, and a signal processing system.

[0002] Flow cytometry has been known for some time as a technique for counting, sorting, and analyzing the characteristics of samples such as cells using laser light. For example, Patent Document 1 listed below discloses a flow cytometry system including a first sensor arranged axially relative to a light source and detecting forward-scattered components, and a second sensor arranged at an angle relative to the first sensor and detecting side-scattered components and / or fluorescent components. Furthermore, Non-Patent Document 1 listed below discloses a technique for fitting a histogram of fluorescent signals using a Gaussian distribution in cell cycle analysis using a DNA-stained sample.

[0003] Special Publication No. 2013-504051

[0004] James V. Watson, et al., “A Pragmatic Approach to the Analysis of DNA Histograms”, Cytometry 8:l-8 (1987)

[0005] In conventional flow cytometry, when analyzing the intensity signal values ​​output by the sensor, the calculations required for analysis tend to be complicated due to the presence of various factors that affect the variability of the intensity signal values. As a result, the calculation time required to evaluate the variability of the intensity signal values ​​tends to be long. On the other hand, conventional fitting techniques may not be able to properly evaluate the distribution of the intensity signal values.

[0006] Therefore, one aspect of the embodiments has been made in consideration of such problems, and aims to provide a signal processing method, a signal processing device, and a signal processing system that are capable of accurately evaluating data variability in flow cytometry through simple calculations.

[0007] A signal processing method according to a first aspect of an embodiment is a signal processing method for processing the output of a photodetector constituting a flow cytometer, in which output current signals from the photodetector that detect signal light generated by flow cytometry using a flow cytometer targeting a plurality of test samples are acquired in a time series, a frequency distribution of the intensity of the output current signals is compiled based on the values ​​of the output current signals in the time series to generate a histogram, a ρ value representing the variability of each peak in the frequency distribution is calculated by fitting the histogram using an exponential function, and data analysis is performed using the ρ value for the output current signals from the photodetector obtained by flow cytometry targeting a plurality of target samples.

[0008] Alternatively, a signal processing device according to a second aspect of the embodiment is a signal processing device that processes the output of a photodetector that constitutes a flow cytometer and includes a processor, and the processor is configured to acquire, in a time series, output current signals from the photodetector that detect signal light generated by flow cytometry using a flow cytometer targeting a plurality of test samples, generate a histogram by aggregating the frequency distribution of the intensity of the output current signals based on the values ​​of the time series output current signals, calculate a ρ value that represents the variation of each peak in the frequency distribution by fitting the histogram using an exponential function, and perform data analysis using the ρ value for the output current signals from the photodetector obtained by flow cytometry targeting a plurality of target samples.

[0009] Alternatively, a signal processing system according to a third aspect of the embodiment includes the above-described signal processing device, a photodetector, and an optical system that guides signal light to the photodetector.

[0010] According to the first, second, or third aspect, a histogram is generated by chronologically aggregating the values ​​of current signals output by a photodetector of a flow cytometer when multiple test samples are used. The histogram is then fitted with an exponential function to calculate a ρ value representing the variability of each peak in the frequency distribution. Data analysis is then performed using the ρ value for the current signals output by the photodetector of the flow cytometer when a target sample is used. This allows for accurate evaluation of the variability of the output values ​​of the photodetector through simple calculations. As a result, the computational cost of data analysis can be reduced while maintaining the accuracy of the data analysis.

[0011] According to any one of the aspects of the present invention, it is possible to accurately evaluate the variation in output signals for a plurality of fluorescent components.

[0012] FIG. 1 is a schematic diagram of a flow cytometer system 1, which is a flow cytometer according to an embodiment. FIG. 2 is a block diagram showing an example of the hardware configuration of the data processing device 12 of FIG. 1. FIG. 3 is a block diagram showing the functional configuration of the data processing device 12. FIG. 4 is a graph showing an example of a histogram generated by the calculation unit 202 of FIG. 5. FIG. 6 is a graph showing a fitting function p single 4 is a graph showing the error between the ρ value calculated by equation (5) and the ρ value calculated by equations (2) to (4). FIG. 5 is a graph showing an example of a dot plot generated and output in data analysis by the analysis unit 203 of FIG. 3. FIG. 6 is a graph showing an example of a dot plot generated and output in data analysis by the analysis unit 203 of FIG. 3. FIG. 7 is a flowchart showing the procedure of a signal processing method according to an embodiment. FIG. 8 is a flowchart showing the procedure of a signal processing method according to an embodiment. FIG. 9 is a graph showing the error between the ρ value calculated by equation (5) and the ρ value calculated by equations (7) to (8).

[0013] In the first and second aspects, it is preferable that the exponential function is a function that exponentially increases in a predetermined range of a variable x corresponding to intensity and becomes zero outside the predetermined range of the variable x. In this case, the variation in data of the output value of the photodetector can be evaluated with higher accuracy.

[0014] In the first and second aspects, the exponential function is represented by the formula (1): [In the above formula, A is a parameter that determines the upper limit of the range of the distribution, B is a parameter that determines the lower limit of the range of the distribution, τ is a parameter that determines the time constant of the distribution, and u is a step function.] ｓｉｎｇｌｅ In this case, the variation in the data of the output value of the photodetector can be evaluated with higher accuracy.

[0015] In the first aspect, when fitting an exponential function, the parameters A and B are determined by defining the range of one peak in the frequency distribution, and then the function p ｓｉｎｇｌｅ It is also preferable to determine the parameter τ by fitting (x) to one peak. In addition, in the second aspect, when fitting an exponential function, the processor determines the parameters A and B by defining the range of one peak in the frequency distribution, and then determines the parameter τ by fitting the function p ｓｉｎｇｌｅ It is also preferable to determine the parameter τ by fitting (x) to one peak. In this case, when fitting an exponential function to a histogram based on the values ​​of the current signal output from the photodetector, a stably fitted exponential function can be obtained by determining the parameters A, B, and τ in that order. As a result, the ρ value that matches the peak of the frequency distribution can be stably calculated.

[0016] In the first aspect, it is also preferable that the processor calculates the ρ value of each peak by fitting each peak of the histogram with an exponential function, and then selecting from a plurality of approximation formulas that approximate the width of the peak of the exponential function and performing calculations. In the second aspect, it is also preferable that the processor calculates the ρ value of each peak by fitting each peak of the histogram with an exponential function, and then selecting from a plurality of approximation formulas that approximate the width of the peak of the exponential function and performing calculations. In this case, it is possible to select an approximation formula that is suitable for approximating the width of the peak of the exponential function and calculate that width, so that the ρ value that matches the multiple peaks included in the frequency distribution can be accurately evaluated by a simple calculation.

[0017] In the first aspect, each peak of the histogram is calculated using a function p ｓｉｎｇｌｅ After fitting with (x), the function p ｓｉｎｇｌｅ It is also preferable that the ρ value of each peak is calculated by selecting one approximation formula from a plurality of approximation formulas that approximate the width of the peak of (x) in accordance with the parameters A and τ. ｓｉｎｇｌｅ After fitting with (x), the function p ｓｉｎｇｌｅ It is also preferable to calculate the ρ value of each peak by selecting one approximation formula from a plurality of approximation formulas that approximate the width of the peak of (x) according to the parameters A and τ. In this case, an approximation formula suitable for approximating the width of the exponential function peak can be selected according to the two parameters A and τ determined by fitting, and the width can be calculated, so that the ρ value that matches the plurality of peaks included in the frequency distribution can be accurately evaluated by a simple calculation.

[0018] In the first aspect or the second aspect, the plurality of approximate expressions include approximate expressions represented by formulas (2) to (4). In this case, an approximation formula suitable for approximating the width of the peak of the exponential function can be selected from formulas (2) to (4) according to the two parameters A and τ determined by fitting, and the width can be calculated, so that the ρ value that matches multiple peaks included in the frequency distribution can be evaluated with high accuracy by simple calculation.

[0019] In the first or second aspect, it is also preferable that the data analysis includes a gating process for defining the boundaries of the population to be analyzed, which can reduce the computational cost of the gating process and maintain the accuracy of the gating process.

[0020] The signal processing method of the embodiment is [1] "a signal processing method for processing the output of a photodetector constituting a flow cytometer, the signal processing method comprising: acquiring, in a time series, output current signals of the photodetector that detect signal light generated by flow cytometry using the flow cytometer for a plurality of test samples; generating a histogram by aggregating a frequency distribution of the intensity of the output current signals based on the time series values ​​of the output current signals; calculating a ρ value that represents the variability of each peak in the frequency distribution by fitting the histogram using an exponential function; and performing data analysis using the ρ value for the output current signals of the photodetector obtained by flow cytometry for a plurality of target samples."

[0021] The signal processing method of the embodiment may be [2] "the signal processing method according to the above [1], wherein the exponential function is a function that exponentially increases with respect to a variable x corresponding to the intensity within a predetermined range of the variable x, and becomes zero outside the range of the predetermined variable x."

[0022] In the signal processing method of the embodiment, [3] "the exponential function is a function p ｓｉｎｇｌｅ The signal processing method may be the signal processing method according to the above item [2], wherein (x) is satisfied.

[0023] The signal processing method of the embodiment includes the steps of: [4] "When fitting the exponential function, parameters A and B are determined by defining the range of one peak in the frequency distribution, and then the function p ｓｉｎｇｌｅ The signal processing method may be the signal processing method according to the above item [3], in which the parameter τ is determined by fitting (x) to the one peak.

[0024] The signal processing method of the embodiment may be [5] "the signal processing method according to any one of the above [1] to [4], in which, after fitting each peak of the histogram with the exponential function, the ρ value of each peak is calculated by selecting from a plurality of approximation formulas that approximate the width of the peak of the exponential function and performing calculation."

[0025] The signal processing method of the embodiment includes the steps of: [6] "subtracting each peak of the histogram from the function p ｓｉｎｇｌｅ After fitting with (x), the function p ｓｉｎｇｌｅ The signal processing method may be the one described in [3] or [4] above, in which one approximation formula is selected from a plurality of approximation formulas that approximate the width of the peak of (x) in accordance with the parameter A and the parameter τ, and the ρ value of each peak is calculated by performing calculation.

[0026] The signal processing method of the embodiment may be [7] "the signal processing method according to the above [6], in which the plurality of approximate expressions include approximate expressions expressed by expressions (2) to (4)."

[0027] The signal processing method of the embodiment may be [8] "the signal processing method according to any one of [1] to [7] above, wherein the data analysis includes a gating process that defines the boundaries of the population to be analyzed."

[0028] The signal processing device of the embodiment is [9] "a signal processing device that processes the output of a photodetector constituting a flow cytometer and includes a processor, wherein the processor is configured to acquire, in a time series, output current signals of the photodetector that detect signal light generated by flow cytometry using the flow cytometer for a plurality of test samples, to generate a histogram by aggregating a frequency distribution of the intensity of the output current signals based on the time series values ​​of the output current signals, to calculate a ρ value that represents the variation of each peak in the frequency distribution by fitting the histogram using an exponential function, and to perform data analysis using the ρ value for the output current signals of the photodetector obtained by flow cytometry for a plurality of target samples."

[0029] The signal processing device of the embodiment may be

[10] "the signal processing device according to the above [9], wherein the exponential function is a function that increases exponentially with respect to a variable x corresponding to the intensity within a predetermined range of the variable x, and becomes zero outside the range of the predetermined variable x."

[0030] The signal processing device of the embodiment is

[11] "The exponential function is a function p ｓｉｎｇｌｅ The signal processing device may be the signal processing device according to the above

[10] , wherein (x) is satisfied.

[0031] The signal processing device of the embodiment is

[12] "When fitting the exponential function, parameters A and B are determined by defining the range of one peak in the frequency distribution, and then the function p ｓｉｎｇｌｅ The signal processing method may be the signal processing method according to the above

[11] , in which the parameter τ is determined by fitting (x) to the one peak.

[0032] The signal processing device of the embodiment may be

[13] "the signal processing device according to any one of the above [9] to

[12] , wherein the processor calculates the ρ value of each peak by fitting each peak of the histogram with the exponential function and then selecting from a plurality of approximation formulas that approximate the width of the peak of the exponential function."

[0033] In the signal processing device of the embodiment,

[14] "the processor calculates each peak of the histogram by the function p ｓｉｎｇｌｅ After fitting with (x), the function p ｓｉｎｇｌｅ The signal processing device may be the one described in

[11] or

[12] above, which calculates the ρ value of each peak by selecting one approximation formula from a plurality of approximation formulas that approximate the width of the peak of (x) in accordance with the parameter A and the parameter τ.

[0034] The signal processing device of the embodiment may be

[15] "the signal processing device according to the above

[14] , in which the plurality of approximate expressions include approximate expressions expressed by expressions (2) to (4)."

[0035] The signal processing device of the embodiment may be the signal processing device according to any one of the above items [9] to

[15] , wherein

[16] the data analysis includes a gating process for defining the boundaries of the population to be analyzed.

[0036] Hereinafter, embodiments of the present invention will be described in detail with reference to the accompanying drawings. In the description, the same elements or elements having the same functions will be denoted by the same reference numerals, and redundant description will be omitted.

[0037] 1 is a schematic diagram of a flow cytometer (signal processing system) according to an embodiment, which is a flow cytometer system 1. The flow cytometer system 1 is a system for performing flow cytometry, and is composed of a fluid system 2, an optical system (optical system) 3, and an electronic system (signal processing device) 4.

[0038] The fluid system 2 includes a flow cell 6 into which a sample fluid containing analytes such as cells or particles is injected, and which allows the analytes (target sample) contained in the sample fluid to be aligned and passed through a narrow channel 5. The flow cell 6 also has a function (not shown) for sorting (classifying and distributing) the gated analytes by controlling an electric field or the like.

[0039] The optical system 3 optically analyzes an analyte passing through a flow cell 6 by flow cytometry. The optical system 3 includes a laser light source 7, a lens 8, filters 9a, 9b, 9c, and 9d, dichroic mirrors 10b and 10c, and photomultiplier tubes (photodetectors) 11a, 11b, 11c, and 11d. Various types of light generated from the analyte by flow cytometry are guided to the photomultiplier tubes 11a, 11b, 11c, and 11d. The laser light source 7 is a light source device that generates laser light (excitation light) in a single wavelength band at a specific frequency. The lens 8 focuses the laser light emitted from the laser light source 7 onto a channel 5 in the flow cell 6. The filter 9a transmits forward-scattered light generated from the sample fluid by irradiation with the laser light. The dichroic mirror 10b reflects side-scattered light generated from the sample fluid by irradiation with the laser light and transmits fluorescence generated from the sample fluid. The dichroic mirror 10c reflects fluorescence in a first wavelength band from the fluorescence that has passed through the dichroic mirror 10b and transmits fluorescence in the remaining wavelength band from the transmitted fluorescence. The filter 9b transmits side-scattered light reflected by the dichroic mirror 10b, and the filter 9c transmits first fluorescence in the first wavelength band reflected by the dichroic mirror 10c. The filter 9d transmits second fluorescence in a second wavelength band from the fluorescence that has passed through the dichroic mirror 10c. The photomultiplier tubes 11a, 11b, 11c, and 11d are disposed on the optical axes of the forward scattered light, side scattered light, first fluorescence, and second fluorescence, respectively, and measure the intensities of the forward scattered light, side scattered light, first fluorescence, and second fluorescence.

[0040] The electronic system 4 includes a data processing device 12 and is a device for analyzing the light intensity measured by the optical system 3. Specifically, the data processing device 12 is electrically connected to the multiple photomultiplier tubes 11a, 11b, 11c, and 11d, and performs data analysis to create histograms or dot plots (also called cytograms) based on intensity signals indicating the intensities detected in each channel of the multiple photomultiplier tubes 11a, 11b, 11c, and 11d, as well as gating processing. Furthermore, the data processing device 12 sorts (classifies and distributes) analytes contained in the sample fluid based on the gating processing.

[0041] Next, the configuration of the data processing device 12 will be described with reference to Figures 2 and 3. Figure 2 is a block diagram showing an example of the hardware configuration of the data processing device 12, and Figure 3 is a block diagram showing the functional configuration of the data processing device 12.

[0042] 2, the data processing device 12 is physically a computer or the like including a processor such as a CPU (Central Processing Unit) 101, a recording medium such as a RAM (Random Access Memory) 102 or a ROM (Read Only Memory) 103, a communication module 104, and an input / output module 106, all of which are electrically connected to one another. The data processing device 12 may include input / output devices such as a display, a keyboard, a mouse, a touch panel display, or a data recording device such as a hard disk drive or semiconductor memory. The data processing device 12 may also be composed of multiple computers.

[0043] As shown in FIG. 3 , the data processing device 12 includes, as functional components, a signal acquisition unit 201, a calculation unit 202, and an analysis unit 203. Each functional unit of the data processing device 12 shown in FIG. 3 is realized by loading a program onto hardware such as the CPU 101 and RAM 102, which operates the communication module 104 and the input / output module 106 under the control of the CPU 101, and reads and writes data from and to the RAM 102. The CPU 101 of the data processing device 12 executes the program to cause each functional unit of FIG. 3 to function and sequentially executes processing corresponding to the signal processing method described below. The CPU 101 may be a standalone piece of hardware or may be implemented in a programmable logic device such as an FPGA, like a software processor. The RAM and ROM may also be standalone pieces of hardware or may be built into a programmable logic device such as an FPGA. Various data required for executing the program and various data generated by the program are all stored in built-in memory such as the ROM 103 and RAM 102, or in a recording medium such as a hard disk drive. The functionality of the functional components of data processing device 12 will now be described in detail.

[0044] The signal acquisition unit 201 acquires intensity signals (output current signals) output from each channel of the multiple photomultiplier tubes 11a, 11b, 11c, and 11d. The acquired intensity signals are analog signals obtained by detecting currents due to multiplied electrons corresponding to the intensity of signal light, such as forward scattered light, side scattered light, or fluorescence, generated by flow cytometry in each photomultiplier tube. The signal acquisition unit 201 converts the acquired intensity signals of each channel into digital values ​​DN and outputs them to the calculation unit 202. Note that the A / D conversion function of the signal acquisition unit 201 may be realized by an external circuit unit of the data processing device 12.

[0045] The calculation unit 202 has a function of calculating parameters for defining a group of photons incident on each channel, using the digital values ​​DN of each channel output from the signal acquisition unit 201 when a plurality of test samples contained in the sample fluid are aligned and passed through the channel 5. Details of the parameter calculation function performed by the calculation unit 202 will be described below.

[0046] The calculation unit 202 generates a histogram by aggregating the frequency distribution of the intensity indicated by the intensity signal based on the time-series digital values ​​DN of one channel acquired by the signal acquisition unit 201. Fig. 4 shows an example of the histogram generated by the calculation unit 202. In this way, the histogram generated by the calculation unit 202 shows the distribution of the frequency of each digital value DN indicating the intensity, with multiple peaks indicating groups of photons appearing.

[0047] The calculation unit 202 performs fitting using a fitting function, which is an exponential function, for each of the multiple peaks that appear in the generated histogram. At this time, the calculation unit 202 uses the fitting function p single (x); Here, in the above formula (1), x represents a variable corresponding to the intensity, A represents the upper limit of the range of the distribution, B represents a positive parameter that determines the lower limit of the range of the distribution, τ represents a positive parameter that determines the time constant of the distribution, and u represents a step function, that is, a function where u(x) = 0 when x < 0 and u(x) = 1 when x ≥ 0. This fitting function p single (x) is a function whose value increases exponentially in the range of B≦x≦A with respect to the variable x, and has a value of zero outside that range. single (x) is a graph showing the fitting function p single The width of the distribution peak of (x) is determined by parameters A and B, and the slope of the peak is determined by parameter τ (the larger τ is, the gentler the slope becomes, and the smaller τ is, the steeper the slope becomes).

[0048] Specifically, the calculation unit 202 performs fitting as follows. First, the calculation unit 202 selects one peak in the histogram and determines the width PW of that peak (see FIG. 4 ). The determination of the width PW may be performed automatically based on the frequency distribution, or may be performed via user input based on a graph of the frequency distribution displayed on an output device such as a display. Next, the calculation unit 202 uses the determined width PW to set parameters A and B to values ​​corresponding to PW.

[0049] Thereafter, the calculation unit 202 calculates a fitting function p single (x) is fitted to the one peak in the histogram, and the value of the parameter τ is sequentially changed to find a fitting function p single (x). At this time, the calculation unit 202 searches for the fitting function p single By multiplying (x) by the integral value of the histogram within the range PW, the fitting function p single (x). The calculation unit 202 then calculates a fitting function p single The parameters A, B, and τ at the time when (x) was found, and the fitting function p single The position on the intensity axis of (x) is stored.

[0050] Furthermore, the calculation unit 202 uses the fitting results for one peak in the histogram to calculate a ρ value representing the variation (width) of the peak and the average intensity value of the peak. The ρ value is the ratio of the standard deviation σ to the average intensity of the peak.

[0051] The inventors of the present invention have derived the theoretical value of the ρ value to obtain the fitting function p single The ρ value for the peak represented by (x) is calculated by the following formula (5): The inventors of the present application have also found that the ρ value calculated by the above formula (5) can also be approximated by the approximation formulas expressed by the following formulas (2) to (4).

[0052] 6 is a graph showing the error between the ρ value calculated by the above formula (5) and the ρ value calculated by the above formulas (2) to (4). This graph shows the change in the value of each of formulas (2) to (4) normalized by the value of formula (5) when the ratio τ / A of the two parameters A and τ is changed. The inventors of the present application have clarified that the value calculated by formula (2) has a small error in the range of τ / A≦predetermined value TH1 (=0.4 to 0.6), the value calculated by formula (3) has a small error in the range of τ / A≦predetermined value TH2 (=0.2 to 0.3), and the value calculated by formula (4) has a small error in the range of τ / A≦predetermined value TH3 (=0.1 to 0.2).

[0053] Using the above property, the calculation unit 202 selects and uses formula (4) to calculate the ρ value based on the values ​​of two parameters A and τ, which are fitting results for one peak, when τ / A≦predetermined value TH3. On the other hand, the calculation unit 202 selects and uses formula (3) to calculate the ρ value when predetermined value TH3<τ / A≦predetermined value TH2. Furthermore, the calculation unit 202 selects and uses formula (2) to calculate the ρ value when predetermined value TH2<τ / A≦predetermined value TH1. In other cases, the calculation unit 202 selects and uses formula (5) to calculate the ρ value. In addition, the calculation unit 202 uses the fitting results for one peak to calculate the average value (expected value) E at that peak using formula (6) below, and applies the fitting function p used during fitting to the average value E. single The average intensity at the peak is calculated by reflecting its position on the (x) intensity axis.

[0054] Furthermore, the calculation unit 202 repeatedly performs fitting on a plurality of peaks in the histogram, calculates the ρ value and the average value of the intensity for each peak, and stores them.

[0055] Referring back to FIG. 3 , the analysis unit 203 has a function of performing data analysis to analyze groups of photons incident on each channel using the digital values ​​DN of each channel output from the signal acquisition unit 201 when multiple target samples contained in the sample fluid are aligned and passed through the channel 5. That is, the analysis unit 203 performs data analysis based on the average intensity and ρ value for each channel calculated by the calculation unit 202. Specifically, the analysis unit 203 generates dot plots based on the digital values ​​DN of the multiple channels and outputs them to an input / output device. Furthermore, the analysis unit 203 performs gating processing on the generated dot plots using the average intensity and ρ value for each channel to define boundaries between different target sample populations. Furthermore, the analysis unit 203 can also control the fluid system 2 to perform a sorting process to classify and allocate target sample populations based on their boundaries.

[0056] 7 and 8 are graphs showing examples of dot plots generated and output during data analysis by the analysis unit 203. FIG. 7 shows a dot plot plotting the relationship between the virtual photon number of the fluorescence channel corresponding to antibody A and the virtual photon number of the fluorescence channel corresponding to antibody B, with the boundary defined by the gating process indicated by a solid line. Thus, the gating process calculates and outputs the proportions of populations within the entire analyte to be detected, with the proportion of populations negative for both antibody A and antibody B being "36.6%," the proportion of populations positive for only antibody B being "34.7%, the proportion of populations positive for only antibody A being "27.7%, and the proportion of populations positive for both antibodies A and B being "1.05%." FIG. 8 shows a dot plot plotting the relationship between the virtual photon number of the fluorescence channel using the fluorescent dye Cy5 and the virtual photon number of the fluorescence channel using the fluorescent dye TR. In this way, when defining the boundaries of a population by gating processing, it is possible to automatically define the boundaries of the range W of the gate section, which is determined by the ρ value corresponding to the average value of the virtual photon number of the population and is centered on the average value of the virtual photon number of the population. The range W determined by the ρ value is set to a range calculated by, for example, (average intensity value)±3×ρ×(average intensity value).

[0057] 9 and 10, a procedure for processing the output signal of the photomultiplier tube using the flow cytometer system 1 will be described. Fig. 9 shows the parameter calculation process for each channel by the flow cytometer system 1, and Fig. 10 shows the analysis process by flow cytometry of a sample fluid.

[0058] 9 , the signal acquisition unit 201 of the data processing device 12 acquires time-series digital values ​​DN for each channel of a plurality of test samples contained in a sample fluid (step S101). Then, the calculation unit 202 of the data processing device 12 generates a histogram based on the time-series digital values ​​DN for one channel (step S102). The calculation unit 202 then determines a width PW, which is a fitting range for each peak of the histogram, and thereby determines parameters A and B for each peak (step S103).

[0059] Then, the calculation unit 202 calculates a fitting function p single The parameter τ of each peak is determined by fitting (x) (step S104). Furthermore, the calculation unit 202 calculates the ratio τ / A based on the two parameters A and τ of each peak (step S105). Then, the calculation unit 202 selects one approximation formula for calculating the ρ value from formulas (2) to (5) for each peak according to the calculated ratio τ / A (step S106).

[0060] Thereafter, the calculation unit 202 calculates the average intensity of each peak based on the average value E of each peak calculated using equation (6), and calculates the ρ value of each peak using the selected approximation equation, and the average intensity value and ρ value of each peak are stored in the data processing device 12 (step S107). The processes of steps S102 to S107 are repeated for the time-series digital values ​​of each channel (step S108), and the parameter calculation process is completed. However, the processes of steps S102 to S107 may be performed for only one channel.

[0061] 10, when flow cytometry detection of a sample fluid containing multiple target samples is initiated in the flow cytometer system 1, the data processor 12 acquires digital values ​​DN for each channel (step S201). Next, the data processor 12 generates a dot plot using the digital values ​​DN for each channel (step S202).

[0062] The data processing device 12 then performs gating on the generated dot plot using the calculated average intensity values ​​and ρ values ​​for each channel (step S203), and classifies the analyte population (target population) in the data shown in the dot plot based on the results of the gating (step S204).

[0063] Next, if the data processing device 12 is set to perform sorting (step S205; Yes), the data processing device 12 is controlled to perform sorting on the classified target population (S206). On the other hand, if the data processing device 12 is set not to perform sorting (step S205; No), the data processing device 12 performs data analysis processing on the classified target population, such as calculating the proportion of the population to the whole (S207).

[0064] The effects of the flow cytometer system 1 according to the embodiment described above will now be described.

[0065] In the flow cytometer system 1, a histogram is generated by chronologically aggregating the digital values ​​DN of the intensity signals output by the photomultiplier tubes 11a, 11b, 11c, and 11d when multiple test samples are used. The histogram is then fitted with an exponential function to calculate a ρ value representing the variability of each peak in the frequency distribution. Data analysis is then performed using the ρ value for the digital values ​​DN of the intensity signals output by the photomultiplier tubes 11a, 11b, 11c, and 11d when multiple target samples are used. This allows for accurate evaluation of the variability of the photomultiplier tube output data through simple calculations. As a result, the computational cost of data analysis can be reduced while maintaining the accuracy of the data analysis. In other words, the flow cytometer system 1 enables highly accurate analysis of populations of analytes that were previously difficult to analyze, enabling population analysis using practical computational processing. For example, data analysis can also be performed by inputting the ρ value acquired by the flow cytometer system 1 into a machine learning model.

[0066] Furthermore, in the flow cytometer system 1, the function expressed by equation (1) is used as the exponential function used in fitting. In this case, the variation in the data of the output values ​​of the photomultiplier tube can be evaluated with higher accuracy.

[0067] Furthermore, in the flow cytometer system 1, when fitting an exponential function to a histogram based on the digital values ​​DN of the intensity signals output from the photomultiplier tube, a stably fitted exponential function can be obtained by determining the parameters A, B, and τ in that order. As a result, the ρ value that matches the peak of the frequency distribution can be stably calculated.

[0068] In the flow cytometer system 1, the function p ｓｉｎｇｌｅAn approximation formula suitable for approximating the width of the peak in (x) can be selected from formulas (2) to (5) according to the two parameters A and τ determined by fitting, and the width can be calculated. Therefore, the ρ value that matches the multiple peaks included in the frequency distribution can be accurately evaluated by simple calculation.

[0069] Furthermore, the flow cytometer system 1 can reduce the calculation cost of the gating process and maintain the accuracy of the gating process.

[0070] Various embodiments of the present invention have been described above, but the present invention is not limited to the above embodiments, and may be modified or applied to other things within the scope that does not change the gist of the claims.

[0071] For example, the gating process in the dot plot is not limited to being automatically set by the electronic system 4, but may be set by input from an operator, or the automatically set process may be adjusted by input from an operator.

[0072] Furthermore, the photomultiplier tubes in the embodiments may be replaced with other types of photodetectors, such as an APD (avalanche photodiode) or a SiPM (silicon photomultiplier).

[0073] Furthermore, the calculation unit 202 of the data processing device 12 in the above embodiment may function to calculate the ρ value expressed by the above equation (5) using an approximate equation by polynomial expansion expressed by the following equations (7) and (8).

[0074] 11 is a graph showing the error between the ρ value calculated by the above formula (5) and the ρ value calculated by the above formulas (7) and (8). This graph shows the change in the values ​​of formulas (7) and (8) normalized by the value of formula (5) when the ratio τ / A of the two parameters A and τ is changed. The inventors of the present application have clarified that the value calculated by formula (8) has a small error in the range of τ / A≦predetermined value TH4 (≈0.5), and the value calculated by formula (7) has a small error in the range of τ / A≦predetermined value TH5 (≈0.2).

[0075] Using the above property, the calculation unit 202 may function as follows. That is, based on the values ​​of two parameters A and τ, which are fitting results for one peak, the calculation unit 202 selects and uses equation (7) to calculate the ρ value when τ / A≦predetermined value TH5. On the other hand, the calculation unit 202 selects and uses equation (8) to calculate the ρ value when predetermined value TH5<τ / A≦predetermined value TH4. In all other cases, the calculation unit 202 selects and uses equation (5) to calculate the ρ value.

[0076] The calculation unit 202 of the data processing device 12 of the above embodiment may also function as follows: In other words, based on the values ​​of two parameters A and τ, which are fitting results for one peak, when τ / A≦predetermined value TH3, instead of selecting and using equation (4) to calculate the ρ value, the average value E and standard deviation √V may be calculated using the approximation equations shown in the following equations (9) and (10).

[0077] 1...flow cytometer system, 2...fluidic system, 3...optical system, 4...electronic system (signal processing device), 5...channel, 6...flow cell, 7...laser light source, 8...lens, 9a, 9b, 9c, 9d...filter, 10b, 10c...dichroic mirror, 11a, 11b, 11c, 11d...photomultiplier tube (photodetector), 12...data processing device, 201...signal acquisition unit, 202...calculation unit, 203...analysis unit.

Claims

1. A signal processing method for processing the output of a photodetector constituting a flow cytometer, comprising the steps of: acquiring, in a time series, output current signals from the photodetector that detect signal light generated by flow cytometry using the flow cytometer for a plurality of test samples; generating a histogram by aggregating the frequency distribution of the intensity of the output current signals based on the time series values ​​of the output current signals; calculating a ρ value that represents the variability of each peak in the frequency distribution by fitting the histogram using an exponential function; and performing data analysis using the ρ value for the output current signals from the photodetector obtained by flow cytometry for a plurality of target samples.

2. The signal processing method according to claim 1, wherein the exponential function is a function that increases exponentially within a predetermined range of a variable x corresponding to the intensity, and becomes zero outside the predetermined range of the variable x.

3. The exponential function is expressed by the formula (1); [In the above formula, A is a parameter that determines the upper limit of the range of the distribution, B is a parameter that determines the lower limit of the range of the distribution, τ is a parameter that determines the time constant of the distribution, and u is a step function.] ｓｉｎｇｌｅ The signal processing method according to claim 2 , wherein (x) is satisfied.

4. When fitting the exponential function, parameters A and B are determined by defining the range of one peak in the frequency distribution, and then the function p ｓｉｎｇｌｅ The signal processing method according to claim 3 , wherein the parameter τ is determined by fitting (x) to the one peak.

5. A signal processing method according to any one of claims 1 to 4, wherein after fitting each peak of the histogram with the exponential function, the ρ value of each peak is calculated by selecting from a plurality of approximation formulas that approximate the width of the peak of the exponential function and performing calculations.

6. Each peak of the histogram is calculated by the function p ｓｉｎｇｌｅ After fitting with (x), the function p ｓｉｎｇｌｅ 5. The signal processing method according to claim 3, wherein the ρ value of each peak is calculated by selecting one approximation formula from a plurality of approximation formulas that approximate the width of the peak of (x) in accordance with the parameters A and τ.

7. The plurality of approximate expressions include approximate expressions represented by equations (2) to (4).

7. The signal processing method according to claim 6.

8. A signal processing method according to any one of claims 1 to 7, wherein the data analysis includes a gating process for defining boundaries of a population to be analyzed.

9. A signal processing device that processes the output of a photodetector that constitutes a flow cytometer and includes a processor, wherein the processor is configured to: acquire, in a time series, output current signals from the photodetector that detect signal light generated by flow cytometry using the flow cytometer on a plurality of test samples; generate a histogram by aggregating the frequency distribution of the intensity of the output current signals based on the time series values ​​of the output current signals; calculate a ρ value that represents the variability of each peak in the frequency distribution by fitting the histogram using an exponential function; and perform data analysis using the ρ value on the output current signals from the photodetector obtained by flow cytometry on a plurality of target samples.

10. The signal processing device according to claim 9, wherein the exponential function is a function that increases exponentially within a predetermined range of a variable x corresponding to the intensity, and becomes zero outside the predetermined range of the variable x.

11. The exponential function is expressed by the formula (1); [In the above formula, A is a parameter that determines the upper limit of the range of the distribution, B is a parameter that determines the lower limit of the range of the distribution, τ is a parameter that determines the time constant of the distribution, and u is a step function.] ｓｉｎｇｌｅ The signal processing device according to claim 10, wherein (x) is satisfied.

12. When fitting the exponential function, the processor determines parameters A and B by defining the range of one peak in the frequency distribution, and then fits the function p ｓｉｎｇｌｅ The signal processing apparatus according to claim 11 , wherein the parameter τ is determined by fitting (x) to the one peak.

13. A signal processing device according to any one of claims 9 to 12, wherein the processor calculates the ρ value of each peak by fitting each peak of the histogram with the exponential function, and then selecting from a plurality of approximation formulas that approximate the width of the peak of the exponential function and performing calculations.

14. The processor calculates each peak of the histogram by the function p ｓｉｎｇｌｅ After fitting with (x), the function p ｓｉｎｇｌｅ 13. The signal processing device according to claim 11, wherein the ρ value of each peak is calculated by selecting one approximation formula from a plurality of approximation formulas that approximate the width of the peak of (x) in accordance with the parameters A and τ.

15. The plurality of approximate expressions include approximate expressions represented by equations (2) to (4). The signal processing device according to claim 14.

16. A signal processing device according to any one of claims 9 to 15, wherein the data analysis includes a gating process for defining boundaries of a population to be analyzed.

17. A signal processing system comprising: a signal processing device according to any one of claims 9 to 16; the photodetector; and an optical system that guides the signal light to the photodetector.

Images