Information processing apparatus, information processing method, program, and information processing system

Abstract

The present invention can analyze data that increases in size due to the use of multiple colors. An information processing apparatus according to one embodiment of the present invention is provided with a dimension compression unit (214) that performs dimension compression on each of a plurality of sets of spectral data including a fluorescent component radiated from respective particles labeled with one or more fluorescent dyes, an initial value determination unit (214) that determines an initial value of each of a plurality of nodes based on a result of the dimension compression, and a clustering unit (214) that performs clustering on the respective sets of spectral data using the initial values.

Description

TECHNICAL FIELD

[0001] The present disclosure relates to an information processing apparatus, an information processing method, a program, and an information processing system. BACKGROUND

[0002] In the fields of medicine, biochemistry, and the like, a flow cytometer is often used to quickly measure a property of each of a large number of particles. A flow cytometer is an apparatus that irradiates a particle such as a cell or a bead flowing in a flow cell with a light beam to detect fluorescence, scattered light, or the like emitted from the particle, and optically measures a property of each particle.

[0003] LIST OF CITATIONS

[0004] PATENT LITERATURE

[0005] Patent Literature 1: JP 2007-132921 A

[0006] Patent Literature 2: JP 2016-511397 A SUMMARY

[0007] TECHNICAL PROBLEM

[0008] In recent flow cytometers, multi-colors in which a particle such as a cell is subjected to multiple fluorescent dye staining have been developed. However, with the development of multi-colors, the number of fluorescent substances that can be measured at a time increases, and a combinatorial explosion occurs. Due to this, there is a problem that the amount of data to be processed increases and analysis becomes difficult.

[0009] Therefore, the present disclosure proposes an information processing apparatus, an information processing method, a program, and an information processing system that can analyze data that increases by multi-colors.

[0010] SOLUTION TO PROBLEM

[0011] To solve the above problem, an information processing apparatus according to one aspect of the present disclosure includes a dimension compression unit that performs dimension compression on each piece of spectral data of a plurality of pieces of spectral data, the spectral data including a fluorescent component emitted from each of a plurality of particles labeled with one or more fluorescent dyes; an initial value determination unit that determines an initial value for each of a plurality of nodes based on a result of the dimension compression; and a clustering unit that performs clustering on the plurality of pieces of spectral data using the initial value.

[0012] Note that the above effects are not necessarily limited, and any effect described in this specification or other effects that can be grasped from this specification can be exhibited together with or in place of the above effects. BRIEF DESCRIPTION OF DRAWINGS

[0013] Figure 1 is a block diagram illustrating a configuration of an information processing system according to a first embodiment.

[0014] Figure 2 FIG. 1 is a graph showing an example of a fluorescence spectrum.

[0015] Figure 3 FIG. 2 is a graph showing an example of a pretreatment parameter table.

[0016] Figure 4 FIG. 3 is a graph showing an example of a spectrum graph.

[0017] Figure 5 FIG. 4 is a graph showing an example of a clustering result.

[0018] Figure 6 FIG. 5 is a graph explaining a fluorescence separation process.

[0019] Figure 7 FIG. 6 is a graph showing a display example of an analysis result.

[0020] Figure 8 FIG. 7 is a flowchart showing a pretreatment flow of a pretreatment unit.

[0021] Figure 9 FIG. 8 is a flowchart showing a processing flow of a clustering processing unit.

[0022] Figure 10 FIG. 9 is a flowchart showing a flow of a fluorescence separation process of a fluorescence separation unit.

[0023] Figure 11 FIG. 10 is a schematic diagram for explaining a series of flows from initialization to learning in FlowSOM.

[0024] Figure 12 FIG. 11 is a block diagram showing a configuration of an information processing system according to a second embodiment.

[0025] Figure 13 FIG. 12 is a flowchart showing an operation example of a pretreatment / fluorescence separation unit according to the second embodiment.

[0026] Figure 14 FIG. 13 is a flowchart showing an example of a clustering process performed by a clustering processing unit according to the second embodiment.

[0027] Figure 15 FIG. 14 is a graph for explaining an operation at the time of generating a representative node vector of each representative node described in step S222 of Figure 14 FIG. 15 is a graph for explaining an operation at the time of performing clustering by batch learning described in step S223 of

[0028] Figure 16 FIG. 16 is a flowchart showing an example of clustering by batch learning according to the present embodiment described in step S223 of Figure 14 FIG. 17 is a flowchart showing an example of clustering by batch learning according to the present embodiment described in step S223 of

[0029] Figure 17 is for explainingFigure 14 a diagram of an operation of determining the number of clusters using consensus clustering described in step S224 of FIG. 8.

[0030] Figure 18 is a block diagram showing a hardware configuration example of the information processing device. DETAILED DESCRIPTION

[0031] Hereinafter, embodiments of the present disclosure will be described in detail with reference to the accompanying drawings. In addition, the present disclosure will be described in the following item order.

[0032] 1. First Embodiment

[0033] 1.1 Configuration of Information Processing System

[0034] 1.2 Operation of Information Processing Device

[0035] 1.3 Effects

[0036] 2. Second Embodiment

[0037] 2.1 Configuration of Information Processing System

[0038] 2.2 Operation Example of Information Processing Device

[0039] 2.2.1 Operation Example of Preprocessing / Fluorescence Separation Unit

[0040] 2.2.2 Operation Example of Cluster Processing Unit

[0041] 2.2.3 Initialization of Representative Node Vector

[0042] 2.2.4 Clustering by Batch Learning

[0043] 2.2.5 Determining Number of Clusters Using Consensus Clustering

[0044] 2.3 Effects

[0045] 3. Hardware Configuration of Information Processing Device

[0046] 1. First Embodiment

[0047] Recent flow cytometers can acquire more abundant information by staining particles such as cells with a plurality of fluorescent dyes and measuring a plurality of fluorescent signals at once. At the same time, due to such multicolor property of flow cytometers, limitations of analysis by manual gating in the related art are pointed out. For example, when n kinds of fluorescent dyes are plotted in a two-dimensional coordinate system, there are nC2 kinds of combination patterns of fluorescent dyes. That is, there are 15 kinds of combination patterns for six colors, and 190 kinds of combination patterns for 20 colors.

[0048] As a method of analyzing such increased data, it is considered to adopt automatic clustering instead of a conventional method of manually setting gates, classifying data by clustering, and then analyzing the data.

[0049] However, in general clustering, a certain piece of data is classified into any cluster. Therefore, even data located at a boundary between two or more clusters and difficult to determine is classified into any cluster. Therefore, when data acquired by a flow cytometer is classified by clustering, a certain degree of misclassification can occur.

[0050] Meanwhile, there are filter-type flow cytometers that detect the intensity of one or more specific wavelengths of fluorescence and spectral-type flow cytometers that can acquire more information by detecting fluorescence emitted from each fluorescent dye as a wavelength spectrum. In the spectral-type flow cytometer, since the intensity of fluorescence of each wavelength is obtained, a large number of features can be obtained from one particle (also referred to as a sample). Therefore, by adopting the spectral-type flow cytometer, more detailed analysis can be performed.

[0051] However, in spectral data obtained by the spectral-type flow cytometer, the intensity of fluorescence of each wavelength exponentially changes rather than linearly. Therefore, in order to display spectral data that is easily recognized by a user, scaling conversion of the spectral data is required. Meanwhile, in clustering, classification is performed based on the distance between pieces of data. Therefore, if the spectral data in which the intensity of fluorescence of each wavelength exponentially changes is clustered as it is, the difference of the part with a large intensity of fluorescence greatly influences the overall distance between pieces of data, and misclassification can occur. As a result, the clustering result can not correctly correspond to the spectral data to be scaled and displayed.

[0052] Therefore, in the present embodiment, an information processing apparatus, an information processing method, a program, and an information processing system that can make the display result of spectral data and the clustering result more correctly correspond to each other will be described as an example.

[0053] Specifically, as a preprocessing of spectral data acquired by a flow cytometer, a log conversion is performed. Then, the spectral data subjected to the log conversion is clustered, and the clustering result obtained by the clustering is displayed to a user. As a result, in the present embodiment, it can be determined whether or not the part with a large value in the spectral data has been clustered. Therefore, clustering can be performed so that the display result of the spectral data and the clustering result more correctly correspond to each other. Hereinafter, an information processing apparatus, an information processing method, a program, and an information processing system according to the present embodiment will be described in detail with reference to the drawings.

[0054] However, the technology according to the present disclosure does not need to solve all the above problems at the same time. Therefore, it should be understood that those which solve some or all of the above problems by including some or all of the configurations described below are included in the technical scope of the technology according to the present disclosure.

[0055] 1.1 Configuration of information processing system

[0056] Figure 1 is a block diagram illustrating a configuration of an information processing system according to the present embodiment. As shown in Figure 1 , the information processing system 4 includes the information processing device 1 and a measurement device 3.

[0057] The measurement device 3 is a measurement device capable of detecting a fluorescence spectrum of each color from a cell or the like to be measured. The measurement device 3 is, for example, a flow cytometer. The measurement sample measured by the flow cytometer can be a biologically derived particle such as a cell, a microorganism, or a biologically related particle. For example, the cell can be an animal cell (for example, a cell based on a small body) or a plant cell. For example, the microorganism can be a bacterium such as Escherichia coli, a virus such as tobacco mosaic virus, a fungus such as yeast. The biologically related particle can be a particle constituting a cell such as a chromosome, a liposome, a mitochondrion, or various organelles. Note that the biologically related particle can include a biologically related polymer such as a nucleic acid, a protein, a lipid, a sugar chain, or a complex thereof. Each of these biologically derived particles can have a spherical shape or a non-spherical shape, and is not particularly limited in size and mass.

[0058] In addition, the measurement sample can be an industrially synthesized particle such as a latex particle, a gel particle, or an industrial particle. For example, the industrially synthesized particle can be a particle synthesized from an organic resin material such as polystyrene or polymethyl methacrylate, an inorganic material such as glass, silica, or a magnetic body, or a metal such as colloidal gold or aluminum. Similarly, each of these industrially synthesized particles can have a spherical shape or a non-spherical shape, and is not particularly limited in size and mass.

[0059] Before measuring the fluorescence spectrum, the measurement sample can be labeled (stained) with one or more fluorescent dyes. The measurement sample can be labeled with the fluorescent dye by a known method. Specifically, when the measurement sample is a cell, a fluorescently labeled antibody that selectively binds to an antigen present on the surface of the cell is mixed with the cell to be measured, and the fluorescently labeled antibody binds to the antigen on the surface of the cell. Thereby, the cell to be measured can be labeled with the fluorescent dye. Alternatively, the cell to be measured can be labeled with the fluorescent dye by mixing a fluorescent dye that is selectively taken up into a specific cell with the cell to be measured.

[0060] The fluorescently labeled antibody is an antibody to which a fluorescent dye as a label is bound. The fluorescently labeled antibody can be an antibody directly bound with a fluorescent dye. Alternatively, the fluorescently labeled antibody can be an antibody obtained by binding a fluorescent dye bound with avidin to an antibody labeled with biotin through an avidin-biotin reaction. Note that, as the antibody, a polyclonal antibody or a monoclonal antibody can be used.

[0061] The fluorescent dye used for labeling the cells is not particularly limited, and at least one or more kinds of known dyes used for staining cells and the like can be used. For example, as the fluorescent dye, phycoerythrin (PE), fluorescein isothiocyanate (FITC), PE-Cy5, PE-Cy7, PE-Texas Red (registered trademark), allophycocyanin (APC), APC-Cy7, ethidium bromide, propidium iodide, Hoechst (registered trademark) 33258, Hoechst (registered trademark) 33342, 4',6-diamidino-2-phenylindole (DAPI), acridine orange, chromomycin, mithramycin, olivomycin, pyronine Y, thiazole orange, rhodamine 101, isothiocyanate, BCECF, BCECF-AM, C.SNARF-1, C.SNARF-1-AMA, aequorin, Indo-1, Indo-1-AM, Fluo-3, Fluo-3-AM, Fura-2, Fura-2-AM, oxonol, Texas Red (registered trademark), rhodamine 123, 10-N-nonyl acridine orange, fluorescein, fluorescein diacetate, carboxyfluorescein, carboxyfluorescein diacetate, carboxy dichlorofluorescein, and carboxy dichlorofluorescein diacetate can be used. In addition, derivatives and the like of the above fluorescent dyes can also be used.

[0062] The flow cytometer includes a laser light source that emits laser light having a wavelength capable of exciting a fluorescent dye used to label a measurement sample S, a flow cell through which the measurement sample S flows in one direction, and a photodetector that receives one or more of fluorescence, phosphorescence, and scattered light from the measurement sample S irradiated with the laser light.

[0063] The laser light source is, for example, a semiconductor laser light source that emits laser light having a predetermined wavelength. A plurality of laser light sources can be provided. When a plurality of laser light sources are provided, the positions at which laser light from the laser light sources is irradiated can be the same as or different from each other in the flow cell. However, in a case where different positions are irradiated with laser light from a plurality of laser light sources, light from the measurement sample S can be detected by different photodetectors. In such a case, even in a case where a dye that emits fluorescence having a wavelength close to that of the laser light is used, it is possible to measure the fluorescence spectrum of each kind of fluorescence without color mixing. Note that the laser light emitted from the laser light source can be pulsed light or continuous light. For example, as the laser light source, a plurality of semiconductor laser light sources that emit laser light having a wavelength of 480 nm and laser light having a wavelength of 640 nm can be used.

[0064] The flow cell is a flow path through which a plurality of measurement samples S flow in a straight line in one direction. Specifically, through the flow cell, sheath liquid that surrounds the measurement sample S flows at high speed in the form of laminar flow, so that a plurality of measurement samples S flow in a straight line in one direction. The flow cell can be formed in a microchip or a cuvette.

[0065] The photodetector detects light from the measurement sample S irradiated with laser light through photoelectric conversion. The light from the measurement sample S can include at least one of fluorescence, phosphorescence, and scattered light.

[0066] For example, the photodetector can include a detector that detects scattered light LS including forward scattered light and side scattered light from the measurement sample S, and an array of light-receiving elements that detects fluorescence from the measurement sample S.

[0067] The detector can be, for example, a known photoelectric conversion element such as a charge-coupled device (CCD), a complementary metal-oxide semiconductor (CMOS), or a photodiode. The array of light-receiving elements can be configured, for example, by arranging a plurality of independent detection channels having different detected light wavelength ranges. Specifically, the array of light-receiving elements can be an array of light-receiving elements in which a plurality of photomultiplier tubes (PMTs) or photodiodes having different detected wavelength ranges are arranged in one dimension, an image sensor in which pixels are arranged in a two-dimensional lattice pattern, or the like. The array of light-receiving elements photoelectrically converts fluorescence from the measurement sample S that is spectrally split by a spectral splitting element such as a prism or a grating.

[0068] In the flow cytometer having the above-described configuration, first, the measurement sample S flowing in the flow cell is irradiated with laser light emitted from the laser light source. The measurement sample S emits scattered light and fluorescence (or phosphorescence) by being irradiated with the laser light. The detector detects the scattered light emitted from the measurement sample S. At the same time, the fluorescence emitted from the measurement sample S is spectrally split into light of each wavelength by the spectral splitting element, and then received by the array of light-receiving elements. Thereby, the spectrum of the fluorescence emitted from the measurement sample S is detected.

[0069] Note that in the following description, it is assumed that cells are measured. The measuring apparatus 3 is, for example, a spectral flow cytometer that makes fluorescently stained cells flow at high speed through a flow cell and irradiates the flowing cells with a light beam to detect the fluorescence spectrum of each fluorescent dye emitted from the cells. Figure 2 is a graph that represents an example of a fluorescence spectrum. As shown in Figure 2 , the fluorescence spectrum is represented by the fluorescence intensity of each channel corresponding to the wavelength.

[0070] The measuring apparatus 3 outputs the detected fluorescence spectrum as measurement data 2. The measurement data 2 includes the fluorescence spectrum data of each cell. The measuring apparatus 3 transmits the measurement data 2 to, for example, the information processing apparatus 1.

[0071] The information processing apparatus 1 acquires and analyzes the measurement data 2 measured by the measuring apparatus 3, and displays the analysis result. Note that the information processing apparatus 1 and the measuring apparatus 3 can be connected to each other via a network, and the information processing apparatus 1 can acquire the measurement data 2 via the network.

[0072] The information processing apparatus 1 includes a pre-processing unit 11, a pre-processing parameter table 12, a spectrum output unit 13, a clustering processing unit 14, a clustering result presentation unit 15, a fluorescence separation unit 16, and a conventional analysis presentation unit 17. Note that all or some of these functional units can be in the cloud. For example, the pre-processing parameter table 12, the clustering processing unit 14, and the fluorescence separation unit 16 can be in the cloud. In this case, the measurement data 2 is also transmitted to the cloud.

[0073] The pre-processing unit 11 pre-processes the spectrum data according to the selected parameters. Here, the pre-processing is a coordinate conversion for display from the actual observation values. The conversion can be, for example, a simple log10 conversion or a logical conversion in consideration of the characteristics of the observation apparatus. The parameters include W, T, M, and A.

[0074] W is a value that linearly displays values close to zero. T is the maximum value of the fluorescence intensity, for example, 104. M is the maximum value of the converted display coordinates. A is the minimum negative value to be converted.

[0075] The pre-processing parameter table 12 is a table that stores the pre-processing parameters. Figure 3 is a graph that shows an example of the pre-processing parameter table 12. As shown in Figure 3 , the pre-processing parameter table 12 stores a plurality of combinations (also referred to as parameter sets) of the values of W, T, M, and A. The parameter identifier is an identifier that identifies the parameter combination.

[0076] The preprocessing unit 11 performs a logical conversion using a combination of W, T, M, and A selected by the user from the preprocessing parameter table 12. Note that the preprocessing unit 11 can use default values instead of selecting parameters from the preprocessing parameter table 12. Alternatively, the user can specify values other than the values stored in the preprocessing parameter table 12. The preprocessing unit 11 performs preprocessing each time the user changes the preprocessing parameters. Note that the preprocessing unit 11 can perform a conversion of nonlinear processing, such as a logarithmic conversion or a double exponential conversion, instead of a logical conversion.

[0077] The spectral output unit 13 generates an image of a spectral map using the spectral data that has been preprocessed by the preprocessing unit 11, and displays the generated image. Figure 4 is a graph showing an example of a spectral map. The spectral map represents a detection wavelength (wavelength) on the horizontal axis, a fluorescence intensity on the vertical axis, and information (group information) on the number of fine particles (event number or density) by different color shades, color tones, and the like.

[0078] In Figure 4 , "LD 488" on the vertical axis indicates fluorescence when laser light with a wavelength of 488 nm (nanometers) is emitted, and "_A" indicates that the measurement value is an integrated intensity. Furthermore, in Figure 4 , information on the number of fine particles is indicated by a shade, but in an actual screen, information on the number of fine particles is indicated by color.

[0079] As Figure 4 indicated by the vertical axis in the spectral map, the vertical axis corresponding to the fluorescence intensity is subjected to a logical conversion and displayed. To this end, in Figure 4 , the arrow 91 and the arrow 92 are displayed in the same length, but the ranges indicated by the arrows are greatly different from each other. That is, when the vertical axis is a straight line, the lengths are completely different, and the length of the arrow 91 is much larger than the length of the arrow 92.

[0080] The clustering processing unit 14 clusters cells using the spectral data preprocessed by the preprocessing unit 11. For example, specifying K as K-means, the clustering processing unit 14 classifies the spectral data into K clusters. Alternatively, the clustering processing unit 14 can also automatically determine the number of divisions as in Flow Self-Organizing Map (FlowSOM).

[0081] Alternatively, for example, as in T-SNE, the clustering processing unit 14 can perform dimension compression, and gate the result of the dimension compression to perform clustering. Alternatively, the clustering processing unit 14 can perform two-stage clustering such as meta-clustering, and use two cluster definitions such as a cluster identifier and a meta-cluster identifier. Here, a meta-cluster is a collection of clusters.

[0082] By using the spectrum data subjected to the logical conversion for clustering, the clustering processing unit 14 can easily make the display result of the spectrum data correspond to the clustering result.

[0083] The clustering result presentation unit 15 displays the clustering result of the clustering processing unit 14 on a display device. The clustering result presentation unit 15 visualizes the number of classifications or which cell group belongs to which group. Figure 5 is a graph showing an example of the clustering result. Figure 5 A case where clustering is performed using FlowSOM is shown. In Figure 5 , the circles represent clusters, and the clusters are classified into meta-clusters M#1 to M#5 having different shades. In actual display, the meta-clusters M#1 to M#5 are displayed in different colors. By the display of the clustering result, the user can be informed of the distribution of the clusters and the meta-clusters.

[0084] The fluorescence separation unit 16 acquires the measurement data 2 and performs fluorescence separation processing (also referred to as deconvolution). Figure 6 is a graph explaining the fluorescence separation processing. Figure 6 (a) shows the measured fluorescence spectrum. As Figure 6 (b) shows, the measured fluorescence spectrum is, for example, a superposition of the fluorescence spectra of three fluorescences.

[0085] Therefore, the fluorescence separation unit 16 separates the spectrum into three spectra of fluorescences #1 to #3 using Figure 6 (c) shows the reference spectrum. Here, the reference spectrum is the fluorescence spectrum of each fluorescence. The separated spectra of the fluorescences #1 to #3 are shown in Figure 6 (d). The fluorescence separation unit 16, for example, calculates the fluorescence intensity by taking a weighted average using the spectrum separated for each fluorescence. Figure 6 (e) shows the intensities of the fluorescences #1 to #3 calculated by the fluorescence separation unit 16.

[0086] The conventional analysis presentation unit 17 performs analysis using the fluorescence intensities separated by the fluorescence separation unit 16, and displays the analysis result on a display device. Figure 7 is a graph showing an example of the display of the analysis result. Figure 7 shows a two-dimensional graph having two axes, APC-Cy7::CD 24 and PE-Dazzle 594::CD 38. Here, APC-Cy7::CD 24 and PE-Dazzle 594::CD 38 are fluorescence dye-labeled antibodies for measuring fluorescence intensity. “APC-Cy7” and “PE-Dazzle 594” are fluorescence dyes, and “CD 24” and “CD 38” are antibodies. By the two-dimensional graph, the user can be informed of the cell distribution with respect to the two fluorescence dyes.

[0087] Note that the information processing apparatus 1 can perform pre-processing and clustering on the cell groups included in the partial region of the two-dimensional map in response to the user's selection of the partial region, and display the clustering result. Further, the information processing apparatus 1 can perform pre-processing and clustering on the cell groups included in the partial region of the spectral map in response to the user's selection of the partial region, and display the clustering result. In particular, a feature portion such as a dense portion in the spectral map is often selected by the user.

[0088] 1.2 Operation of the information processing apparatus

[0089] Next, the operation of the information processing apparatus 1 will be described with reference to Figure 8 to Figure 10 FIG. 8. Figure 8 is a flowchart illustrating a pre-processing flow of the pre-processing unit 11. As illustrated in Figure 8 , the pre-processing unit 11 selects pre-processing parameters in accordance with the user's instruction (step S1).

[0090] Then, the pre-processing unit 11 pre-processes the spectral data using the pre-processing parameters (step S2). Then, the pre-processing unit 11 determines whether the user has changed the pre-processing parameters (step S3). If the user has changed the pre-processing parameters, the pre-processing unit 11 changes the pre-processing parameters and pre-processes the spectral data (step S4). The process returns to step S3.

[0091] Meanwhile, if the user has not changed the pre-processing parameters, the pre-processing unit 11 determines whether all of the target cells have been processed (step S5). If there is a target cell that has not been processed, the process returns to step S2, and another cell is processed. Meanwhile, if all of the target cells have been processed, the pre-processing unit 11 instructs the spectral output unit 13 to present the pre-processed spectral data to the user (step S6). Further, the pre-processing unit 11 transfers the pre-processed spectral data to the clustering processing unit 14.

[0092] As described above, the pre-processing unit 11 pre-processes the spectral data, so that the clustering processing unit 14 can make the clustering result correspond to the display result of the spectral data.

[0093] Figure 9 is a flowchart illustrating a processing flow of the clustering processing unit 14. As illustrated in Figure 9 , the clustering processing unit 14 acquires the pre-processed spectral data from the pre-processing unit 11 (step S11), and performs clustering processing (step S12). Then, the clustering processing unit 14 instructs the clustering result presentation unit 15 to present the clustering result (step S13).

[0094] As described above, since the clustering processing unit 14 performs clustering using the pre-processed spectral data, the clustering result can be made to correspond to the display result of the spectral data. ​

[0095] Figure 10 is a flowchart showing a flow of the fluorescence separation process by the fluorescence separation unit 16. As shown in Figure 10 the fluorescence separation unit 16 acquires the reference spectrum (step S21). Then, the fluorescence separation unit 16 acquires the spectrum data of one cell (step S22), and performs the fluorescence separation process (step S23). Then, the fluorescence separation unit 16 determines whether all the cells have been processed (step S24). If there is a cell that has not been processed, the process returns to step S22. If all the cells are processed, the fluorescence separation process ends.

[0096] As described above, since the fluorescence separation unit 16 performs the fluorescence separation process, the conventional analysis presentation unit 17 can analyze the fluorescence data and display the analysis result.

[0097] 1.3 Action and Effect

[0098] As described above, according to the embodiment, the preprocessing unit 11 acquires the spectrum data and performs the logical conversion as the preprocessing. Then, the clustering processing unit 14 performs the clustering using the spectrum data subjected to the logical conversion. Then, the clustering result presentation unit 15 displays the clustering result on the display device. Thus, the information processing apparatus 1 can prevent the clustering of the portion having a large value in the spectrum data. Therefore, the information processing apparatus 1 can perform the clustering so that the display result of the spectrum data and the clustering result correspond to each other.

[0099] 2. Second Embodiment

[0100] As described above, recent flow cytometers have become multispectral in which particles such as cells are stained with a plurality of fluorescent dyes.

[0101] With the multispectralization of the flow cytometer, the number of fluorescent substances that can be measured at a time increases, whereby a combinatorial explosion occurs. Therefore, it is difficult to perform the analysis by hand. For example, in a case where N (N is a natural number) colors are measured at a time, when two-dimensional processing is performed for each color, there are n (n-1) / 2 combinations, and the number of combinations increases in proportion to half of the square of the number of colors N.

[0102] As a data analysis method that increases by such a combinatorial explosion, for example, a method of classifying data by clustering such as FlowSOM and then analyzing the data is considered as described above.

[0103] Figure 11 is a schematic diagram for explaining a series of flows from initialization to learning in FlowSOM. Figure 11The illustrated example represents a case where 100 vectors (coordinate values) included in the data group 900 to be analyzed are classified as being in an initial state at nine representative nodes N11 to N33 arranged in three rows and three columns in a two-dimensional coordinate system.

[0104] In Figure 11 In the illustrated example, first, a predetermined number (nine in this example) of data segments are selected from the data group 900 by random sampling, and the selected data segments are set as initial node vectors N_ij of the representative nodes N11 to N33 (step S1).

[0105] Next, learning is performed to update the representative node vector N_ij of each of the representative nodes N11 to N33. Specifically, first, one piece of data is acquired from the data group 900 by random sampling (step S2).

[0106] Subsequently, the representative node closest to the acquired data (vector) is subjected to a process represented by the following equation (1), thereby updating the representative node vector N_ij of the representative node (step S3). Note that in equation (1), a represents a learning rate. The learning rate a can be set based on an empirical rule or the like, and can be a value such as 0.1 or 0.01, for example.

[0107] New representative node vector N_ij = (1 - a) x old representative node vector N_ij + a x vector of acquired data (1)

[0108] Subsequently, the representative nodes (hereinafter referred to as peripheral representative nodes) around the representative node having the representative node vector N_ij updated in step S3 are subjected to a process further multiplying the amount of change by equation (1) multiplied by β, represented by the following equation (2), thereby updating the representative node vector (hereinafter referred to as a peripheral representative node vector) N_ij of the peripheral representative nodes (step S4). Note that in equation (2), β(r) can be a parameter of an algorithm determined according to an empirical rule, and can be a value such as 0.1 or 0.01, for example. Further, β(r) can be a fixed value, or can vary depending on the number of times of learning. For example, the β(r) value can be changed between the first learning and the second learning. The same applies to a.

[0109] New peripheral representative node vector N_ij = (1 - β(r)) x old peripheral representative node vector + β(r) x vector of acquired data (2)

[0110] Note that in formula (2), r can be a distance from the node to be updated, or can be a variable in which β is weighted with a function related to the distance. For example, β(r) = 0.9 can be set for a node adjacent to the node to be updated, and β(r) = 0.5 can be set for a node adjacent to the node adjacent to the node to be updated. As described above, β(r) can change depending on the number of times of learning, for example, the value of β can decrease as the number of times of learning increases.

[0111] Thereafter, the processing in steps S2 to S4 is repeated until the processing of all data included in the data group 900 is completed.

[0112] However, in such a FlowSOM, since the SOM algorithm is used, the following problems exist.

[0113] First, when initializing representative nodes (hereinafter, also referred to as SOM nodes) of the SOM, as described above, since the initial node vectors are determined by random sampling of data (step Sl), there is a problem that the clustering result varies depending on the execution situation.

[0114] Second, in the SOM algorithm, since clustering is performed while learning is performed by random sampling (steps S2 to S4), there is a problem that the learning result varies depending on the sampling order, and thus the clustering result varies depending on the execution situation.

[0115] Third, as a derivative of the second problem, since learning is performed by the random sampling order (steps S2 to S4), the processing is indivisible processing in which the update of the vectors of the SOM nodes cannot be parallelized (divided). Thus, there is a problem that the processing efficiency or processing speed cannot be improved.

[0116] Further, although the multispectral flow cytometer makes more detailed analysis possible, the amount of data to be processed increases. Thus, there is also a problem that when data is classified by clustering, an increase in processing time or a dimension disaster easily occurs.

[0117] Further, in recent years, since the measurement speed of the flow cytometer is improved, the number of data to be analyzed increases. However, when the number of data increases, not only does the problem of an increase in plotting time in data analysis exist, but also the following problem exists: when additionally analyzing a cluster generated by a user, in addition to an increase in plotting time, since the number of target clusters simply increases, the analysis time increases.

[0118] Therefore, in the present embodiment, an information processing apparatus, an information processing method, a program, and an information processing system capable of efficiently performing clustering capable of suppressing variation depending on the execution situation in the result will be described by an example.

[0119] However, the technology according to the present disclosure does not need to solve all the above problems at the same time. Therefore, it should be understood that those which solve some or all of the above problems by including some or all of the configurations described below are included in the technical scope of the technology according to the present disclosure.

[0120] Further, in the following description, the same reference signs are given to configurations and operations similar to those of the configuration and operation of the above-described first embodiment, and are not described here again.

[0121] 2.1 Configuration of information processing system

[0122] Figure 12 is a block diagram showing a configuration of an information processing system according to the present embodiment. As shown in Figure 12 , the information processing system 200 includes an information processing device 201 and a measurement device 3. In the present embodiment, the measurement device 3 and the measurement sample can be similar to those described in the first embodiment, and are not described here again.

[0123] The information processing device 201 includes a pre-processing / fluorescence separation unit 211, a pre-processing parameter table 12, a spectral output unit 13, a clustering processing unit 214, a clustering result presentation unit 15, and a routine analysis presentation unit 17. In this configuration, the pre-processing parameter table 12, the clustering result presentation unit 15, and the routine analysis presentation unit 17 can be similar to those according to the first embodiment. Further, the pre-processing / fluorescence separation unit 211 can have the functions of both the pre-processing unit 11 and the fluorescence separation unit 16 according to the first embodiment. Note that, Figure 12 the spectral output unit 13 is omitted in the present embodiment, but can be added similar to the spectral output unit 13 of the first embodiment.

[0124] Similar to the fluorescence separation unit 16 according to the first embodiment, the pre-processing / fluorescence separation unit 211 separates the measurement data 2 as spectral data into the fluorescence spectrum of each fluorescent dye by using the reference spectrum. Then, the pre-processing / fluorescence separation unit 211 performs pre-processing such as logical conversion on the unmixed fluorescence spectrum. The pre-processed fluorescence spectrum of each fluorescent dye is presented to the user by the routine analysis presentation unit 17.

[0125] Note that the parameter set of the conversion parameter used in the pre-processing can be selected for use from among the parameter sets managed in the pre-processing parameter table 12 (see Figure 3 ) by a parameter set specified by the user, as in the first embodiment. Further, the parameter set managed in the pre-processing parameter table 12 can be finely adjusted by the user.

[0126] The clustering processing unit 214 according to the present embodiment performs clustering processing on the preprocessed fluorescence spectrum output by the preprocessing / fluorescence separation unit 211, and the clustering result generated by the clustering processing unit 214 is presented to the user by the clustering result presentation unit 15.

[0127] 2.2 Operation example of information processing apparatus

[0128] Next, the operation of the information processing apparatus 201 according to the present embodiment will be described.

[0129] 2.2.1 Operation example of preprocessing / fluorescence separation unit

[0130] First, an operation example of the preprocessing / fluorescence separation unit 211 according to the present embodiment will be described. Figure 13 is a flowchart showing an operation example of the preprocessing / fluorescence separation unit according to the present embodiment. As shown in Figure 13 , in the present embodiment, the preprocessing / fluorescence separation unit 211 first selects preprocessing parameters according to an instruction from the user (step S201), as in Figure 8 step S1.

[0131] Next, similarly to the fluorescence separation processing described with reference to Figure 10 in the first embodiment, the preprocessing / fluorescence separation unit 211 performs fluorescence separation processing on the spectral data of all cells included in the measurement data 2 using the reference spectrum (step S202).

[0132] Next, as in steps S2 to S5 of Figure 8 , the preprocessing / fluorescence separation unit 211 performs preprocessing on the spectral data of all cells included in the measurement data 2 using the preprocessing parameters selected or changed by the user (steps S203 to S206).

[0133] Thereafter, if the preprocessing of the spectral data of all cells is completed (Yes in step S206), the preprocessing / fluorescence separation unit 211 inputs the preprocessed spectral data to the clustering processing unit 214 (step S207), and ends the present operation. Note that the preprocessing / fluorescence separation unit 211 can instruct the spectral output unit 13 to present the preprocessed spectral data to the user, for example, as in step S6 of Figure 8 .

[0134] Note that, in the present operation, the spectral data of all cells included in the measurement data 2 is the target, but the target of the present operation is not limited to this, and only the spectral data of cells equal to or more than the number of divisions of nodes in the clustering described below (hereinafter, also referred to as the necessary number) is required. The same applies to the operation of the clustering processing unit described below.

[0135] 2.2.2 Operation example of the clustering processing unit

[0136] Next, an operation example of the clustering processing unit 214 according to the present embodiment will be described. The basic operation flow performed by the clustering processing unit 214 according to the present embodiment can be similar to the operation described with reference to FIG. 12, for example. However, in the present embodiment, the clustering processing performed in step S12 of FIG. 12 is replaced with the clustering processing described later with reference to FIG. 13. Figure 9 Figure 9 The clustering processing performed in step S12 of FIG. 12 is replaced with the clustering processing described later with reference to FIG. 13. Figure 14

[0137] Figure 14 is a flowchart showing an example of the clustering processing performed by the clustering processing unit according to the present embodiment. Note that, as can be seen from the operation of the clustering processing unit 214 shown in Figure 14 , the clustering processing unit 214 according to the present embodiment can also function as one or more of a dimension compression unit, an initial value determination unit, a clustering unit, a mean value calculation unit, one or more assignment units, an update unit, a node number change unit, a meta-clustering unit, a division unit, and a node setting unit.

[0138] As shown in Figure 14 , in the present embodiment, the clustering processing unit 214 first performs principal component analysis on the spectral data of all cells (or a number of cells equal to or greater than a necessary number) included in the measurement data 2 (step S221).

[0139] Note that the target spectral data in step S221 can be the preprocessed spectral data after fluorescence separation. Furthermore, in the present specification, the number of principal components is two, but is not limited thereto, and can be three or more. Furthermore, in step S221, not only principal component analysis can be performed, but also various types of dimension compression using a statistical data analysis method such as T-SNE can be performed.

[0140] Next, the clustering processing unit 214 determines the initial value of the representative node vector of the SOM node based on the first principal component and the second principal component determined in step S221 and the values (vectors) of all cells (or a number of cells equal to or greater than a necessary number) included in the measurement data 2 (step S222). Note that details of step S222 will be described later with reference to Figure 15

[0141] Next, the clustering processing unit 214 performs SOM clustering processing by performing batch learning (step S223). Note that details of step S223 will be described later with reference to Figure 16

[0142] ​​​​Next, the cluster processing unit 214 performs a process of determining the number of clusters such as a consistency clustering based on the result of the SOM clustering process performed in step S223 (hereinafter, referred to as meta-clustering) (step S224). Thereby, the number of clusters in the clustering process is determined. Note that the details of step S223 in the case of using the consistency clustering will be described later with reference to Figure 17 Details of step S223 in the case of using the consistency clustering will be described.

[0143] Thereafter, the cluster processing unit 214 instructs the cluster result presentation unit 15 to present the cluster result to the user (step S225). For example, for the visualization of presenting the cluster result to the user, a minimum spanning tree (MST) method can be used.

[0144] 2.2.3 Initialization of representative node vector

[0145] Figure 15 is a diagram for explaining the operation in the initialization of the representative node vector of each representative node described in step S222 according to the present embodiment. Note that Figure 14 the example shown in Figure 15 shows the 100 pieces of processed spectral data (vectors (coordinate values)) on which the pre-processing / fluorescence separation unit 211 performs the fluorescence separation and the pre-processing are classified as nine representative nodes N11 to N33 arranged in three rows and three columns in a coordinate system in the initial state, similarly to Figure 11 shown in the example.

[0146] As shown in Figure 15 , in the initialization of the representative node vector according to the present embodiment, first, the cluster processing unit 214 calculates the average value 302 of each dimension with respect to the data group (hereinafter, simply referred to as data group) including the spectral data after the fluorescence separation and the pre-processing (hereinafter, referred to as processed spectral data) (step S301). For example, in the case where the dimension number of data is 10, for each of 1 to 10 dimensions, the values in each dimension of all the processed spectral data included in the data group 302 are added, and the average value thereof is calculated. Note that the dimension number is a value corresponding to the number of data types, and for example, in the case of spectral data, it can correspond to the number of channels. Therefore, for example, in the case where the light receiving element array has 32 channels, that is, the light receiving element array detects the fluorescence by dividing the entire detection range into 32 wavebands, the dimension number of the spectral data acquired by the light receiving element array is 32.

[0147] Subsequently, clustering processing unit 214 performs dimensionality compression on all processed spectral data included in data group 302 to determine the first eigenvalue, the second eigenvalue, the variance of the first eigenvalue, and the variance of the second eigenvalue (step S302). For dimensionality compression of the data, statistical data analysis methods such as principal component analysis can be used. Note that the first and second eigenvalues ​​can be arbitrarily selected according to predetermined rules or from the dimensionally compressed dimensions. The number of eigenvalues ​​is not limited to two of the first and second eigenvalues; it can be, for example, one, three, or more.

[0148] Subsequently, clustering processing unit 214 calculates the initial node vector for each of the representative nodes N11 to N33 using the following equation (3) (step S303). Note that in equation (3), the initial value (initial node vector) of the representative node Nij with coordinates (i, j) is represented by N_ij. In equation (3), the row partition number is the number of representative nodes arranged along the row direction, which is three in this example. Similarly, the column partition number is the number of representative nodes arranged along the column direction, which is three in this example.

[0149] The initial node vector N_ij = average value per dimension + (first eigenvalue / number of row partitions) × i + (second eigenvalue / number of column partitions) × j (3)

[0150] By determining the initial node vector (initial value) for each representative node as described above, the same initial node vector can always be set for data groups 302 that belong to the same cluster. This avoids the clustering results from varying depending on the execution (corresponding to the first problem mentioned above).

[0151] 2.2.4 Clustering through Batch Learning

[0152] Figure 16 It is shown Figure 14 The flowchart of an example of clustering by batch learning according to this embodiment is described in step S223. Note that... Figure 16 The example shown illustrates how 100 processed spectral data (vectors (coordinate values)) from fluorescence separation and preprocessing performed by the preprocessing / fluorescence separation unit 211 are classified into nine representative nodes N11 to N33 arranged in a three-row, three-column grid in a two-dimensional coordinate system in the initial state, similar to... Figure 15 The example shown.

[0153] like Figure 16As illustrated, in clustering by batch learning according to the present embodiment, first, the clustering processing unit 214 performs the following processing: the processed spectral data is selected from the data group 302 one by one, and the selected processed spectral data is assigned to the representative node closest thereto (step S304). This assignment processing is repeated until the assignment of all the processed spectral data in the data group 302 (or the number equal to or greater than the necessary number) is completed. In Figure 16 In the example illustrated, the processed spectral data of the cell identifiers "1" and "3" are assigned to the representative node N11, the processed spectral data of the cell identifier "2" is assigned to the representative node N12, and the processed spectral data of the cell identifier "100" is assigned to the representative node N33 (the cell identifiers = "4" to "99" are not described).

[0154] Note that the cell identifier can be an identifier uniquely identifying the cell to which the processed spectral data registered in the data group 302 corresponds. Further, the order in which the processed spectral data is selected from the data group 302 is not particularly limited, and can be, for example, various orders such as an ascending order of the cell identifier for uniquely identifying the cell of the processed spectral data registered in the data group 302 and the registration order in the data group 302.

[0155] Next, the clustering processing unit 214 performs processing of updating the representative node vector of each of the representative nodes N11 to N33 based on the processed spectral data assigned to each of the representative nodes N11 to N33 (step S305). In the update of the representative node vector, for example, the updated new representative node vector is calculated in accordance with the following equation (4).

[0156] New representative node vector N_ij = (1 - a) x old representative node vector N_ij + a x average value of assigned processed spectral data (4)

[0157] Next, the clustering processing unit 214 performs, on the peripheral representative nodes located around the representative node having the representative node vector N_ij updated in step S305, processing represented by the following equation (5) of multiplying the amount of change by equation (4) by β, thereby updating the peripheral representative node vector N_ij of the peripheral representative node (step S306). Note that in equation (5), β(r) can be similar to β(r) in equation (2) described above.

[0158] New peripheral representative node vector N_ij = (1 - β(r)) x old peripheral representative node vector N_ij + β(r) x average value of processed spectral data assigned to the representative node (5)

[0159] In this embodiment, the representative node vector of each of the representative nodes N11 to N33 is updated by repeating the processing in steps S304 to S306 a predetermined number of times (e.g., once or more). Note that in the repetition of the processing in steps S304 to S306, the learning ratio α and / or β can be changed (e.g., decreased or increased) each time the processing is repeated.

[0160] As described above, by assigning all the processed spectral data to be analyzed to any one of the representative nodes N11 to N33, and then updating the representative node vector of each of the representative nodes N11 to N33 using the processed spectral data assigned to each of the representative nodes N11 to N33, random sampling during learning can be eliminated. Therefore, the clustering results can be avoided from varying depending on the execution (corresponding to the second question).

[0161] Furthermore, since the update of the representative node vector is performed after all the processed spectral data has been allocated, the update of the representative node vector from the allocation of processed spectral data can be subdivided and performed by different information processing devices.

[0162] For example, the processed spectral data included in data group 302 can be divided into multiple groups, and the spectral data belonging to each group can be assigned to representative nodes by different information processing devices. For example, the clustering processing unit 214 according to this embodiment includes an allocation unit that assigns multiple processed spectral data to any one of the representative nodes N11 to N33. The number of allocation units included in the clustering processing unit 214 can be equal to or less than the number of groups, and the allocation units can be executed in different information processing devices.

[0163] Therefore, since the processing of spectral data can be performed in parallel, processing efficiency and speed can be improved.

[0164] Furthermore, since the amount of data handled by each information processing device can be reduced by subdividing the processing, problems such as increased processing time or the curse of dimensionality during clustering can also be mitigated.

[0165] 2.2.5 Using Consistent Clustering to Determine the Number of Clusters

[0166] Figure 17 It is used to illustrate its use. Figure 14 The graph of the consistent clustering operation described in step S224 to determine the number of clusters. Note that... Figure 17 The example shown illustrates how 100 processed spectral data (vectors (coordinate values)) after fluorescence separation and preprocessing by the preprocessing / fluorescence separation unit 211 are initially arranged in a two-dimensional coordinate system as nine representative nodes N11 to N33 in a three-row, three-column configuration, similar to...Figure 15 and Figure 16 the example shown in (a) of FIG. 12.

[0167] In the above description, as shown in (a) of FIG. 11, it has been exemplified that the number of classification groups of the processed spectral data included in the data group 302 (i.e., the number of representative nodes) is set to nine in the initial setting. However, the number of representative nodes (nine) in the initial setting is not necessarily optimal, depending on the processed spectral data included in the data group 302. Figure 17 Therefore, in the present embodiment, in step S224 of FIG. 10, by performing meta-clustering such as consistency clustering, a more appropriate division number of representative nodes and a combination of representative nodes (hereinafter referred to as meta-representative nodes) after meta-clustering are determined as shown in (b) of FIG. 11.

[0168] Figure 14 Figure 17

[0169] Figure 17 The example shown in (b) of FIG. 11 exemplifies a case where, as a result of meta-clustering, the representative nodes N11, N12, and N21 constitute one meta-representative node NN1, the representative nodes N22, N31, and N32 constitute one meta-representative node NN2, and the representative nodes N13, N23, and N33 constitute one meta-representative node NN3.

[0170] Note that, in the present embodiment, the number of node divisions (nine in the present example) set when clustering is performed can be a division number determined by default, or can be a division number set by a user. In a case where the initial number of node divisions is set by a user, the information processing apparatus 201 can further include an operation input unit that functions as a node setting unit that sets the number of node divisions by a user.

[0171] 2.3 Effects

[0172] As described above, according to the present embodiment, since the initial node vector (initial value) of each representative node is determined based on the average value of the processed spectral data of each dimension and the feature value obtained by dimension compression of the processed spectral data, the same initial node vector can be set for the data group 302 having the same group at all times. Thereby, it is possible to avoid variation of the clustering result depending on the execution situation.

[0173] Further, according to the present embodiment, by assigning all of the processed spectral data to be analyzed to any one of the representative nodes N11 to N33, and then updating the representative node vector of each of the representative nodes N11 to N33 using the processed spectral data assigned to each of the representative nodes N11 to N33, it is possible to avoid variation of the clustering result depending on the execution situation.

[0174] ​​​Further, in the present embodiment, since the update of the representative node vector is performed after the allocation of all the processed spectral data is completed, the process from the spectral data after the allocation process to the update of the representative node vector can be subdivided and performed by different information processing devices.

[0175] Further, since the amount of data to be handled by each information processing device can be reduced by the subdividing process, it is also possible to suppress problems such as an increase in processing time or a disaster dimension occurring at the time of clustering.

[0176] Note that, in the present embodiment, the case where the spectral data after the fluorescence separation is clustered by the clustering processing unit 214 is exemplified, but the clustering target is not limited thereto, and for example, the measurement data 2 acquired by the measurement device 3 can be subjected to the clustering process by the clustering processing unit 214. In this case, the pre-processing / fluorescence separation unit 211 performs the pre-processing on the measurement data 2, and inputs the spectral data before the fluorescence separation obtained by the pre-processing by the clustering processing unit 214. Then, the clustering processing unit 214 performs the above-described clustering process on the input spectral data before the fluorescence separation.

[0177] Further, in the present embodiment, the case where the pre-processing / fluorescence separation unit 211 performs the pre-processing on the fluorescence spectrum of each fluorescent dye obtained by performing the fluorescence separation process on the spectral data included in the measurement data 2 has been exemplified, but the operation of the pre-processing / fluorescence separation unit 211 is not limited thereto, and for example, the pre-processing / fluorescence separation unit 211 can perform the pre-processing on the spectral data included in the measurement data 2 first, and then can perform the fluorescence separation process.

[0178] Other configurations, operations, and effects can be similar to those of the above-described first embodiment, and thus will not be described here.

[0179] 3. Hardware configuration of information processing device

[0180] Next, the hardware configuration of the information processing devices 1 and 201 according to the above-described embodiments will be described with reference to Figure 18 Figure 18 is a block diagram illustrating an example of the hardware configuration of the information processing device according to the embodiments of the present disclosure. Note that, in the following description, the information processing device 1 will be exemplified, but the same applies to the information processing device 201.

[0181] As shown in Figure 18 , the information processing device 1 includes a central processing unit (CPU) 901, a read only memory (ROM) 903, and a random access memory (RAM) 905. Further, the information processing device 1 includes an auxiliary storage device 907, a communication interface (I / F) 909, a drive 911, a display 913, a keyboard 915, and a mouse 917. Figure 1 ​The host bus 907, the bridge 909, the external bus 911, the interface 913, the input device 915, the output device 917, the storage device 919, the drive 921, the connection port 925, and the communication device 929 are included. Instead of or in addition to the CPU 901, the information processing device 1 can include a processing circuit called a digital signal processor (DSP) or an application specific integrated circuit (ASIC).

[0182] The CPU 901 functions as an arithmetic processing device and a control device, and controls the overall operation in the information processing device 1 or a part thereof in accordance with various programs recorded in the ROM 903, the RAM 905, the storage 919. For example, the CPU 901 controls the overall operation of each functional unit in the information processing device 1 included in the above-described embodiments. The ROM 903 stores programs used by the CPU 901, operation parameters, and the like. The RAM 905 mainly stores programs used in the execution of the CPU 901, parameters appropriately changed in the execution, and the like. The CPU 901, the ROM 903, and the RAM 905 are connected to each other by the host bus 907 constituted by an internal bus such as a CPU bus. Further, the host bus 907 is connected to the external bus 911 such as a peripheral component interconnect / interface (PCI) bus via the bridge 909.

[0183] The input device 915 is a device operated by a user, such as a mouse, a keyboard, a touch panel, a button, a switch, or a lever. The input device 915 can be, for example, a remote control device using infrared rays or other radio waves, or an external connection device 927 such as a mobile phone corresponding to the operation of the information processing device 1. The input device 915 includes an input control circuit that generates an input signal in accordance with information input by the user and outputs the input signal to the CPU 901. By operating the input device 915, the user inputs various types of data to the information processing device 1 or instructs the information processing device 1 to perform a processing operation.

[0184] The output device 917 is constituted by a device capable of visually or aurally notifying a user of acquired information. The output device 917 can be, for example, a display device such as an LCD, a PDP, an OELD, or the like, a sound output device such as a speaker, a headphone, or the like, or a printer device. The output device 917 outputs a result obtained by the processing of the information processing device 1 as a video such as text or an image, or a sound such as audio.

[0185] The storage device 919 is a data storage device configured as one example of a memory of the information processing device 1. The storage device 919 is configured by, for example, a magnetic storage device such as a hard disk drive (HDD), a semiconductor storage device, an optical storage device, or a magneto-optical storage device. The storage device 919 stores programs executed by the CPU 901 and various types of data, various types of data acquired from the outside, and the like.

[0186] The drive 921 is a reader / writer for a removable recording medium 923 such as a magnetic disk, an optical disk, a magneto-optical disk, or a semiconductor memory, and is built in or externally attached to the information processing device 1. The drive 921 reads information recorded in the attached removable recording medium 923 and outputs the information to the RAM 905. In addition, the drive 921 writes a record in the attached removable recording medium 923.

[0187] The connection port 925 is a port for directly connecting a device to the information processing device 1. The connection port 925 can be, for example, a universal serial bus (USB) port, an IEEE 1394 port, or a small computer system interface (SCSI) port. In addition, the connection port 925 can be an RS-232C port, an optical audio terminal, a high-definition multimedia interface (HDMI) (registered trademark) port, or the like. By connecting an external connection device 927 to the connection port 925, various types of data can be exchanged between the information processing device 1 and the external connection device 927.

[0188] The communication device 929 is, for example, a communication interface configured by a communication device or the like for connecting to a communication network NW. The communication device 929 can be, for example, a communication card for a wired or wireless local area network (LAN), Bluetooth (registered trademark), or wireless USB (WUSB). In addition, the communication device 929 can be a router for optical communication, a router for asymmetric digital subscriber line (ADSL), a modem for various types of communication, or the like. The communication device 929 transmits and receives signals and the like to and from the Internet or another communication device using a predetermined protocol such as TCP / IP. In addition, the communication network NW connected to the communication device 929 is a network connected in a wired or wireless manner, and is, for example, the Internet, a home LAN, infrared communication, radio wave communication, or satellite communication.

[0189] Note that the technical scope of the present disclosure is not limited to the above-described embodiments, and various modifications can be made without departing from the spirit of the present disclosure. Furthermore, components of different embodiments and modifications can be appropriately combined with each other.

[0190] For example, in the above embodiments, the information processing system 4 includes the information processing device 1 or 201 and the measurement device 3, but the present technology is not limited to such an example. For example, the information processing device 1 or 201 can have a function (a measurement function) of the measurement device 3. In this case, the information processing system 4 is realized by the information processing device 1 or 201. Further, the measurement device 3 can have a function of the information processing device 1 or 201. In this case, the information processing system 4 is realized by the measurement device 3. Further, the measurement device 3 can have some functions of the information processing device 1 or 201, and the information processing device 1 or 201 can have some functions of the measurement device 3.

[0191] Note that the present technology can also have the following configurations. (1)

[0193] An information processing device, comprising:

[0194] a dimension compression unit that performs dimension compression on each piece of spectral data of a plurality of pieces of spectral data, the spectral data including a fluorescent component emitted from each of a plurality of particles labeled with one or more fluorescent dyes;

[0195] an initial value determination unit that determines an initial value for each of a plurality of nodes based on a result of the dimension compression; and

[0196] a clustering unit that clusters the plurality of pieces of spectral data using the initial value. (2)

[0198] The information processing device according to (1), further comprising an average value calculation unit configured to calculate an average value of the plurality of pieces of spectral data for each dimension, wherein

[0199] The initial value determination unit determines the initial value for each of the plurality of nodes based on the average value of the plurality of pieces of spectral data for each dimension in addition to the result of the dimension compression. (3)

[0201] The information processing device according to (1) or (2), wherein

[0202] The clustering unit includes:

[0203] an assignment unit that assigns each piece of spectral data of the plurality of pieces of spectral data to any one of the plurality of nodes; and

[0204] an update unit that updates a node vector of each of the plurality of nodes based on the spectral data assigned to each of the plurality of nodes. (4)

[0206] The information processing device according to (3), wherein

[0207] The clustering unit includes:

[0208] a node number changing unit that changes the number of nodes by performing consistent clustering on the node vector of each of the plurality of nodes updated by the updating unit; and

[0209] a meta-clustering unit that performs meta-clustering on the nodes after changed by the node number changing unit, based on the node vector of each of the plurality of nodes before changed. (5)

[0211] The information processing apparatus according to (4), in which

[0212] The clustering unit further includes a division unit that divides the plurality of pieces of spectral data into two or more groups, and

[0213] the allocation unit allocates each of the plurality of pieces of spectral data to any of the plurality of nodes in each of the two or more groups. (6)

[0215] The information processing apparatus according to (5), in which

[0216] The clustering unit includes the allocation unit equal to or less than the number of groups, and

[0217] The allocation unit is arranged in a different information processing apparatus. (7)

[0219] The information processing apparatus according to any one of (1) to (6), in which

[0220] The clustering unit performs clustering using a self-organizing map (SOM) algorithm. (8)

[0222] The information processing apparatus according to any one of (1) to (7), further including

[0223] a node setting unit that causes a user to set the number of nodes. (9)

[0225] The information processing apparatus according to any one of (1) to (8), in which

[0226] The dimension compression unit performs principal component analysis on the plurality of pieces of spectral data as the dimension compression. (10)

[0228] The information processing apparatus according to any one of (1) to (9), in which

[0229] The spectral data is spectral data measured by a spectral flow cytometer. (11)

[0231] The information processing apparatus according to any one of (1) to (10), further comprising

[0232] a preprocessing unit that performs a scale conversion on each of the plurality of pieces of spectral data, wherein

[0233] The dimension compression unit performs dimension compression on each of the plurality of pieces of spectral data on which the scale conversion has been performed. (12)

[0235] The information processing apparatus according to (11), further comprising

[0236] a fluorescence separation unit that separates each of the plurality of pieces of spectral data into a fluorescence spectrum for each of the types of fluorescent dyes, wherein

[0237] The preprocessing unit performs a scale conversion on each of the plurality of fluorescence spectra. (13)

[0239] The information processing apparatus according to (11) or (12), wherein

[0240] The preprocessing unit performs a nonlinear process conversion as the scale conversion. (14)

[0242] The information processing apparatus according to (13), wherein

[0243] The preprocessing unit performs a logical conversion, a logarithmic conversion, or a double exponential conversion as the scale conversion. (15)

[0245] The information processing apparatus according to any one of (1) to (14), further comprising

[0246] a display control unit that displays a result of the clustering performed by the clustering unit. (16)

[0248] The information processing apparatus according to any one of (11) to (14), further comprising

[0249] a display control unit that displays a result of the clustering performed by the clustering unit, wherein

[0250] The display control unit displays data on which a scale conversion has been performed by the preprocessing unit. (17)

[0252] An information processing method comprising:

[0253] performing dimension compression on each of the plurality of pieces of spectral data including a fluorescent component emitted from each of a plurality of particles labeled with one or more fluorescent dyes;

[0254] determining an initial value for each of the plurality of nodes based on a result of the dimension compression; and

[0255] performing clustering on the plurality of pieces of spectral data using the initial value. (18)

[0257] A program causes a computer to execute:

[0258] performing dimension compression on each of the plurality of pieces of spectral data including a fluorescent component emitted from each of a plurality of particles labeled with one or more fluorescent dyes;

[0259] determining an initial value for each of the plurality of nodes based on a result of the dimension compression; and

[0260] performing clustering on the plurality of pieces of spectral data using the initial value. (19)

[0262] An information processing system comprising:

[0263] a measurement device that detects each of a plurality of pieces of spectral data including a fluorescent component emitted from each of a plurality of particles labeled with one or more fluorescent dyes; and

[0264] an information processing device that clusters the plurality of pieces of spectral data detected by the measurement device, wherein

[0265] the information processing device comprises:

[0266] a dimension compression unit that performs dimension compression on each of the plurality of pieces of spectral data detected by the measurement device;

[0267] an initial value determination unit that determines an initial value for each of the plurality of nodes based on a result of the dimension compression; and

[0268] a clustering unit that performs clustering on the plurality of pieces of spectral data using the initial value.

[0269] List of Reference Signs

[0270] 1, 201 Information processing device

[0271] 2 Measurement data

[0272] 3 Measurement device

[0273] 4, 200 Information processing system

[0274] 11 preprocessing unit

[0275] 12 preprocessing parameter table

[0276] 13 spectral output unit

[0277] 14, 214 cluster processing unit

[0278] 15 cluster result presentation unit

[0279] 16 fluorescence separation unit

[0280] 17 routine analysis presentation unit

[0281] 211 preprocessing / fluorescence separation unit

[0282] 302 data set

Claims

1. An information processing device, comprising: A dimensionality compression unit performs dimensionality compression on each of multiple spectral data points, the spectral data including fluorescent components emitted from each of multiple particles labeled with one or more fluorescent dyes, wherein the dimensionality compression includes a statistical data analysis method using t-distributed random neighbor embedding (T-SNE). An initial value determination unit determines the initial value of each of the multiple nodes based on the result of the dimensional compression. A clustering unit performs clustering on multiple spectral data using the initial values, wherein the clustering unit comprises: The allocation unit assigns each of the multiple spectral data points to any one of the multiple nodes; and The update unit updates the node vector of each of the plurality of nodes based on the spectral data assigned to each of the plurality of nodes; and The preprocessing unit performs scaling transformation on each of the multiple spectral data sets, wherein the dimensionality compression unit performs dimensionality compression on each of the multiple spectral data sets that have already undergone scaling transformation.

2. The information processing device according to claim 1, further comprising: The average value calculation unit calculates the average value of multiple spectral data points for each dimension, wherein... In addition to the result of the dimensionality compression, the initial value determination unit also determines the initial value of each of the plurality of nodes based on the average value of the multiple spectral data for each dimension.

3. The information processing device according to claim 1, wherein, The clustering unit includes: The node count changing unit changes the node count by performing consistent clustering on the node vector of each of the plurality of nodes updated by the updating unit; and The meta-clustering unit performs meta-clustering on the nodes after they have been changed by the node number change unit, based on the node vector of each of the multiple nodes before the change.

4. The information processing device according to claim 3, wherein, The clustering unit further includes a partitioning unit, which divides the multiple spectral data into two or more groups, and The allocation unit assigns each of the multiple spectral data to any one of the multiple nodes in each of the two or more groups.

5. The information processing device according to claim 4, wherein, The clustering unit includes the allocation unit that is equal to or less than the number of groups, and The allocation units are respectively arranged in different information processing devices.

6. The information processing device according to claim 1, wherein, The clustering unit uses the self-organizing map algorithm to perform clustering.

7. The information processing device according to claim 1, further comprising: The node setting unit allows users to set the number of nodes.

8. The information processing device according to claim 1, wherein, The dimensionality compression unit performs principal component analysis on multiple spectral data sets as dimensionality compression.

9. The information processing device according to claim 1, wherein, The spectral data were measured using a spectral flow cytometer.

10. The information processing device according to claim 1, further comprising: The fluorescence separation unit separates each of the multiple spectral data points into the fluorescence spectrum of each fluorescent dye, wherein... The preprocessing unit performs the scaling transformation on each of the multiple fluorescence spectra.

11. The information processing device according to claim 1, wherein, The preprocessing unit performs nonlinear processing as a scaling transformation.

12. The information processing device according to claim 11, wherein, The preprocessing unit performs logical transformations, logarithmic transformations, or double exponential transformations as scaling transformations.

13. The information processing device according to claim 1, further comprising: The display control unit displays the results of clustering performed by the clustering unit.

14. The information processing device according to claim 1, further comprising: The display control unit displays the clustering results performed by the clustering unit, wherein... The display control unit displays the data that has undergone scaling transformation by the preprocessing unit.

15. An information processing method, comprising: Dimensional compression is performed on each of multiple spectral data points, the spectral data including fluorescent components emitted from each of multiple particles labeled with one or more fluorescent dyes, wherein the dimensional compression includes a statistical analysis method using t-distributed random neighbor embedding (T-SNE). Based on the result of the dimensionality compression, determine the initial value of each of the multiple nodes; and Clustering is performed on the plurality of spectral data using the initial values, wherein each of the plurality of spectral data is assigned to any one of the plurality of nodes; and the node vector of each of the plurality of nodes is updated based on the spectral data assigned to each of the plurality of nodes; and A scaling transformation is performed on each of the plurality of spectral data, wherein the dimensionality compression is performed on each of the plurality of spectral data for which the scaling transformation has been performed.

16. A computer-readable storage medium storing a program configured to cause a computer to execute: A step of performing dimensionality compression on each of multiple spectral data points, wherein the spectral data includes fluorescent components emitted from each of a plurality of particles labeled with one or more fluorescent dyes, wherein... The dimensionality compression includes a statistical analysis method using t-distributed random neighbor embedding T-SNE; The step of determining the initial value of each of the multiple nodes based on the result of the dimensional compression; and The initial value is used to perform a clustering step on the multiple spectral data, wherein each of the multiple spectral data is assigned to any one of the multiple nodes; and the node vector of each of the multiple nodes is updated based on the spectral data assigned to each of the multiple nodes; and A scaling transformation is performed on each of the plurality of spectral data, wherein the dimensionality compression is performed on each of the plurality of spectral data for which the scaling transformation has been performed.

17. An information processing system, comprising: A measuring device that detects multiple spectral data, the spectral data including fluorescent components emitted from each of a plurality of particles labeled with one or more fluorescent dyes; as well as The information processing device performs clustering on multiple spectral data points detected by the measuring device, wherein... The information processing device includes: The dimension compression unit performs dimension compression on each of the multiple spectral data detected by the measurement device, wherein the dimension compression includes a statistical data analysis method using t-distributed random neighbor embedding (T-SNE). The initial value determination unit determines the initial value of each of the multiple nodes based on the result of the dimensionality compression; and A clustering unit performs clustering on multiple spectral data using the initial values, wherein the clustering unit comprises: The allocation unit assigns each of the multiple spectral data points to any one of the multiple nodes; and The update unit updates the node vector of each of the plurality of nodes based on the spectral data assigned to each of the plurality of nodes; and The preprocessing unit performs scaling transformation on each of the multiple spectral data sets, wherein the dimensionality compression unit performs dimensionality compression on each of the multiple spectral data sets that have already undergone scaling transformation.

Images