Spectral data acquisition method, cell classification method, and raman spectroscopy system

Abstract

To provide a spectral data acquisition method which enables more uniform laser irradiation while reducing damage to cells further.SOLUTION: A spectral data acquisition method is provided, comprising circularly irradiating an area having cells 50 therein with laser light from a laser light source 12, receiving Raman scattered light emitted from the cells 50 caused by the irradiation using a spectroscope 14, and detecting a Raman spectrum corresponding to the received Raman scattered light using a CCD (charge-coupled device) detector 16.SELECTED DRAWING: Figure 1

Description

[Technical field]

[0001] The present invention relates to a method for acquiring spectral data, a method for classifying cells, and a Raman spectroscopy system. [Background technology]

[0002] Conventionally, there are microscope systems that use Raman scattered light to evaluate cells. Raman scattered light obtained from cells indicates the distribution and amount of intracellular molecules such as proteins and lipids. Therefore, it is possible to determine the state of the cells non-invasively and without staining based on the Raman spectrum obtained from the Raman scattered light.

[0003] For example, Non-Patent Document 1 discloses a technique for acquiring Raman scattered light by specifying a square region within a cell based on an image obtained from a quantitative phase microscope (QPM) and irradiating it with laser light using a random scan system. Non-Patent Document 2 discloses a cell sorting technique for measuring the entire region of a cell by linear scanning to create a Raman image, and using the spectrum of the cell region extracted from the Raman image. [Prior art documents] [Non-patent literature]

[0004] [Non-Patent Document 1] Nicolas Pavillon, Nicholas I. Smith, “Maximizing throughput in label-free microspectroscopy with hybrid Raman imaging”, Journal of Biomedical Optics, January 2015, Vol. 20(1), 016007-1~016007-10 [Non-Patent Document 2] Taro Ichimura, “Non-label immune cell state prediction using Raman spectroscopy”, Scientific Reports 6,37562; DOI: 10.1038 / srep37562 (2016) Summary of the Invention [Problem to be solved by the invention]

[0005] However, in the line irradiation method as in Non-Patent Document 2, the selection of cells using Raman images takes a long time for measurement and analysis, and also increases damage to the cells. Since the Raman scattered light from cells is very weak and has low sensitivity, extending the exposure time to the laser light may increase the sensitivity of the spectrum, but it also causes problems such as greater damage to the cells and higher noise. In addition, in the random scanning method as in Non-Patent Document 1, it is difficult to uniformly irradiate a specific region of a cell with a laser.

[0006] The present invention has been made in view of the above problems. [Means for solving the problem]

[0007] In order to solve the above problems, one aspect of the present invention is a spectral data acquisition method including the steps of receiving Raman dispersed light emitted from a cell by irradiating a region in which the cell is present with a laser light in a circular manner, and detecting a Raman spectrum corresponding to the received Raman dispersed light.

[0008] Another aspect of the present invention is a method for classifying cells, comprising the step of classifying cells based on spectral data acquired by the above method. Another aspect of the present invention is a Raman spectroscopy system including a laser light source that emits laser light, a two-axis galvanometer mirror that reflects the laser light so that the laser light is irradiated in a circular manner onto an area where cells are present, a spectrometer that receives Raman dispersed light of the laser light emitted from the cells, and a CCD (charge-coupled device) detector that detects a Raman spectrum according to the received Raman dispersed light. Effect of the Invention

[0009] Laser light can be irradiated in a circular or arbitrary pattern within any region of a single cell, and Raman scattering light can be obtained from the entire irradiation range. This makes it possible to obtain Raman scattering light from any region within a single cell in a short time while minimizing damage to the cell.

[0010] In addition, by expanding the measurement area, it is possible to obtain Raman scattered light from multiple cells at once. In addition, by treating the obtained Raman scattered light as spectral data and performing multivariate analysis or machine learning, it is possible to classify the type and activation state of cells, which makes it possible to select cells in a short time, non-invasively, and without staining. [Brief description of the drawings]

[0011] [Figure 1] FIG. 2 is a diagram illustrating an example of a light path in a Raman spectroscopy system according to an embodiment of the present invention. [Diagram 2] FIG. 1 is a diagram illustrating an example of the configuration of a Raman spectroscopy system according to an embodiment of the present invention. [Diagram 3] FIG. 13 is a diagram showing an example of a state in which a specific, approximately circular region of a cell is irradiated with laser light. [Figure 4] FIG. 1 shows an example of a spectrum of solid palmitic acid. [Diagram 5] FIG. 13 is a diagram showing an example of a graph showing the obtained Raman spectrum. [Figure 6]FIG. 1 is a diagram for explaining an example of Raman measurement by a point irradiation method and an irradiation method according to an embodiment of the present invention. [Figure 7] FIG. 13 is a diagram showing an example of a graph showing the obtained Raman spectrum. [Figure 8] FIG. 1 is a diagram illustrating an example of a comparison between the obtained Raman spectrum and the spectrum of palmitic acid. [Figure 9] FIG. 1 is a diagram illustrating an example of a comparison between the obtained Raman spectrum and the spectrum of palmitic acid. [Figure 10] FIG. 1 is a diagram illustrating an example of a comparison between the obtained Raman spectrum and the spectrum of palmitic acid. [Figure 11] FIG. 1 is a diagram illustrating an example of cell type classification performed on cells in suspension using an irradiation method according to one embodiment of the present invention. [Figure 12] FIG. 1 is a diagram illustrating an example of cell type classification performed on cells in suspension using an irradiation method according to one embodiment of the present invention. [Figure 13] FIG. 1 is a diagram illustrating an example of cell type classification performed on cells in suspension using an irradiation method according to one embodiment of the present invention. [Figure 14] FIG. 1 is a diagram illustrating an example of cell type classification performed on cells in suspension using an irradiation method according to one embodiment of the present invention. [Figure 15] FIG. 1 is a diagram illustrating an example of cell type classification performed on cells in suspension using an irradiation method according to one embodiment of the present invention. [Figure 16] FIG. 1 is a diagram illustrating an example of cell type classification performed on cells in suspension using an irradiation method according to one embodiment of the present invention. [Figure 17] FIG. 1 illustrates an example of cell type classification of adherent cells using an irradiation method according to an embodiment of the present invention. [Figure 18] FIG. 1 illustrates an example of cell type classification of adherent cells using an irradiation method according to an embodiment of the present invention. [Figure 19]FIG. 1 illustrates an example of cell type classification of adherent cells using an irradiation method according to an embodiment of the present invention. [Figure 20] FIG. 1 illustrates an example of cell type classification of adherent cells using an irradiation method according to an embodiment of the present invention. [Figure 21] FIG. 1 illustrates an example of cell type classification of adherent cells using an irradiation method according to an embodiment of the present invention. [Figure 22] FIG. 1 illustrates an example of cell type classification of adherent cells using an irradiation method according to an embodiment of the present invention. [Diagram 23] FIG. 1 is a diagram for explaining an example of classification of identical cells in different states, performed using an irradiation method according to one embodiment of the present invention. [Figure 24] FIG. 1 is a diagram for explaining an example of classification of identical cells in different states, performed using an irradiation method according to one embodiment of the present invention. [Diagram 25] FIG. 1 is a diagram for explaining an example of classification of identical cells in different states, performed using an irradiation method according to one embodiment of the present invention. [Figure 26] FIG. 1 is a diagram for explaining an example of classification of identical cells in different states, performed using an irradiation method according to one embodiment of the present invention. [Figure 27] FIG. 1 is a diagram for explaining an example of classification of identical cells in different states, performed using an irradiation method according to one embodiment of the present invention. [Figure 28] FIG. 1 is a diagram for explaining an example of classification of identical cells in different states, performed using an irradiation method according to one embodiment of the present invention. [Figure 29] FIG. 1 is a diagram for explaining an example of classification of identical cells in different states (stimulation with PMA and A23187) performed using an irradiation method according to one embodiment of the present invention. [Diagram 30] FIG. 1 is a diagram for explaining an example of classification of identical cells in different states (stimulation with PMA and A23187) performed using an irradiation method according to one embodiment of the present invention. [Diagram 31]FIG. 1 is a diagram for explaining an example of classification of identical cells in different states (stimulation with PMA and A23187) performed using an irradiation method according to one embodiment of the present invention. [Diagram 32] FIG. 1 is a diagram for explaining an example of classification of identical cells in different states (stimulation with PMA) performed using an irradiation method according to one embodiment of the present invention. [Diagram 33] FIG. 1 is a diagram for explaining an example of classification of identical cells in different states (stimulation with PMA) performed using an irradiation method according to one embodiment of the present invention. [Diagram 34] FIG. 1 is a diagram for explaining an example of classification of identical cells in different states (stimulation with PMA) performed using an irradiation method according to one embodiment of the present invention. [Diagram 35] FIG. 1 is a diagram for explaining an example of classification of identical cells in different states (stimulation with PHA) performed using an irradiation method according to one embodiment of the present invention. [Diagram 36] FIG. 1 is a diagram for explaining an example of classification of identical cells in different states (stimulation with PHA) performed using an irradiation method according to one embodiment of the present invention. [Figure 37] FIG. 1 is a diagram for explaining an example of classification of identical cells in different states (stimulation with PHA) performed using an irradiation method according to one embodiment of the present invention. [Figure 38] FIG. 1 is a diagram illustrating an example of an irradiation method and machine learning used to predict the state of Jurkat cells according to one embodiment of the present invention. [Figure 39] FIG. 1 is a diagram illustrating an example of an irradiation method and machine learning used to predict the state of Jurkat cells according to one embodiment of the present invention. [Diagram 40] FIG. 1 is a diagram illustrating an example of measurement of a plurality of cells performed using an irradiation method according to one embodiment of the present invention. [Diagram 41] FIG. 1 is a diagram illustrating an example of measurement of a plurality of cells performed using an irradiation method according to one embodiment of the present invention. [Diagram 42] FIG. 1 is a diagram illustrating an example of measurement of a plurality of cells performed using an irradiation method according to one embodiment of the present invention. [Diagram 43]FIG. 1 is a diagram illustrating an example of measurement of a plurality of cells performed using an irradiation method according to one embodiment of the present invention. [Diagram 44] FIG. 1 is a diagram illustrating an example of measurement of a plurality of cells performed using an irradiation method according to one embodiment of the present invention. [Diagram 45] FIG. 1 is a diagram illustrating an example of measurement of a plurality of cells performed using an irradiation method according to one embodiment of the present invention. [Figure 46] FIG. 1 is a diagram illustrating an example of measurement of a plurality of cells performed using an irradiation method according to one embodiment of the present invention. [Figure 47] FIG. 1 is a diagram illustrating an example of measurement of a plurality of cells performed using an irradiation method according to one embodiment of the present invention. [Figure 48] FIG. 1 is a diagram illustrating an example of measurement of a plurality of cells performed using an irradiation method according to one embodiment of the present invention. [Figure 49] FIG. 1 is a diagram illustrating an example of measurement of a plurality of cells performed using an irradiation method according to one embodiment of the present invention. [Figure 50] FIG. 1 is a diagram illustrating an example of measurement of a plurality of cells performed using an irradiation method according to one embodiment of the present invention. [Figure 51] FIG. 1 is a diagram illustrating an example of measurement of a plurality of cells performed using an irradiation method according to one embodiment of the present invention. [Figure 52] FIG. 1 is a diagram illustrating an example of measurement of a plurality of cells performed using an irradiation method according to one embodiment of the present invention. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0012] Hereinafter, embodiments of the present invention will be described in detail with reference to the drawings. (Light path of the Raman spectroscopy system) FIG. 1 is a diagram for explaining an example of an optical path in a Raman spectroscopy system according to the present embodiment. The Raman spectroscopy system 1 according to the present embodiment is a system capable of acquiring a spectrum from a wide area of ​​a cell at high speed in order to avoid fluctuations in the measurement area due to the complex structure of the cell. The Raman spectroscopy system 1 according to the present embodiment may be configured to include a diode-pumped solid-state (DPSS) laser light source 12, a spectrometer 14, a CCD (charge-coupled device) detector (cooled CCD camera) 16, an inverted microscope 18, and a Raman optical system 20. The laser light emitted from the DPSS laser light source 12 enters the optical path via a laser optical fiber 32, and passes through a neutral density (ND) filter 204, a shutter 205, and a bandpass filter 206 via a lens 202. The laser light is then reflected by a flat mirror 208 and a dichroic mirror 210, and reaches a two-axis galvanometer mirror 30 in which galvanometer mirrors are arranged on each of the orthogonal X-axis and Y-axis. By rotating the two-axis galvanometer mirror 30 at high speed, it is possible to irradiate the laser light in a circular manner to the area where the cell 50 is present, which is predetermined from the bright field image. In addition, as the laser light output from the DPSS laser light source 12, for example, various laser lights such as 266 nm (nanometers) (Deep UV laser), 320 nm (DPSS laser), 325 nm (He-Cd laser), 442 nm (He-Cd laser), 477 nm (air-cooled Ar+ laser), 488 nm (air-cooled Ar+ laser), 514 nm (air-cooled Ar+ laser), 532 nm (DPSS laser), 633 nm (He-Ne laser), 785 nm (near-infrared semiconductor laser), 839 nm (air-cooled semiconductor laser), and 1064 nm (air-cooled DPSS laser) can be adopted, and a person skilled in the art can appropriately select it.

[0013] FIG. 3 is a diagram showing an example of a state in which a specific approximately circular region of a cell is irradiated with laser light. As shown in FIG. 3, it is possible to irradiate a region within the cell with laser light for each individual cell (one cell). When measuring an individual cell, the irradiation region (irradiation area) can be specified as any region within the cell. For example, the irradiation region can be specified as wide as possible. For example, when a circular cell (such as a human T cell) with a diameter of 8 to 15 μm is to be measured, a circular region with a diameter of about 10 μm can be set as the irradiation region. In particular, since different Raman scattered light is generated in the nucleus region and the region other than the nucleus (cytoplasm) of the cell, it is preferable to specify the region covering the nucleus and cytoplasm as the measurement region. This makes it possible to accumulate Raman scattered light originating from the nucleus and cytoplasm in this system and detect it as one Raman signal. As another example, it is also possible to set the irradiation region to a limited region within the cell. For example, when an adhered cell is to be measured, it is also possible to limit the irradiation region to the region where the nucleus is present. It is also possible to limit the irradiation region to the region other than the nucleus (cytoplasm).

[0014] Alternatively, it is possible to set a wide area to be irradiated with the laser light, and irradiate the laser light to a plurality of cells present in the area at once (FIG. 40, etc.). The term "approximately circular" is intended to allow the case where the shape is not a perfect circle (true circle) but has some distortion (almost circular). In FIG. 3, the Raman spectroscopy system 1 according to this embodiment irradiates the laser light to an approximately circular area of ​​one or more cells, but it is also possible to irradiate the laser light to a quadrangular area (for example, various quadrangular shapes such as a square, a rectangle, a rhombus, a parallelogram, a trapezoid, etc.) of one or more cells. Furthermore, the Raman spectroscopy system 1 can irradiate the laser light in a circular manner to an area of ​​any shape, such as a triangle, a hexagon, etc., of one or more cells. Furthermore, the Raman spectroscopy system 1 according to this embodiment may irradiate the laser light in a circular manner from the outside to the inside (i.e., from the outer periphery toward the center of the area) to various shaped areas of one or more cells, or may irradiate the laser light in a circular manner from the inside to the outside. In addition, the Raman spectroscopy system 1 according to the present embodiment may irradiate various shape regions of one or more cells with laser light from the outside to the inside to near the center of the shape region in a spiral manner, and then irradiate the laser light from the inside to the outside of the shape region in a spiral manner in the same or opposite rotation direction, or may repeat irradiating the shape region from the outside to the inside and from the inside to the outside. This allows the shape region to be irradiated sufficiently without omission. In addition, the rotation direction when irradiating the laser light in a spiral manner may be clockwise or counterclockwise. In addition, the interval of the spiral line when rotating can be appropriately adjusted. In addition, "in a spiral manner (irradiating the laser light)" means irradiating the laser light in a spiral manner.

[0015] In the Raman spectroscopy system 1 of this embodiment, the range in which the laser light can be irradiated is, for example, a substantially circular area with a radius of 0.25 to 150 μm (micrometers), a rectangular area with a minor axis (the X-axis or Y-axis direction of the two-axis galvanometer mirror 212) of 0.5 to 300 μm, a marking speed of 0.002 to 2 mm / ms (millimeters / millisecond), etc. The moving speed of the laser light can be controlled by the vibration speed of the two-axis galvanometer mirror 30.

[0016] Returning to FIG. 1, the scattered light from the cell 50 irradiated with the laser light travels in the opposite direction along the same path as the incident path of the laser light, passing again through the dichroic mirror 210. The scattered light further passes through a high-pass filter 212, through which only the high-frequency components of the scattered light pass. As a result, only the Raman scattered light is focused on the spectrometer 14 via the lens 214, the shutter 215, and the optical fiber 34. The Raman scattered light is detected by the CCD detector 16 as the Raman shift (wave number, cm) on the horizontal axis. -1 ) and is detected as a Raman spectrum with the intensity of scattered light (au) on the vertical axis. (Configuration of Raman spectroscopy system) FIG. 2 is a diagram showing an example of the configuration of a Raman spectroscopy system according to this embodiment. In this configuration example, elements that are the same as or equivalent to those shown in FIG. 1 are given the same reference numerals. The Raman spectroscopy system 1 according to this embodiment may be configured to include a DPSS laser light source 12, a spectrometer 14, a CCD detector 16, an inverted microscope 18, a Raman optical system 20, and a shutter control device 38. The shutter controller 38 is a controller for controlling the shutter 205 and the shutter 215 shown in FIG. 1. The Raman spectroscopy system 1 may further include a joystick 44 operated by a user to adjust the stage 40 of the inverted microscope 18, and a stage controller 42 for controlling the stage of the inverted microscope 18 in response to the movement of the joystick 44. The Raman spectroscopy system 1 may further include an epi-illumination controller 46 for controlling the epi-illumination 48 of the inverted microscope 18.

[0017] The Raman spectroscopy system 1 may further include a computer device 50, an input / output user interface such as a keyboard 52 and a mouse 54 for a user to operate the computer device 50, and a monitor device 56. The computer device 50 may acquire spectral data of cells detected by the CCD detector 16 and display the spectral data on the monitor device 56 in a form such as a graph. The computer device 50 may also execute machine learning software. The machine learning software may be unsupervised or supervised machine learning software that takes spectral data as input and outputs an answer such as a classification of cells, as described below. The computer device 50 may be realized by a hardware configuration similar to that of a general computer device.

[0018] The computer device 50 may include, for example, a processor, a RAM (Random Access Memory), a ROM (Read Only Memory), an internal hard disk device, an external hard disk device, a removable memory such as a CD, a DVD, a USB memory, a memory stick, or an SD card, an input / output user interface (a touch panel, a speaker, a microphone, a lamp, etc.), and a wired / wireless communication interface capable of communicating with other computer devices. The computer device 50 may, for example, have a processor that reads various programs such as a spectral data analysis program and a machine learning program stored in advance in a hard disk device, a ROM, or a removable memory into a memory such as a RAM, and executes the spectral data of the cells detected by the CCD detector 16 while appropriately reading them from the hard disk device, the ROM, the removable memory, etc.

[0019] The Raman spectroscopy system and the spectral data acquisition method described above can acquire Raman spectra over a wide range in a short time, which makes it possible to irradiate cells with the laser more efficiently and uniformly while minimizing damage to the cells. (Cell Classification Methods) The present invention also provides a method for classifying cells. As described above, the method for classifying cells includes a step of classifying cells based on spectral data acquired from the cells. Here, the classification of cells includes classification of cell types and classification of cell states. The cell states include the physiological state of the cells, the activation state of the cells (due to endogenous or exogenous stimulants, signal transducing substances, other compounds, etc.), and the like. Such states include, but are not limited to, cell division, cell death, canceration of the cells, cell differentiation, dedifferentiation of the cells, changes in metabolic state, changes in shape or morphology, infection of the cells by foreign cells (infectious cells such as viruses and bacteria), and the like. The cells may be in either a floating state or an adherent state.

[0020] The cells used as the subject of analysis are not particularly limited, and include animal cells, plant cells, microbial cells, and the like. Examples of animal cells include cells derived from mammals, birds, reptiles, amphibians, fish, and the like, and preferably cells derived from humans, dogs, mice, rats, cats, horses, goats, sheep, cows, frogs, and the like. Examples of cells include, but are not limited to, undifferentiated cells, totipotent cells, pluripotent cells, differentiated cells, germ cells, and cells included in somatic cells. More specifically, examples of cells include, but are not limited to, Jurkat cells (human T cell line), mouse embryonic fibroblasts (MEF), embryonic stem cells, immortalized human adipose-derived mesenchymal stem cells (hMSC), and human induced pluripotent stem cells (hiPSC). Examples of plant cells include cells derived from angiosperms, gymnosperms, ferns, mosses, algae, and the like, and preferably cells derived from vegetables, flowers, trees, and the like. Microbial cells may be either prokaryotes or eukaryotes. Examples include, but are not limited to, Escherichia coli, Staphylococcus aureus, halophiles, acidophiles, thermophiles, yeasts, fungi, and bacteria.

[0021] The step of classifying the cells includes a step of comparing the spectral data acquired from the cell to be analyzed with standard spectral data. Here, the standard spectral data is Raman spectral data acquired from a cell of known type or state. The standard spectral data may be acquired at any time. For example, the standard spectral data may be acquired at any stage before, during, or after the analysis.

[0022] The standard spectral data is obtained for each of cells of different types or states. When the standard spectral data of each cell has a characteristic spectral region, it is preferable to specify the characteristic spectral region. The characteristic spectral region may be one or more. When there are multiple characteristic spectral regions, all of the characteristic spectral regions can be used to classify the cells, but it is also possible to select particularly relevant features to classify the cells. Those skilled in the art can appropriately select such spectra.

[0023] The step of classifying cells can be performed by machine learning. Machine learning may be either supervised learning or unsupervised learning. By using machine learning, learning is performed autonomously from data acquired by analysis, so that the more the analysis is repeated, the higher the accuracy of cell classification can be expected.

[0024] The present invention will be described in more detail with reference to the following examples. However, the examples are provided for the understanding of the present invention and are not intended to limit the scope of the present invention. [Example] (Specific configuration of Raman spectroscopy system) The inventors performed cell measurements using the Raman spectroscopy system according to the present embodiment. In this measurement, a Ti2-U microscope manufactured by Nikon Corporation was used as the inverted microscope 18. A DPSS 532 nm green laser (DL 532-100, maximum output 120 mW (milliwatts)) coupled to an optical fiber was used as the excitation method. A Lambda 40xC / NA 0.95 objective lens (Nikon, CFI Plan Apo Lambda) was used at a laser light output of about 40 mW on the sample stage, with a laser spot size of less than 450 nm. The backscattered Raman signal was dispersed by a spectrometer equipped with a 1200 g grating (Irix, STR200-2LC) and detected by a CCD camera (Andor Technology, iVac316). The detectable Raman spectral range was 46-3110 cm -1 In addition, a two-axis galvanometer mirror was placed in the laser beam path so that the laser beam could move in a specific area. (Raman data collection) To calibrate the spectrometer, the spectrum of sulfur, a Raman shift standard, was measured before the sample measurement. The sulfur spectrum was detected by a point scan system with an exposure time of 1 second using a neutral density (ND) filter of 1 or 5%. For correction, the three detected characteristic spectral peaks at 153.8, 219.1, and 473.2 cm were detected. -1 (ASTM E1840-96, DOI:10.1520 / E1840-96). (Comparison of Raman spectra between the point method and the irradiation method according to the present embodiment) A comparison was made between a conventional point irradiation method and the irradiation method using the Raman spectroscopy system 1 of the present embodiment.

[0025] (1) Comparison when palmitic acid is used as the analyte High-purity (<95%) solid palmitic acid was used as the analysis target. The spectrum of solid palmitic acid was obtained by the point irradiation method (laser spot diameter <450 nm) (FIG. 4(a)). On the other hand, in the irradiation method according to the present embodiment, the spectrum of solid palmitic acid was obtained from a circular area with a diameter of 5, 10, or 20 μm (FIGS. 4(b) to (d)). The laser exposure time was set to 1 second in each irradiation method, and the oscillation speed of the two-axis galvanometer mirror 30 was adjusted so that the laser could irradiate the entire area (1 mm / ms) even in a circle with a diameter of 20 μm.

[0026] In both methods, the peaks derived from palmitic acid were observed at 1059, 1124, 1293, 1419, 1435, 1462, 2842, 2879, and 2921 cm. -1 These spectra were detected at 100 nm (FIG. 5). When these spectra were compared with the spectrum database provided by Wiley Science Solutions (KnowItAll, https: / / sciencesolutions.wiley.com / ), it was confirmed that they matched the spectrum of palmitic acid (code: FFRX#478) (FIGS. 8 to 10). Therefore, it was demonstrated that the irradiation method of this embodiment generated spectral data comparable to the conventional point irradiation method without reducing the intensity.

[0027] (2) Comparison when live cells are used as the analysis subject Next, Raman measurements of live cells using the irradiation method according to the present embodiment were verified using Jurkat cells, a human T cell line. In this irradiation method, the exposure time was set to 3 seconds, and the spectrum was measured in a circular region with a diameter of 10 μm inside the cell (FIG. 6(b)). On the other hand, when the conventional point irradiation method was used, three spectra were obtained at random positions with the same exposure time of 3 seconds (FIG. 6(a)). In both the irradiation method of the present embodiment and the point irradiation method, Raman measurements were obtained in the fingerprint region of 600-1800 cm. -1 , nucleic acid (685, 748, 783 cm -1 ), protein (1003, 1450cm -1 ), lipids (1125, 1580, 1654cm -1) (Figure 7). The spectrum obtained by the irradiation method of this embodiment did not show a decrease in sensitivity compared to the spectrum obtained by point irradiation. Furthermore, the three spectra obtained by point irradiation had characteristic peaks such as 748 cm -1 Or 1125cm -1 There were several Raman shift regions with different intensities. Classification of cell types using the irradiation method according to the present embodiment To determine whether the irradiation method according to the present invention can be used to classify cell types, spectra were obtained from Jurkat cells, mouse embryonic fibroblasts (MEFs), immortalized human adipose-derived mesenchymal stem cells (hMSCs), and human induced pluripotent stem cells (hiPSCs).

[0028] (1) Cells in suspension MEFs, hMSCs, and hiPSCs, except for Jurkat cells, are usually cultured in an adherent state on a culture dish. Therefore, in order to match the state of Jurkat cells (suspension state), we cultured the cells in TrypLE TM Cells were measured in suspension after treatment with Express Enzyme® or Accutase (both from Thermo Fisher Scientific, Waltham, MA, USA). For each cell, Raman spectra were acquired from measuring a circular area of ​​10 μm diameter. The suspension measurements are shown in Figure 11. For each cell type, 50 cells were measured in suspension (total of 200 samples). The spectra obtained were smoothed, baseline corrected, and normalized. The average spectral data for each cell is shown in Figure 12.

[0029] To investigate the discrimination of each cell type, we used the acquired spectra as multivariate data for Principal Component Analysis (PCA), an unsupervised machine learning method. PCA extracted the first five principal components (PCs) that contributed to the variance of the total cells, and the distribution of cells relative to the five PCs was plotted as a pair plot or shown by kernel density estimation distribution (Figure 13). The contribution of the five PCs to the cell variance in the suspension state was 64.89%, 10.52%, 3.86%, 2.83%, and 1.63%. The loadings of the five PCs against wavenumber to the cell variance in the suspension state were shown in a graph (Figure 14). PC1 and PC2 in the suspension state allowed visual classification (Figure 15). Jurkat cells, MEFs, hMSCs, and iPSCs showed different cell distribution densities. In particular, it was shown that PC1 contributed to the classification of four types in the suspension state.

[0030] We then verified the classification by using machine learning to train on the spectral data. For this classification, we used Linear Discriminant Analysis (LDA), a supervised learning method. LDA is a method that extracts features between classes to create classification axes and optimize class separation. To verify the accuracy of the classification, we randomly split all the spectral data (200 floating cells) into 80% of the training data and 20% of the test data. LDA used the training data to define new three or two classification axes (the number of extracted axes is the number of classes minus one) that best classified the cell types. We then applied the test data to these extracted axes to verify the accuracy of the classification. Figure 16 shows the distribution of cells with respect to the LD1 and LD2 axes by cell type. The cells from the training data used to extract the classification axes are shown as kernel density distributions, and the cells used in the test data are shown as plots. LDA classified the four cell types more clearly than PCA, and the cells in the test data matched 100% with their original labels. By combining machine learning, it was possible to classify cell types with higher accuracy under floating conditions.

[0031] (2) Adherent cells MEF, hMSC, and hiPSC were also measured in the adherent state on a culture dish (Figures 17 to 22). In the adherent state, 100 cells of each cell type were measured (total of 300 samples). The obtained spectra were smoothed, baseline corrected, and normalized. The average spectral data for each cell is shown in Figure 18. It is possible to classify cell types in adherent culture.

[0032] To investigate the discrimination of each cell type, we used the acquired spectra as multivariate data for Principal Component Analysis (PCA), an unsupervised machine learning method. PCA extracted the first five principal components (PCs) that contributed to the variance of the total cells, and the distribution of cells relative to the five PCs was plotted as a pair plot or kernel density estimation distribution (Figure 19). The contribution of the five PCs to the cell variance in the adherent state was 49.7%, 12.3%, 7.16%, 3.93%, and 1.69%. The loadings of the five PCs to the cell variance in the adherent state against the wave number were shown graphically (Figure 20). PC2 and PC3 in the adherent state allowed the most visual classification (Figure 21). MEFs, hMSCs, and iPSCs showed different cell distribution densities. In particular, PC2 contributed to the classification of hMSCs and hiPSCs, and PC3 contributed to the classification of MEFs and hMSCs.

[0033] We then verified the classification by using machine learning to train on the spectral data. For this classification, we used Linear Discriminant Analysis (LDA), a supervised learning method. LDA is a method that extracts features between classes to create classification axes and optimize class separation. To verify the accuracy of the classification, we randomly split all the spectral data (300 cells in an adherent state) into 80% of the training data and 20% of the test data. LDA used the training data to define new three or two classification axes (the number of extracted axes is the number of classes minus one) that best classified the cell types. We then applied the test data to these extracted axes to verify the accuracy of the classification. Figure 22 shows the distribution of cells with respect to the LD1 and LD2 axes for each cell type. The cells in the training data used to extract the classification axes are shown as kernel density distributions, and the cells used in the test data are shown as plots. LDA classified the three cell types more clearly than PCA, and the cells in the test data matched 100% with their original labels. By combining machine learning, it was possible to classify cell types under adherent culture with higher accuracy.

[0034] (Classification of identical cells in different states) We investigated whether it is possible to distinguish between the same cell type in different states from the spectrum obtained by the irradiation method according to this embodiment. Jurkat cells (human T cell line) were stimulated to obtain activated cells. The activated cells and naive cells were analyzed. Magnetic beads bound to CD3 and CD28 antibodies were used to stimulate the cells (Figure 23). qPCR was performed to confirm that the cells were activated. It was confirmed that the gene expression of activation markers IL-2 and TNF-α was increased in the activated cells compared to the naive cells.

[0035] Three hundred spectra were obtained from each of the activated and naive cells by this irradiation method. To avoid the influence of the magnetic beads on the spectrum, the spectra were measured in a circular area with a diameter of 5 μm. The Raman shifts of 600–2980 cm for the activated and naive cells were measured. -1The average of the spectral data in this region is shown in FIG.

[0036] We next applied supervised machine learning to the spectral data to more clearly classify the activation states. Discriminant analysis-based methods Linear Discriminant Analysis (LDA) and Partial Least Squares Regression-Discriminant Analysis (PLS-DA), as well as artificial neural network-based methods Multilayer Perceptron (MLP) and Support Vector Machine (SVM) were tested to classify activated and naive cells with high accuracy.

[0037] For supervised learning, all data of activated and naive cells were randomly split into 80% of training data and 20% of test data. Using the training data, axes for classifying the two cell types using the four classification methods were extracted, and the test data was classified against these defined axes. The false positive ratio (FPR) and true positive rate (TPR) were calculated from the classified labels and original labels of the test data, and receiver operating characteristic (ROC) curves were constructed from the FPR and TPR. In addition, the area under the ROC curve, which means the classification accuracy, sensitivity, and specificity, and the root mean square error (RMSE) (Table 1) were calculated. Table 1 shows the average values ​​of 10 repetitions of the classification test.

[0038] [Table 1]

[0039] The AUCs of LDA, PLS-DA, SVM, and MLP were 0.93, 0.95, 0.97, and 0.98, respectively, which were all high. Using PLS-DA, which was used as a dimensionality reduction method, two features that contribute to the classification of cell states were extracted as PLS scores. Using scores 1 and 2, which extracted activated and naive cells, the training data was shown as a kernel density distribution. Then, the test data (shown as dots) was overlaid on the training data (Figure 25). By combining supervised learning methods, it was possible to classify activated and naive cells with higher accuracy. (Classification of identical cells in different states (stimulation with purified CD3 / CD28 antibodies)) We investigated whether it is possible to distinguish between the same cell type in different states from the spectrum obtained by the irradiation method according to this embodiment. Jurkat cells (human T cell line) were stimulated to obtain activated cells. The cells and naive cells were analyzed. Purified CD3 and CD28 antibodies were used to stimulate the cells. Here, "purified CD3 and CD28 antibodies" means free (not bound to magnetic beads) CD3 and CD28 antibodies. To confirm that the cells were activated, the positive rate of the activation marker CD69 was measured with a flow cytometer (Figure 26). It was confirmed that the positive rate of CD69 increased in the stimulated cells compared to the control (naive cells).

[0040] One hundred spectra were obtained from each of the activated and naive cells by this irradiation method. The spectra were measured in a circular area with a diameter of 10 μm. The Raman shifts of the activated and naive cells were 600–2980 cm -1 The average of the spectral data in this region is shown in FIG.

[0041] We next applied supervised machine learning to the spectral data to more clearly classify the activation states. Discriminant analysis-based methods Linear Discriminant Analysis (LDA) and Partial Least Squares Regression-Discriminant Analysis (PLS-DA), as well as artificial neural network-based methods Multilayer Perceptron (MLP) and Support Vector Machine (SVM) were tested to classify activated and naive cells with high accuracy.

[0042] For supervised learning, all data of activated and naive cells were randomly split into 80% of training data and 20% of test data. Using the training data, axes for classifying the two cell types using the four classification methods were extracted, and the test data was classified against these defined axes. The false positive ratio (FPR) and true positive rate (TPR) were calculated from the classified and original labels of the test data, and receiver operating characteristic (ROC) curves were constructed from the FPR and TPR. In addition, the area under the ROC curve, which means the classification accuracy, sensitivity, and specificity, and the root mean square error (RMSE) (Table 2) were calculated. Table 2 shows the average values ​​of 10 repetitions of the classification test.

[0043] [Table 2]

[0044] The AUCs of LDA, PLS-DA, SVM, and MLP were 0.99, 1.00, 1.00, and 0.99, respectively, which were all high. Using PLS-DA, which was used as a dimensionality reduction method, two features that contribute to the classification of cell states were extracted as PLS scores. Using scores 1 and 2, which extracted activated and naive cells, the training data was shown as a kernel density distribution. Then, the test data (shown as dots) was overlaid on the training data (Figure 28). By combining supervised learning methods, it was possible to classify activated and naive cells with higher accuracy. (Classification of cells in different states (stimulated with PMA and A23187)) We investigated whether the spectrum obtained by the irradiation method according to this embodiment can distinguish between the same cell type in different states. Jurkat cells (human T cell line) were stimulated to obtain activated cells. The cells and naive cells were analyzed. The cells were also stimulated with PMA (phorbol-12-myristate-13-acetate (known as a tumor promoter), 20 ng / ml) and calcium ionophore A23187 (1 μM). After that, the positive rate of the activation marker CD69 was measured with a flow cytometer for the treated cells (FIG. 29). It was confirmed that the positive rate of CD69 was almost 100% in the stimulated cells.

[0045] One hundred spectra were obtained from each of the activated and naive cells by this irradiation method. The spectra were measured in a circular area with a diameter of 10 μm. The Raman shifts of the activated and naive cells were 600–2980 cm -1 The average of the spectral data in this region is shown in FIG.

[0046] We next applied supervised machine learning to the spectral data to more clearly classify the activation states. Discriminant analysis-based methods Linear Discriminant Analysis (LDA) and Partial Least Squares Regression-Discriminant Analysis (PLS-DA), as well as artificial neural network-based methods Multilayer Perceptron (MLP) and Support Vector Machine (SVM) were tested to classify activated and naive cells with high accuracy.

[0047] For supervised learning, all data of activated and naive cells were randomly split into 80% of training data and 20% of test data. Using the training data, axes for classifying the two cell types using the four classification methods were extracted, and the test data was classified against these defined axes. The false positive ratio (FPR) and true positive rate (TPR) were calculated from the classified labels and original labels of the test data, and the receiver operating characteristic (ROC) curves were constructed from the FPR and TPR. In addition, the area under the ROC curve, which means the classification accuracy, sensitivity, and specificity, and the root mean square error (RMSE) (Table 3) were calculated. Table 3 shows the average values ​​of 10 repetitions of the classification test.

[0048] [Table 3]

[0049] The AUCs of LDA, PLS-DA, SVM, and MLP were 0.94, 0.97, 0.97, and 0.95, respectively, which were all high. Using PLS-DA used as a dimensionality reduction method, two features that contribute to the classification of cell states were extracted as PLS scores. Using scores 1 and 2, which extracted activated and naive cells, the training data was shown as a kernel density distribution. Then, the test data (shown as dots) was overlaid on the training data (Figure 31). By combining supervised learning methods, it was possible to classify activated and naive cells with higher accuracy. (Classification of identical cells in different states (stimulated with PMA)) We investigated whether it is possible to distinguish between the same cell type in different states from the spectrum obtained by the irradiation method according to this embodiment. Jurkat cells (human T cell line) were stimulated to obtain activated cells. The cells and naive cells were analyzed. The cells were also stimulated with PMA (20 ng / ml). After that, the positive rate of the activation marker CD69 was measured with a flow cytometer for the cells after the treatment (Figure 32). It was confirmed that the positive rate of CD69 was almost 100% for the cells after stimulation.

[0050] One hundred spectra were obtained from each of the activated and naive cells by this irradiation method. The spectra were measured in a circular area with a diameter of 10 μm. The Raman shifts of the activated and naive cells were 600–2980 cm -1 The average of the spectral data in this region is shown in FIG.

[0051] We next applied supervised machine learning to the spectral data to more clearly classify the activation states. Discriminant analysis-based methods Linear Discriminant Analysis (LDA) and Partial Least Squares Regression-Discriminant Analysis (PLS-DA), as well as artificial neural network-based methods Multilayer Perceptron (MLP) and Support Vector Machine (SVM) were tested to classify activated and naive cells with high accuracy.

[0052] For supervised learning, all data of activated and naive cells were randomly split into 80% of training data and 20% of test data. Using the training data, axes for classifying the two cell types using the four classification methods were extracted, and the test data was classified against these defined axes. The false positive ratio (FPR) and true positive rate (TPR) were calculated from the classified labels and original labels of the test data, and receiver operating characteristic (ROC) curves were constructed from the FPR and TPR. In addition, the area under the ROC curve, which means the classification accuracy, sensitivity, and specificity, and the root mean square error (RMSE) (Table 4) were calculated. Table 4 shows the average values ​​of 10 repetitions of the classification test.

[0053] [Table 4]

[0054] The AUCs of LDA, PLS-DA, SVM, and MLP were 0.97, 1.00, 1.00, and 0.99, respectively, which were all high. Using PLS-DA used as a dimensionality reduction method, two features that contribute to the classification of cell states were extracted as PLS scores. Using scores 1 and 2, which extracted activated and naive cells, the training data was shown as a kernel density distribution. Then, the test data (shown as dots) was overlaid on the training data (Figure 34). By combining supervised learning methods, it was possible to classify activated and naive cells with higher accuracy. (Classification of identical cells in different states (stimulation with PHA)) We investigated whether it is possible to distinguish between the same cell type in different states from the spectrum obtained by the irradiation method according to this embodiment. Jurkat cells (human T cell line) were stimulated to obtain activated cells. The cells and naive cells were analyzed. The cells were also stimulated with PHA (1 μg / ml). After that, the positive rate of the activation marker CD69 was measured with a flow cytometer for the cells after the treatment (FIG. 35). It was confirmed that the positive rate of CD69 in the stimulated cells was almost 100%.

[0055] 100 spectra were obtained from each of the activated cells and naive cells by the irradiation method according to the present embodiment. The spectra were measured in a circular area with a diameter of 10 μm. For each of the activated cells and naive cells, the Raman shift was 600-2980 cm -1 The average of the spectral data in this region is shown in FIG.

[0056] We next applied supervised machine learning to the spectral data to more clearly classify the activation states. Discriminant analysis-based methods Linear Discriminant Analysis (LDA) and Partial Least Squares Regression-Discriminant Analysis (PLS-DA), as well as artificial neural network-based methods Multilayer Perceptron (MLP) and Support Vector Machine (SVM) were tested to classify activated and naive cells with high accuracy.

[0057] For supervised learning, all data of activated and naive cells were randomly split into 80% of training data and 20% of test data. Using the training data, axes for classifying the two cell types using the four classification methods were extracted, and the test data was classified against these defined axes. The false positive ratio (FPR) and true positive rate (TPR) were calculated from the classified labels and original labels of the test data, and the receiver operating characteristic (ROC) curves were constructed from the FPR and TPR. In addition, the area under the ROC curve, which means the classification accuracy, sensitivity, and specificity, and the root mean square error (RMSE) (Table 5) were calculated. Table 5 shows the average values ​​of 10 repetitions of the classification test.

[0058] [Table 5]

[0059] The AUCs of LDA, PLS-DA, SVM, and MLP were 0.94, 0.97, 0.97, and 0.95, respectively, which were all high. Using PLS-DA used as a dimensionality reduction method, two features that contribute to the classification of cell states were extracted as PLS scores. Using scores 1 and 2, which extracted activated and naive cells, the training data was shown as a kernel density distribution. Then, the test data (shown as dots) was overlaid on the training data (Figure 37). By combining supervised learning methods, it was possible to classify activated and naive cells with higher accuracy. (Prediction of the state of Jurkat cells using the irradiation method and machine learning according to this embodiment) We tested whether the state of cells could be inferred using the spectrum measured by the irradiation method according to this embodiment. Jurkat cells (human T cell line) were stimulated with magnetic beads bound to CD3 and CD28 antibodies. It was expected that the cells would be activated through various states during stimulation. After stimulating the cells, they were sampled at 0, 0.5, 3, 6, and 24 hours (FIG. 38). For each sampled cell, the Raman spectrum of a circular area with a diameter of 5 μm was measured.

[0060] Next, we used LDA to distinguish between the activation states of cells. We used all the data of cells as training data to create a classification axis. Then, as test data, we showed a visual classification by the distribution of cells for scores 1 and 2 extracted by LDA (Figure 39). The cells used in the training data were shown as kernel density distributions, and the unknown cells used in the test data were plotted on a line. It was possible to classify cells in different activation stages. (Measurement of multiple cells) By the irradiation method according to this embodiment, multiple cells were measured at once.

[0061] Using Jurkat cells, we verified Raman measurement of living cells using this irradiation method. The measurement area was set wide so that multiple cells were present in the measurement area, and the Raman spectrum of an area with a diameter of 40 μm was measured. In the measurement, the spectrum of a circular area with a diameter of 10 μm was measured for each cell present in the circular area with a diameter of 40 μm (Figure 40(a)). In addition, the spectrum of the cells present in the circular area with a diameter of 40 μm was measured collectively (Figure 40(b)).

[0062] The results are shown in Figure 41. In both cases, the fingerprint area of ​​600-1800 cm -1 , nucleic acid (685, 748, 783 cm -1 ), protein (1003, 1450cm -1 ), lipids (1125, 1580, 1654cm -1 ) It was therefore possible to measure not only individual cells, but also all the cells in a specified area at once. (Measurement of multiple cells) By the irradiation method according to this embodiment, multiple cells were measured at once.

[0063] We verified the Raman measurement of living cells using this irradiation method using human pancreatic cancer cell line (Capan-1) cells and hMSCs. Jurkat cells and hMSCs were cultured (both in an adherent state), and Raman spectra were measured from 10 points in an area with a diameter of 80 μm. In the measurement, the spectrum of the cells present in the area with a diameter of 80 μm was measured all at once (Figure 42).

[0064] The average spectral data for each cell is shown in Figure 43. Capan-1 cells and hMSCs can be classified even when multiple cells are measured collectively. To investigate the discrimination of each cell type, we used the acquired spectra as multivariate data for Principal Component Analysis (PCA), an unsupervised machine learning method. PCA extracted the first two principal components (PCs) that contributed to the variance of the total cells. The contribution of the PCs was 58.98% and 11.68%. In PC1 and PC2, Capan-1 cells and hMSCs showed different cell distribution densities (Figure 44).

[0065] Therefore, it was confirmed that cell types can be classified by measuring all the cells present in a specified area at once. (Measurement of multiple cells) By the irradiation method according to this embodiment, multiple cells were measured at once.

[0066] To investigate whether it is possible to distinguish between the same cell type in different states, we stimulated Jurkat cells (a human T cell line) to obtain activated cells, and also stimulated the cells with PHA (phytohematoglutinin, also known as T cell activating factor) (1 μg / ml).

[0067] The cells activated as described above and naive cells were each cultured at a cell number of 1.75 × 10 7 The cells were seeded at 100 μm / ml in separate wells. Using this irradiation method, the cells in an area of ​​80 μm diameter were collectively measured by spectrum. For each cell, Raman spectra were acquired from a total of 18 areas (9 areas per well) (Figure 45).

[0068] The average spectral data for each cell is shown in Figure 46. Even when multiple cells are measured at the same time, it is possible to classify activated cells and naive cells. To investigate the discrimination of activated and naive cells, we used the acquired spectra as multivariate data for Principal Component Analysis (PCA), an unsupervised machine learning method. PCA extracted the first two principal components (PCs) that contributed to the total cell variance. The contribution of the PCs was 10.21% and 4.16%. According to PC1 and PC2, activated and naive cells showed different cell distribution densities (Figure 47).

[0069] We then applied supervised machine learning to the spectral data to classify cells. We used partial least squares regression-discriminant analysis (PLS-DA). For supervised learning, we randomly split all data from activated and naive cells into 80% training data and 20% testing data.

[0070] Two features that contribute to classifying the cell state were extracted as PLS scores. Using scores 1 and 2, which extracted activated and naive cells, the training data was plotted as a kernel density distribution. The test data (shown as dots) was then overlaid on the training data (Figure 48).

[0071] Therefore, it was confirmed that activated cells and naive cells can be classified by measuring all the cells present in a specified area at once. (Measurement of multiple cells) By the irradiation method according to this embodiment, multiple cells were measured at once.

[0072] We investigated whether it is possible to distinguish between the same cell type in different states. In order to normally maintain and subculture hiPSCs (201B7 strain) in a feeder cell-free state, mTeSR1 medium (modified Tenneille Serum Replacer 1, obtained from STEM CELL) is often used. In this case, hiPSCs form rounded colonies (Figure 49). On the other hand, when cultured in DMEM (Dulbecco's Modified Eagle Medium), which is commonly used for differentiated cells, hiPSCs are unable to maintain normal colony morphology and show abnormal morphology, such as flattening of the outer periphery of the colony (Figure 49).

[0073] Using this irradiation method, Raman spectrum data was obtained from hiPSC colonies with normal morphology (cultured in mTeSR1 medium) and hiPSC colonies with abnormal morphology (cultured in DMEM medium) obtained by the above-mentioned culture. The Raman spectrum was measured for all cells in an area with a diameter of 80 μm (Figure 50). For each cell, Raman spectra were obtained from a total of 20 areas (Figure 51).

[0074] We applied supervised machine learning to the spectral data to classify normal and abnormal cells. We used partial least squares regression-discriminant analysis (PLS-DA). For supervised learning, we randomly split all the data for normal and abnormal cells into 80% training data and 20% test data.

[0075] Two features that contribute to classifying the state of cells were extracted as PLS scores. The training data is shown in a dot plot using scores 1 and 2, which extracted normal and abnormal cells (Figure 52).

[0076] Therefore, it was confirmed that normal and abnormal hiPSCs can be classified by measuring the cells present in a specified area all at once. According to the Raman spectroscopy system, the spectral data acquisition method, and the cell classification method of the present embodiment described above, it is possible to perform laser irradiation more uniformly while further reducing damage to cells.

[0077] Although one embodiment of the present invention has been described above, it goes without saying that the present invention is not limited to the above-described embodiment and may be embodied in various different forms within the scope of the technical concept thereof.

[0078] Moreover, the scope of the present invention is not limited to the exemplary embodiments shown and described, but includes all embodiments that achieve the same effect as the object of the present invention. Furthermore, the scope of the present invention is not limited to the combination of the features of the invention defined by each claim, but may be defined by any desired combination of specific features among all the respective features disclosed. [Explanation of symbols]

[0079] 1. Raman Spectroscopy System 12 DPSS laser light source 14 Spectrometer 16 CCD detector (cooled CCD camera) 18 Inverted Microscope 20 Raman optics 30 2-axis galvanometer mirror 32 Laser Optical Fiber 34 Optical Fiber 38 Shutter Controller 40 Stages 42 Stage Controller 44 Joystick 46 Epi-illumination controller 48 Reflection lighting 50 Computer Equipment 52 Keyboard 54 Mouse 56 Monitor Device 202 Lens 204 ND Filter 205 Shutter 206 Bandpass Filter 208 Flat Mirror 210 Dichroic Mirror 212 High Pass Filter 214 Lens 215 Shutter

Claims

1. A step of receiving Raman dispersed light emitted from a cell by irradiating a region where the cell exists with a laser light in a circular manner; detecting a Raman spectrum according to the received Raman dispersed light; A method for acquiring spectral data comprising:

2. The method of claim 1 , wherein the step of receiving the Raman dispersed light comprises irradiating an area within an individual cell with laser light in a circular fashion.

3. The spectral data acquisition method according to claim 1 , wherein the step of receiving the Raman dispersed light comprises irradiating a region where a plurality of cells are present with laser light in a circular manner.

4. A method for classifying cells, comprising the step of classifying cells based on spectral data acquired by the method according to any one of claims 1 to 3.

5. The method of claim 4 , wherein the step of classifying the cells comprises comparing the acquired spectral data with standard spectral data.

6. The method for classifying cells according to claim 4 or 5, wherein the step of classifying cells is a step of classifying cell types.

7. The method for classifying cells according to claim 4 or 5, wherein the step of classifying cells is a step of classifying cells in different states.

8. The method for classifying cells according to any one of claims 4 to 7, wherein the step of classifying cells is carried out using machine learning.

9. A laser light source that emits laser light; a biaxial galvanometer mirror that reflects the laser light so that the laser light is irradiated in a circular manner onto an area where cells are present; a spectrometer that receives Raman dispersed light of the laser light emitted from the cell; a charge-coupled device (CCD) detector for detecting a Raman spectrum according to the received Raman dispersed light; A Raman spectroscopy system comprising:

10. The Raman spectroscopy system according to claim 9 , wherein the two-axis galvanometer mirror is composed of two galvanometer mirrors arranged on two axes perpendicular to each other.

Images