Imaging flow cytometer, sorting method, and calibration method

Abstract

The present invention provides an imaging flow cytometer with a simple structure and high sorting accuracy that takes into account the speed deviation of an object. The imaging flow cytometer (1) includes: a laser unit (20) that irradiates a first laser and a second laser on a first point (25a) and a second point (25b); first and second imaging units (52a, 52b) that image the first point (25a) and the second point (25b); first and second detectors (32a, 32b) that detect particles (P) passing through the first point (25a) and the second point (25b); a first particle detection unit (70) that issues an imaging timing instruction signal (SG1) to the first and second imaging units (52a, 52b); an image storage unit (82) that receives images (D1) captured by the first and second imaging units (52a, 52b); and a sorting determination unit (84) that determines whether the particles (P) are target particles. The first and second imaging units (52a, 52b) cut out images (D1) of the captured particles (P) based on the imaging timing instruction signal (SG1).

Description

Technical Field

[0001] The present invention relates to an imaging flow cytometer, a sorting method, and a calibration method for sorting objects such as cells using an image. Background Art

[0002] Generally, a flow cytometer for sorting objects such as cells includes: a detection unit that detects signals from the objects; a sorting determination unit that determines whether the objects are specific objects based on the signals obtained by the detection unit; and a sorting unit that sorts the objects based on the determination result of the sorting determination unit (for example, Patent Document 1). In order to improve the sorting accuracy of the objects, it is necessary to make the sorting unit operate in synchronization with the timing when the specific objects reach the sorting unit.

[0003] In existing flow cytometers, the time required for the sorting determination unit to perform signal processing is short, generally within 1 millisecond. In reality, the objects flowing in the flow path have a speed deviation based on the objects. For example, cells of different sizes sometimes have different speeds. In addition, even for cells of the same size, the speed of the cells flowing near the center in the flow path is faster.

[0004] On the other hand, by shortening the distance from the detection unit to the sorting unit, the influence of the speed deviation of the objects can be suppressed. Therefore, in existing flow cytometers, generally, the speed deviation of the objects is not considered, and after the objects pass through the detection unit, the sorting unit is made to operate at a certain delay time.

[0005] On the other hand, in an imaging flow cytometer that uses an image to determine whether an object is a specific object, the signal processing time required to obtain necessary information sometimes becomes long. For example, it sometimes takes about 2 to 32 milliseconds. In this case, it is necessary to extend the distance from the detection unit to the sorting unit. When the sorting unit is made to operate at a certain delay time after the objects pass through the detection unit without considering the speed deviation of the objects, the sorting accuracy is reduced due to the influence of the speed deviation.

[0006] In order to suppress the reduction in sorting accuracy caused by such speed deviation, a structure having a speed measurement unit that measures the speed of each object has been proposed. For example, in Patent Document 1, a speed measurement unit using a grating (diffraction grating) is disclosed. In addition, in Non-Patent Document 1, the following structure is disclosed: laser spots are arranged on the upstream side and the downstream side of the flow path in which the objects flow, and the speed of each object is measured based on the timing when the objects pass through each laser spot.

[0007] In addition, in Patent Document 2, the following method is disclosed: performing time-lapse photography on the objects and calculating their speeds based on the position changes of the objects in each frame, and predicting the timing when each object reaches the sorting unit.

[0008] Prior art documents

[0009] Patent documents

[0010] Patent Document 1: U.S. Patent No. 6,532,061 Gazette

[0011] Patent Document 2: International Publication No. 2011 / 105507

[0012] Non-patent documents

[0013] Non-patent Document 1: N. Nitta et al., Intelligent Image-Activated Cell Sorting. Cell 175, 266-276 (2018)

[0014] In a speed measurement unit using a grating as in Patent Document 1, when multiple objects flow in a flow path in a close state, it is difficult to measure the speed of each object.

[0015] In addition, even if the speed of each object can be measured, in Patent Document 1, the calculation of the speed signal and cell sorting is performed independently, and these results are input to a delay processing circuit to control the sorting unit. Therefore, it is impossible to ensure that each image captured by the camera and the speed of each object measured by the speed measurement unit are accurately and one-to-one correlated. When the processing results and speed information of images from different objects are mis-correlated, it becomes a problem because sorting cannot be performed correctly. Specifically, when objects flow in the flow path at a high throughput, the signal intervals of each object become narrow, and depending on the situation, the signals are repeatedly buried, sometimes causing errors in correlation. Or, when the objects to be measured are small, etc., although they are recognized by the camera but cannot be detected by the speed detection unit, errors may also occur in the one-to-one correspondence of the signals.

[0016] As in Non-patent Document 1, according to the structure using two laser points, even for multiple close objects, the speed of each object can be measured. However, in the structure using two laser points, in order to perform speed detection, in addition to the laser irradiation for cell imaging, two additional laser irradiations for speed detection need to be prepared, and the device scale is large. Furthermore, since the timing of signal emission for speed detection and cell imaging depends on the positions of the respective laser points, as long as the positions of the laser points do not exactly coincide, there will be a timing shift. Calibration for corresponding to this shift is required, and the operation becomes complicated.

[0017] In addition, in order to image or detect particles flowing in a flow cytometer, an objective lens is usually used. However, the field of view obtained from the objective lens is limited. In the method of Non-Patent Document 1, at least two points for velocity measurement and one point for image imaging need to be included within the field of view of the objective lens. However, in order to prevent light leakage and the like, a certain degree of separation is required between the points. Therefore, particularly in the case where multiple points are to be set for image imaging, obtaining the positions of the laser points within the field of view of the objective lens becomes a problem. For example, in the case where it is desired to separately capture fluorescence images obtained when irradiating excitation light of different wavelengths, it is considered to separately set laser points for each excitation light position of different wavelengths. However, in this case, further setting two points for velocity detection becomes a design burden.

[0018] In Patent Document 2, a method is disclosed in which a camera performs time-lapse imaging of cells flowing in a flow path, and reliable cell separation is performed by tracking the cells from this series of images. However, in the method described in Patent Document 2, a large computational load is required to track cells from a series of time-lapse images. Further, in order to track cells, it is necessary to perform imaging under a wide field of view including a cell sorting unit. Therefore, in order to capture cell images with high spatial resolution, the number of pixels in the entire image becomes very large, and the load of information processing becomes high. On the other hand, since reducing the number of pixels sacrifices spatial resolution, there are limitations such as not being able to obtain detailed morphological information of cells. Summary of the Invention

[0019] An object of the present invention is to provide an imaging flow cytometer, a sorting method, and a calibration method that can have a high sorting accuracy considering the velocity deviation of an object with a simple structure.

[0020] A first aspect of the present invention is an imaging flow cytometer having:

[0021] a laser unit that irradiates a first laser and a second laser to a first point and a second point on a flow path where particles flow, respectively;

[0022] a first imaging unit that images the first point;

[0023] a second imaging unit that images the second point;

[0024] a first detector that detects the particles passing through the first point;

[0025] a second detector that detects the particles passing through the second point;

[0026] The first particle detection unit detects the particles based on the signals obtained by the first detector and the second detector, and issues a shooting timing indication signal for instructing the shooting timing to the first imaging unit and the second imaging unit;

[0027] The system time management unit issues the system time to the first particle detection unit;

[0028] The image storage unit receives the images captured by the first imaging unit and the second imaging unit; and

[0029] The sorting judgment unit judges whether the particles are target particles,

[0030] Based on the shooting timing indication signal, the first imaging unit and the second imaging unit cut out (extract, shear) the images of the captured particles.

[0031] The second aspect of the present invention is a sorting method, including:

[0032] The laser unit irradiates the first laser and the second laser to a first point and a second point on the flow path of the particle flow respectively,

[0033] The first detector detects the particles passing through the first point,

[0034] The second detector detects the particles passing through the second point,

[0035] Based on the shooting timing of the signals obtained by the first detector and the second detector, the first imaging unit shoots the first point and the second imaging unit shoots the second point, and cuts out the images of the captured particles,

[0036] The sorting judgment unit judges whether the particles are target particles,

[0037] Based on the timing of the signals obtained by the first detector and the second detector, calculate the delay time until the particles reach the sorting unit,

[0038] According to the judgment result of the sorting judgment unit, the sorting unit sorts the target particles based on the delay time of the particles.

[0039] The third aspect of the present invention is a calibration method, including:

[0040] The laser unit irradiates the first laser and the second laser to a first point and a second point on the flow path of the particle flow respectively,

[0041] The first detector detects the particles passing through the first point,

[0042] The second detector detects the particles passing through the second point.

[0043] Based on the timing of obtaining signals by the first detector and the second detector, calculate the arrival prediction time until the particles pass through the detection position, where the detection position is located downstream of the first point and the second point on the flow path.

[0044] Detect the particles passing through the detection position, and detect the arrival time of the particles at the detection position.

[0045] Adjust the arrival prediction time based on the arrival time.

[0046] Advantages of the Invention

[0047] According to the present invention, it is possible to provide an imaging flow cytometer, a sorting method, and a calibration method that can consider the speed deviation of an object with a simple structure and have high sorting accuracy. Description of the Drawings

[0048] Figure 1 It is a structural diagram of an imaging flow cytometer according to the first embodiment.

[0049] Figure 2 (a) is an enlarged view of the flow path around the point. Figure 2 (b) is a graph of the laser intensity.

[0050] Figure 3 It is a structural diagram of the imaging unit.

[0051] Figure 4 It is a structural diagram of an imaging flow cytometer according to the second embodiment.

[0052] Figure 5 It is a diagram showing an example of the calibration method.

[0053] Description of the Reference Numerals

[0054] 1 Imaging flow cytometer

[0055] 10 Flow path

[0056] 20 Laser unit

[0057] 22a First laser light source

[0058] 22b Second laser light source

[0059] 22c Third laser light source

[0060] 25a First point

[0061] 25b Second point

[0062] 25c Third point

[0063] 25d Fourth point

[0064] 26 Laser control unit

[0065] 30 Detection unit

[0066] 32a First detector

[0067] 32b Second detector

[0068] 32c Third detector

[0069] 34 Fourth detector

[0070] 40 Illumination light source

[0071] 50 Imaging unit

[0072] 52a First imaging section

[0073] 52b Second imaging section

[0074] 52c Third imaging section

[0075] 52d Fourth imaging section

[0076] 60 System time management unit

[0077] 70 First particle detection unit

[0078] 72 Second particle detection unit

[0079] 80 Delay time calculation unit

[0080] 81 Image analysis unit

[0081] 82 Image storage unit

[0082] 84 Sorting judgment unit

[0083] 86 Sorting signal control unit

[0084] 88 Storage unit

[0085] 90 Sorting unit

[0086] 100 APD

[0087] 102 AFE

[0088] 104 ADC

[0089] 106 FPGA

[0090] 108 GbE

[0091] 110 APD Control High-Voltage Power Supply

[0092] P particle

[0093] SG1 Camera Timing Indication Signal

[0094] D1 Image Specific Embodiment

[0095] <First Embodiment>

[0096] Hereinafter, the imaging flow cytometer 1 of the first embodiment will be described with reference to the drawings. Figure 1 It is a structural diagram of the imaging flow cytometer 1 of this embodiment. The imaging flow cytometer 1 of this embodiment determines whether the particle P flowing in the flow path 10 is an object particle to be sorted, and sorts the object particles. The particle P is, for example, a droplet (beads), a cell, and a cell mass (blood cell, bone marrow cell, lymphocyte, circulating cancer cell, vascular endothelial cell, platelet, platelet aggregate, egg, sperm, fertilized egg, spheroid cell, organoid, etc.), an organelle (chromosome, chloroplast, mitochondrion, etc.), a microorganism, a parasite, pollen, an alga (Chlamydomonas, Euglena, etc.). A plurality of particles P flow in a certain direction (axial direction) in the flow path 10.

[0097] The imaging flow cytometer 1 mainly includes a laser unit 20, a detection unit 30, a first particle detection unit 70, a camera unit 50, a system time management unit 60, and an image analysis unit 81.

[0098] The laser unit 20 has at least two laser light sources. The laser unit 20 of this embodiment has a first laser light source 22a, a second laser light source 22b, and a third laser light source 22c. Each laser light source 22a, 22b, 22c preferably irradiates laser light of different wavelengths from each other.

[0099] The first laser light source 22a irradiates the first laser on the first point 25a on the flow path 10. The second laser light source 22b irradiates the second laser on the second point 25b on the flow path 10. The third laser light source 22c irradiates the third laser on the third point 25c on the flow path 10. Figure 2 (a) shows an enlarged view of the flow path 10 around the point.

[0100] Here, the inner diameter W1 of the flow path 10 is about 200 μm. Each point 25a to 25c extends in a direction orthogonal to the axis of the flow path 10. The width of the points 25a to 25c in the direction orthogonal to the axis of the flow path 10 is, for example, 80 μm to 100 μm. The diameter of the particle P is, for example, 20 μm.

[0101] Each of the points 25a to 25c is located at a position separated from each other in the axial direction of the flow path 10. The points 25a to 25c may also be arranged at equal intervals. The arrangement of the points 25a to 25c will be described later.

[0102] Figure 2 (b) shows an example of a graph of the intensity of the laser. The laser preferably has a certain intensity in a direction orthogonal to the axis of the flow path 10.

[0103] The detection unit 30 has at least two detectors. The detection unit 30 of the present embodiment has a first detector 32a, a second detector 32b, and a third detector 32c. The first detector 32a detects the scattered light emitted when the particle P passes through the first point 25a. Similarly, the second and third detectors 32b and 32c respectively detect the scattered light emitted when the particle P passes through the second and third points 25b and 25c.

[0104] Here, the distance L1 between the first point 25a and the second point 25b is determined by the distance between the line sensors (line sensors, row sensors, linear sensors) 100 on the first imaging unit 52a and the line sensors 100 on the second imaging unit 52b and the optical magnification of the imaging optical system described later, and thus can be calculated based on these design values. Therefore, the velocity of the particle P can be calculated for each particle based on the timing information of the particle P passing through the first point 25a and the second point 25b.

[0105] The first particle detection unit 70 detects the timing of the particle P passing through each of the points 25a to 25c based on the scattered light detected by the detection unit 30. Here, as described later, the system time is sent from the system time management unit 60 to the first particle detection unit 70. Therefore, the first particle detection unit 70 obtains the system time of the timing of the particle P passing through each of the points 25a to 25c.

[0106] The first particle detection unit 70 assigns an identification number to each detected particle P in units of particles. In addition, the first particle detection unit 70 outputs the system time of the timing of the particle P passing through each of the points 25a to 25c together with the identification number assigned to the particle P to a delay time calculation unit 80 described later.

[0107] The first particle detection unit 70 sends the imaging timing instruction signal SG1 together with the identification number of each particle P to the imaging unit 50.

[0108] The imaging unit 50 has at least two imaging parts. The imaging unit 50 of the present embodiment has a first imaging part 52a, a second imaging part 52b, a third imaging part 52c, and a fourth imaging part 52d. The first imaging part 52a captures an image of the first point 25a. Similarly, the second and third imaging parts 52b and 52c respectively capture images of the second and third points 25b and 25c. Regarding the fourth imaging part 52d, it will be described later.

[0109] An imaging optical system including a lens is disposed between the first to fourth imaging parts 52a to 52d and the first to fourth points 25a to 25d on the flow path 10. The imaging optical system is configured such that the images on the first to fourth points 25a to 25d are respectively imaged on the line sensors 100 of the first to fourth imaging parts 52a to 52d. Here, the width of the points 25a to 25d is, for example, 80 μm to 100 μm. The width of the line sensors 100 of the first to fourth imaging parts 52a to 52d is, for example, 10 mm to 80 mm. The magnification of the imaging optical system is, for example, 100 times to 1000 times.

[0110] Each of the imaging parts 52a to 52d extracts an image D1 of the particle P based on the imaging timing instruction signal SG1 from the first particle detection unit 70. In addition, each of the imaging parts 52a to 52d sends the extracted image D1 of the particle P and the identification number of the particle P to the image storage unit 82 described later.

[0111] Here, the configuration of each of the points 25a to 25c will be described. The first to third imaging parts 52a to 52c, the imaging optical system, and the flow path 10 are preferably fixed such that the images of the particles P flowing in the flow path 10 are respectively imaged on the line sensors 100 (described later) of the first to third imaging parts 52a to 52c. In this case, the positions of each of the points 25a to 25c are finely adjusted so that the images of the particles P passing through each of the points 25a to 25c are imaged on the line sensors 100 of the first to third imaging parts 52a to 52c. Thus, the distance between each of the points 25a to 25c is determined by the distance between each of the line sensors 100 of the first to third imaging parts 52a to 52c and the optical magnification of the imaging optical system. Thus, the moving speed of each particle P can be calculated based on the measured value of the time difference when the particles P flowing in the flow path 10 pass through each of the points 25a to 25c. Or, conversely, when the moving speed of the particles P flowing in the flow path 10 is known, the time difference when the particle P passes through each of the points 25a to 25c can be calculated.

[0112] For example, mirrors capable of adjusting the angle with respect to each of the first to third lasers are respectively provided between the laser unit 22 and the flow path 10. By adjusting the angle of the mirror, the positions of each of the points 25a to 25c can be adjusted.

[0113] The first to fourth imaging units 52a to 52d each have the same structure. Hereinafter, with reference to Figure 3 the structure of each imaging unit will be described. Figure 3 An example of the structure of the first imaging unit 52a is shown. The first imaging unit 52a includes a line sensor 100. The line sensor 100 is, for example, an avalanche photodiode array (APD array), a photodiode array (PD array), or a photomultiplier tube array (PMT array). The line sensor 100 outputs a signal current or a signal voltage from each element according to the received image. This output is parallel for all pixels, but it is also possible to switch the outputs of multiple pixels and connect them to the subsequent stage in a time-division manner.

[0114] Figure 3 In the example shown, the line sensor 100 is an APD array. The line sensor 100 extends in the direction in which the first point 25a extends. In the case of using an APD array, the APD control high-voltage power supply 110 controls the applied voltage and controls it so that the output is constant with respect to changes in gain characteristics and the like caused by temperature changes and the like. In the case of using a light-receiving element other than an APD array, a control circuit for controlling the bias voltage, bias current, gain, etc. of the connected control element is provided.

[0115] The current or voltage output from the line sensor 100 is amplified in the analog front end (AFE) circuit 102 and converted into a voltage capable of AD conversion. If necessary, a transimpedance amplifier (TIA) circuit that converts the current into a voltage and amplifies it may also be connected between the APD array 100 and the analog front end circuit 102. The analog front end circuit 102 is connected in parallel to the outputs of all pixels of the APD array 100, but it is also possible to switch the outputs of multiple pixels and connect them to the subsequent stage in a time-division manner.

[0116] The output of the analog front end circuit 102 is input to the AD converter 104 as a voltage signal. The input to the AD converter 104 is connected in parallel to the outputs of all pixels of the analog front end circuit 102, and it is also possible to switch the outputs of multiple analog front end circuits 102 and connect them to the subsequent AD converter 104 in a time-division manner.

[0117] When the AD converter 104 is connected in parallel to the outputs of all pixels of the analog front end circuit 102, the sampling rate of the AD converter 104 is, for example, when the speed of the particle P passing through the flow path is V m / s. When the size of one pixel in the image in the flow direction of the particle is set to X μm, it is set to V / X × 10 6 samples / s or more.

[0118] The line image converted into a digital signal by the AD converter 104 is input to the FPGA 106. The FPGA 106 can be an FPGA (Field Programmable Gate Array) equipped with an internal memory serving as an image buffer, a structure combining the FPGA and an external memory such as a DRAM (Dynamic Random Access Memory), a structure replacing the FPGA with a dedicated or general-purpose logic circuit, or a computer system combining a microprocessor and a memory.

[0119] The line image stores a certain amount of data for a moment in the FPGA 106. When the imaging timing signal SG1 is input, the image obtained at a certain time in the previously specified periphery is cut out according to the timing specified by the imaging timing signal SG1, and the time information corresponding to the timing signal SG1 or the identification number of the captured particle P is added and output. At this time, in the FPGA 106, two or more lines of images can also be processed by operations such as accumulation to form a one-line image, extend the bit length of the data, expand the dynamic range, and output. In addition, the image can also be compressed to reduce the data volume. In the compression of data, in addition to variable-length coding, the redundancy in the time direction or spatial direction of the image data can also be utilized. In addition, the image obtained here can also be analyzed, the characteristic quantities corresponding to the structure and shape of the cell can be calculated, etc., and the information attached to the image or the image can be output as independent information together with the identification number.

[0120] The cut-out image D1 or image information is output as a data packet from the network interface 108 to the image storage unit 82 (described later) via the network. The network interface 108 is, for example, a network interface with different bandwidths such as 1 Gb Ethernet (GbE), 10 Gb Ethernet, etc. Alternatively, the cut-out image D1 or image information is output to the image storage unit 82 through a high-speed digital interface such as PCIe, high-speed LVDS, or a general-purpose interface such as USB.

[0121] In addition, a cylindrical lens or a lens array can be arranged in front of the line sensor 100. Thereby, the light collection efficiency can be improved. In addition, a band-pass filter can be arranged in front of the line sensor 100. Thereby, the specificity can be improved. In addition, a slit can be arranged in front of the line sensor 100. Thereby, the spatial resolution in the flow direction can be improved.

[0122] The system time management unit 60 is composed of, for example, a system clock that changes its value at regular intervals and a counter that counts the system clock. The system time management unit 60 maintains the system time of the entire system and issues the system time. The system time issued by the system time management unit 60 is sent to the first particle detection unit 70, the second particle detection unit 72 described later, and the sorting signal control unit 86 described later.

[0123] The image analysis unit 81 includes an image storage unit 82 and a sorting determination unit 84. The image storage unit 82 receives the image D1 captured by the imaging unit 50 together with the identification number of each particle P. The image storage unit 82 may be a network interface or a general interface that inputs the image D1, an FPGA equipped with an internal memory serving as an image buffer, a structure combining the FPGA and an external memory such as a DRAM, a structure replacing the FPGA with a dedicated or general logic circuit, or a computer system combining a microprocessor and a memory.

[0124] The sorting determination unit 84 determines whether the particle P is a target particle. Specifically, the sorting determination unit 84 determines whether the particle P is a target particle by analyzing the image of the particle P output from the image storage unit 82. Then, the sorting determination unit 84 outputs the determination result together with the identification number of the particle P to the sorting signal control unit 86 and the storage unit 88 described later.

[0125] The sorting determination unit 84 can use various image analysis methods according to the type of the target particle. For example, a method of calculating a feature quantity of 1 or more from the image and classifying the particle P according to its distribution, a method of classifying the distribution of the feature quantity by machine learning such as SVM (Support Vector Machine), a method of classifying by applying deep learning to the image, etc.

[0126] In addition, as described in U.S. Patent No. 6,211,955, the sorting determination unit 84 may also use a method of classifying according to the presence or absence of bright spots in the fluorescence image inside the cell nucleus.

[0127] The imaging flow cytometer 1 of the present embodiment further includes a laser control unit 26, a fourth detector 34, an illumination light source 40, a second particle detection unit 72, a delay time calculation unit 80, a sorting signal control unit 86, a storage unit 88, and a sorting unit 90.

[0128] The illumination light source 40 is, for example, a light-emitting diode (LED). The illumination light source 40 is disposed on the side opposite to the imaging unit 50 with the flow path 10 therebetween. The light from the illumination light source 40 is irradiated onto the fourth point 25d on the flow path 10 through an illumination optical system provided between the illumination light source 40 and the flow path 10. The illumination optical system includes, for example, one or more lenses. The illumination optical system can use, for example, Köhler illumination or critical illumination. Preferably, by disposing a slit or a cylindrical lens on the illumination optical system, the fourth point 25d is deformed so as to extend in a direction orthogonal to the axis of the flow path 10.

[0129] The fourth imaging unit 52d images the fourth point 25d. Here, the fourth imaging unit 52d, the imaging optical system, and the flow path 10 are preferably fixed such that the images of the particles P flowing in the flow path 10 are respectively formed on the line sensor 100 of the fourth imaging unit 52d. In this case, the position of the fourth point 25d is finely adjusted so that the particles irradiated at the fourth point 25d are imaged on the fourth imaging unit 52d. When adjusting in this way, the distances between the first to third points 25a to 25c and the fourth point 25d are determined by the distances between the respective line sensors 100 on the first to fourth imaging units 52a to 52d and the optical magnification of the imaging optical system. Utilizing this situation, by measuring the timing when the particles P flowing in the flow path 10 pass through each point, the moving speed of each particle P can be calculated. Or, conversely, when the moving speed of the particles P flowing in the flow path 10 is known, the timing difference when the particle P passes through each point 25a to 25d can be calculated. For example, an optical element such as a mirror capable of angular adjustment or a slit or lens capable of position adjustment is provided on the optical path of the illumination optical system. By adjusting the angles and positions of these optical elements, the position of the fourth point 25d can be adjusted. Or, the position of the illumination light source 40 and the entire illumination optical system can also be adjusted.

[0130] The fourth detector 34 detects the particles P flowing through the detection position X on the flow path 10. The fourth detector 34 is, for example, a camera, a strobe camera, a laser and a detector. The camera or the strobe camera acquires an image of the detection position X and detects the passage of the particles P from the image. In the laser and detector, the laser is made incident on the detector, the signal of the detector at the detection position X is acquired, and the passage of the particles P is detected based on the change in the signal waveform when the particles P pass through the detection position X. The detection position X is preferably near the sorting position. The fourth detector 34 outputs a detection signal to the second particle detection unit 72.

[0131] The detection position X is located downstream of the first to fourth points 25a to 25d. The detection position X is preferably coincident with the sorting unit 90, but the detection position X does not necessarily have to be coincident with the sorting unit 90. For example, the detection position X may be slightly upstream or downstream of the sorting unit 90.

[0132] The second particle detection unit 72 detects the timing when the particle P passes through the detection position X based on the scattered light or fluorescence detected by the fourth detector 34. In addition, the second particle detection unit 72 receives the system time sent from the system time management unit 60. Then, the second particle detection unit 72 stores the system time (arrival time) at the timing when the detection signal is detected in the fourth detector 34 together with the detection signal. That is, a time stamp is given to each detection signal.

[0133] The second particle detection unit 72 processes the input of the detection signal from the fourth detector 34 and the input of the system time from the system time management unit 60 in real time.

[0134] When the fourth detector 34 acquires an image, the second particle detection unit 72 may be an FPGA having an input of an image from a camera and a system time, and having an internal memory serving as an image buffer, a structure combining the FPGA and an external memory such as a DRAM, a structure replacing the FPGA with a dedicated or general logic circuit, or a computer system combining a microprocessor and a memory. For example, when it is known from analyzing the image that the particle P is included in the image, the system time input at the time when the image is acquired can be detected as the passing time of the particle P. When the fourth detector 34 acquires a signal continuous in the time direction like a detector, the second particle detection unit 72 can be composed of an ADC (Analog Digital Converter) that converts the detection signal into a digital signal, an FPGA having an internal memory serving as a detection signal buffer and a processing circuit. Regarding the detection signal buffer and the processing circuit, it may also be a structure combining the FPGA and an external memory such as a DRAM, a structure replacing the FPGA with a dedicated or general logic circuit, or a computer system combining a microprocessor and a memory. For example, the particle P can be detected by a method such as detecting the time when the change of the signal exceeds a certain threshold as the passing time of the particle P.

[0135] The second particle detection unit 72 outputs the system time (arrival time) given to the detection signal corresponding to the passing of the particle P to the delay time calculation unit 80.

[0136] The laser control unit 26 includes a control circuit and a drive circuit for controlling the laser. The laser control unit 26 controls the on / off of the first to third laser light sources 22a to 22c of the laser unit 20 corresponding to the respective points 25a to 25c used for imaging, and the intensities of the first to third lasers.

[0137] The delay time calculation unit 80 calculates the delay time until the particle P reaches the sorting unit 90 based on the system times when the particle P detected by the first particle detection unit 70 passes through the first point 25a and the second point 25b. Specifically, when the system time when the particle P passes through the first point 25a is set as t1, the system time when the particle P passes through the second point 25b is set as t2, the distance between the first point 25a and the second point 25b is set as L1, and the distance between the first point 25a and the sorting unit 90 is set as L2, the delay time ΔT is obtained as ΔT = (L2 / L1) × (t2 - t1). At this time, when the values of L1 and L2 are both known, their values can also be used. As described above, since L1 can be obtained based on the design value, when L2 is uncertain, the value of the coefficient A = L2 / L1 can also be obtained by calibration.

[0138] In addition, the above delay time ΔT is the time from when the particle P passes through the first point 25a to when the particle P reaches the sorting unit 90, but it is not limited to this. The delay time ΔT can also be the time from when the particle P passes through the second point 25b to when it reaches the sorting unit 90.

[0139] The delay time calculation unit 80 outputs the delay time of the particle P and the identification number of the particle P to the sorting signal control unit 86.

[0140] The sorting signal control unit 86 issues a necessary sorting signal for sorting the particle P at the timing calculated based on the delay time of the particle P calculated by the delay time calculation unit 80 according to the judgment result of the sorting judgment unit 84.

[0141] Since the system time is sent from the system time management unit 60 to the sorting signal control unit 86, the sorting signal control unit 86 issues an appropriate sorting signal to the sorting unit 90 in accordance with the timing when the particle P reaches the sorting unit 90.

[0142] When issuing the sorting signal, if the timing calculated based on the delay time is later than the system time, the sorting is successful at the calculated timing, and the sorting signal is output. In the case where the timing calculated based on the delay time is earlier than the system time, the sorting is unsuccessful, and the sorting signal may not be output. Whether the sorting is successful or not is output from the sorting signal control unit 86 to the storage unit 88.

[0143] The sorting signal is, for example, a pulse signal. The pulse signal is emitted in synchronization with the timing of the target particles arriving at the sorting unit 90 and is transmitted to the sorting unit 90. Alternatively, a signal encoding the timing information can be transmitted using a separately prepared signal line.

[0144] In addition, the sorting unit 90 sometimes sorts the target particles into two or more types. For example, there are cases of sorting into two types (2-way sorting) and sorting into four types (4-way sorting). In such cases, the sorting determination unit 84 further classifies the target particles into multiple different types, and the classification results are included in the sorting signal. In this case, for example, the amplitude, sign, etc. of the pulse signal can be used to make the pulse signal contain the classification results.

[0145] Alternatively, a separately prepared signal line consistent with the pulse signal can be used to transmit the classification results to the sorting unit 90 as a digital signal or the like.

[0146] In addition, the sorting signal can also transmit the time width for continuously sorting with the sorting timing as the start time. In this case, the time width for continuously sorting can be transmitted as the pulse width of the pulse signal, or a separately prepared signal line can be used to transmit a signal encoding the time width for continuously sorting.

[0147] In addition, the system time when the particle P arrives at the sorting unit 90 can also be calculated by either the delay time calculation unit 80 or the sorting signal control unit 86.

[0148] It is also possible to make the sorting signal control unit 86 emit a sorting signal only when the particle P is a target particle according to the judgment result of the sorting determination unit 84. In addition, it is also possible to make the sorting signal control unit 86 emit different sorting signals when the particle P is a target particle and when it is not a target particle according to the judgment result of the sorting determination unit 84.

[0149] The storage unit 88 stores the judgment result of the particle P received from the sorting determination unit 84 together with the identification number of the particle P. In addition, the storage unit 88 stores whether the sorting signal control unit 86 can emit a sorting signal at a specified timing, the sorting result of the particle P, and the judgment result together according to the judgment result sent from the sorting determination unit 84.

[0150] The sorting unit 90 sorts the target particles based on the sorting signal sent from the sorting signal control unit 90. The sorting unit 90 uses the information of the sorting window received from the sorting signal control unit 86 or stored inside the sorting unit 90 (described later) based on the timing indicated by the sorting signal. Figure 5Y1 and Y2) to control the switching of the sorting direction. The sorting unit 90 is, for example, a droplet sorter or an on-chip sorter.

[0151] The sorting unit 90 is, for example, the droplet sorter disclosed in M. J. Fulwyler, Science 150, 910 - 911 (1965). In this case, the sorting unit 90 has a nozzle disposed at the front end of the flow path 10, and forms droplets at a certain cycle through the nozzle. By applying an electric field to the liquid near the target particles at the timing of forming the droplets containing the target particles, the droplets containing the target particles are selectively charged. By applying an electrostatic field near the trajectory of the falling droplets, it is possible to selectively shift only the falling trajectory of the droplets containing the charged target particles. By providing a collection container at the falling point, the target particles can be sorted. Furthermore, by controlling the positive and negative of the charge and the amount of charge of the charged droplets, the trajectories of the droplets are assigned to different directions and angles, and it is also possible to collect two or more types of cells in two or more separate containers.

[0152] The sorting unit 90 may also be, for example, the on-chip sorter disclosed in Non-Patent Document 1. In this case, by arranging a dual-membrane pump in a form spanning the flow path 10, at the timing when the target particles pass through, a local flow is generated in a direction orthogonal to the flow path 10 to shift the trajectory of the target particles, and the target particles are sorted by being distributed to flow paths in different directions in the downstream branch flow paths. The dual-membrane pump is described, for example, in S. Sakuma et al., Lab on a Chip 17, 2760 - 2767, 2017.

[0153] Hereinafter, a sorting method using the imaging flow cytometer 1 of the present embodiment will be described. First, a plurality of particles P flow in the flow path 10. The laser unit 20 irradiates the first to third points 25a to 25c with the first to third lasers, respectively. The detection unit 30 detects the particles P passing through the first to third points 25a to 25c. The first particle detection unit 70 detects the particles P based on the signals obtained by the detection unit 30. In addition, the first particle detection unit 70 issues a camera timing indication signal SG1 for instructing the camera timing to the imaging unit 50. During this period, the system time management unit 60 issues the system time to the first particle detection unit 70.

[0154] The imaging unit 50 captures images of the first to fourth points 25a to 25d. In addition, the imaging unit 50 extracts an image D1 of the captured particle P based on the imaging timing instruction signal SG1. The image storage unit 82 receives the image D1 of the particle P captured by the imaging unit 50. The sorting determination unit 84 determines whether the particle P is a target particle.

[0155] The sorting signal control unit 86 issues a sorting signal according to the determination result of the sorting determination unit 84 to sort the particle P in accordance with the timing calculated by the delay time calculation unit 80. The sorting unit 90 sorts the target particles based on the sorting signal.

[0156] The delay time calculation unit 80 calculates the delay time until the particle P reaches the sorting unit 90 based on the system time of the timing when the detection unit 30 obtains the signal.

[0157] As described above, according to the imaging flow cytometer 1 and sorting method of the present embodiment, it is possible to improve the sorting accuracy by considering the velocity deviation of the particle P with a simpler structure than in the past.

[0158] <Second Embodiment>

[0159] Hereinafter, the imaging flow cytometer 1 of the second embodiment will be described. The structures of the fourth detector 34, the second particle detection unit 72, and the delay time calculation unit 80 of the imaging flow cytometer 1 of the present embodiment are different from those of the first embodiment. The other structures are the same as those of the first embodiment. In the present embodiment, the detection position X is located downstream of the sorting unit 90, and it is possible to detect the change in the position that occurs when the particle P is sorted by the sorting unit 90.

[0160] The fourth detector 34 is, for example, an area sensor such as a CCD camera or a CMOS camera. In this case, the fourth laser is irradiated to the detection position X, and after arranging the components so that the scattered light or fluorescence emitted by irradiating the particle P passing through the detection position X with the laser is incident on the fourth detector 34, the signal of the fourth detector 34 is obtained to obtain information about the position where the particle P passes. The fourth detector 34 outputs the detection signal to the second particle detection unit 72. Instead of the laser, it is also possible to use an illumination such as a flash for bright field detection.

[0161] The second particle detection unit 72 obtains the ratio of the particles sorted or not sorted in the sorting unit based on the passing position of the particle P, and outputs this value to the delay time calculation unit 80. Here, the fourth detector 34 does not need to obtain the passing position information for each particle P, and can also obtain it as an accumulated value for a plurality of particles P. For example, when the positions X1 of the particles P sorted in the sorting unit 90 and the positions X2 of the particles P not sorted in the detection position X are detected by the fourth detector 34 respectively, consider obtaining the ratio of the particles passing through the positions X1 and X2 as the accumulation of a plurality of particles. For example, when the particle P can be detected by the fluorescence emitted by laser excitation, by measuring the ratio of the cumulative fluorescence intensity values of the particle P from the positions X1 and X2, the ratio of the sorted particle P and the unsorted particle P can be investigated.

[0162] The delay time calculation unit 80 evaluates the appropriateness of the delay time calculation method by obtaining the ratio of the sorted particle P or the unsorted particle P from the second particle detection unit 72. For example, in the case of the setting where all the particles P are sorted, by adjusting the delay time so that the ratio of the sorted particle P is closest to 100%, the appropriate delay time during sorting can be determined. In addition, it is also possible to evaluate the appropriateness of the delay time calculation method while changing the delay time calculation method, explore the most appropriate delay time calculation method, and perform calibration of the delay time calculation method.

[0163] <Third Embodiment>

[0164] Hereinafter, the imaging flow cytometer 1 of the third embodiment will be described. The imaging flow cytometer 1 of the present embodiment further has a calibration function for the delay time in the imaging flow cytometer 1 of the first embodiment. Specifically, in the imaging flow cytometer 1, a calibration droplet (particle P) is caused to flow through the flow path 10, and the arrival time when the particle P passing through the first point 25a and the second point 25b is detected at the detection position X is measured. Thereby, calibration of the arrival prediction time of the particle P calculated based on the particle P passing through the first point 25a and the second point 25b is performed.

[0165] In the imaging flow cytometer of the present embodiment, the structure of the delay time calculation unit 80 is different from that of the first embodiment. The other structures are the same as those of the first embodiment.

[0166] The delay time calculation unit 80 of the present embodiment calculates the time (arrival prediction time) when the particle P arrives at the detection position X based on the system times t1 and t2 detected by the first particle detection unit 70 when the particle P passes through the first point 25a and the second point 25b. In addition, the distance between the detection position X and the sorting unit 90 is known.

[0167] The delay time calculation unit 80 associates the identification number of the particle P with the arrival time of the particle P output from the second particle detection unit 72. Specifically, the delay time calculation unit 36 associates the identification number of the particle P based on the arrival prediction time of the particle P from the first particle detection unit 70 and the arrival time of the particle P from the second particle detection unit 72.

[0168] The delay time calculation unit 80 outputs the identification number, arrival prediction time, and arrival time of the associated particle P in a group.

[0169] The delay time calculation unit 80 performs calibration based on the arrival prediction time and arrival time of the particle P. Figure 5 An example of the calibration method is shown. The horizontal axis represents the point passage time ΔT calculated by the delay time calculation unit 80. Here, when the times (system times) at which the particle P passes through the first point 25a and the second point 25b are set as t1 and t2, the point passage time ΔT becomes (t2 - t1). The vertical axis represents the arrival time T of the particle P.

[0170] When the distance between the first point 25a and the second point 25b is set as L1, and the distance between the first point 25a and the detection position X is set as L2, in the case where the particle P moves at a constant speed, the arrival prediction time is obtained as (L2 / L1) × (t2 - t1). Here, the value of the coefficient A = L2 / L1 is obtained through calibration, so that the arrival prediction time can be obtained when t1 and t2 are given.

[0171] The delay time calculation unit 80 plots the point passage time ΔT and the arrival time T of the particle P, and fits them in a model formula. The model formula is, for example, a straight line passing through the origin as shown, but is not limited thereto. In the case where the straight line passing through the origin is set as the model formula, the model formula becomes T = A × ΔT. The delay time calculation unit 80 obtains the coefficient A based on the plot of the point passage time ΔT and the arrival time T of the particle P, and thus performs calibration. Figure 5

[0172] Figure 5 In addition, the delay time calculation unit 80 may calibrate the sorting window instead of, or in addition to, the arrival prediction time of the particle P. Here, the sorting window represents the time period during which the sorting unit 90 performs sorting. In Figure 5In the example shown, when the passing time of the point of particle P is ΔT1, it means sorting is performed during the time period of sorting window Y1 that includes the arrival time T1. Additionally, when the passing time of the point of particle P is ΔT2, it means sorting is performed during the time period of sorting window Y2 that includes the arrival time T2. The width (length of time) of the sorting window can be constant regardless of the delay time of particle P, or can be variable depending on the delay time of particle P. Generally, when the speed of particle P is high, the passing time of the point becomes shorter and the delay time becomes smaller. Conversely, when the speed of particle P is low, the passing time of the point becomes longer and the delay time becomes larger. In Figure 5 the example shown, the longer the passing time of the point of particle P, the larger the delay time, and the wider the width of the sorting window. This is because, the larger the delay time, the larger the error, so by expanding the width of the sorting window, the sorting accuracy can be improved.

[0173] The sorting window does not have to be centered on the measured arrival time T. The delay time calculation unit 80 can set the sorting window based on the graph of the passing time ΔT of the particle and the arrival time T.

[0174] Preferably, the detection position X coincides with the sorting unit 90, but the detection position X does not have to coincide with the sorting unit 90. For example, the detection position X can also be slightly upstream or downstream of the sorting unit 90. In this case, in accordance with the offset between the detection position X and the position of the sorting unit 90, the center position of the sorting window can be appropriately offset from the arrival prediction time and adjusted to set the sorting window at the timing when particle P arrives at the sorting unit 90. Additionally, due to the physical operation speed limit of the sorting unit 90, there may be a time lag from when the sorting signal is sent to the sorting unit 90 until particle P is sorted. When this time lag is set as B, it can also be adjusted so that the sorting signal generation timing is earlier by the amount of the time lag B in such a way that sorting is performed at the timing when particle P arrives at the sorting unit 90.

[0175] Hereinafter, the calibration method and sorting method using the imaging flow cytometer 1 of the present embodiment will be described.

[0176] First, a plurality of particles P flow in the flow path 10. The laser unit 20 irradiates the first to third lasers to the first to third points 25a to 25c respectively. The detection unit 30 detects the particles P passing through the first to third points 25a to 25c. The first particle detection unit 70 detects the particles P based on the signals obtained by the detection unit 30. Additionally, the first particle detection unit 70 issues a camera timing indication signal SG1 for indicating the camera timing to the imaging unit 50. During this period, the system time management unit 60 issues the system time to the first particle detection unit 70.

[0177] The second particle detection unit 72 detects the timing when the particle P passes through the detection position X based on the signal obtained by the fourth detector 34.

[0178] The delay time calculation unit 80 calculates the arrival prediction time of the particle P at the detection position X based on the time (system time) when the particle P passes through the first point 25a and the second point 25b.

[0179] The second particle detection unit 72 detects the detection signal corresponding to the passage of the particle P and outputs the arrival time of the particle P at the detection position X to the delay time calculation unit 80.

[0180] The delay time calculation unit 80 calibrates and adjusts the calculation formula of the arrival prediction time based on the arrival time. Then, the sorting signal control unit 86 and the sorting unit 90 adjust the timing of sorting the particle P based on the adjusted arrival prediction time.

[0181] As described above, according to the imaging flow cytometer, sorting method, and calibration method of the present embodiment, the sorting accuracy can be further improved.

[0182] <Fourth Embodiment>

[0183] Hereinafter, the imaging flow cytometer 1 of the fourth embodiment will be described. The imaging flow cytometer 1 of the present embodiment further has a calibration function for the delay time in the imaging flow cytometer 1 of the second embodiment. Specifically, in the imaging flow cytometer 1, the calibration droplet (particle P) is caused to flow through the flow path 10, and the arrival time when the particle P passing through the first point 25a and the second point 25b is detected at the detection position X is measured. Thus, the calibration of the arrival prediction time of the particle P calculated based on the particle P passing through the first point 25a and the second point 25b is performed.

[0184] The structures of the fourth detector 34, the second particle detection unit 72, and the delay time calculation unit 80 of the imaging flow cytometer of the present embodiment are different from those of the third embodiment. The other structures are the same as those of the third embodiment. In addition, the structure of the delay time calculation unit 80 of the imaging flow cytometer of the present embodiment is different from that of the second embodiment, and the other structures are the same as those of the second embodiment. In the present embodiment, the detection position X is located downstream of the sorting unit 90, and the change in the position that occurs when the particle P is sorted by the sorting unit 90 can be detected. Therefore, at the timing when the particle P passes through the sorting unit 90, a sorting signal is sent to the sorting unit 90. Then, by evaluating the ratio of the position shift caused by the sorting of the particle P, the timing of the sorting signal sent to the sorting unit 90 can be evaluated. Taking advantage of this situation, in the present embodiment, while changing the delay time from when the particle P is detected until the sorting signal is sent, the position change based on sorting at the detection position X is evaluated, and thus calibration for obtaining the optimal delay time is performed.

[0185] In this embodiment, the second particle detector 72 calculates the sorting success rate of the particle P based on the position information of the particle P flowing in the flow path 10, and outputs the determination result to the delay time calculation unit 80. In addition, it is not necessary for the second particle detector 72 to receive the system time from the system time management unit 60.

[0186] The delay time calculation unit 80 performs calibration based on the predicted arrival time of the particle P at the sorting unit 90 and the sorting result determination of the particle P. When the times (system times) of the particle P passing through the first point 25a and the second point 25b are set as t1 and t2 respectively, the point passing time ΔT becomes (t2 - t1). When the distance between the first point 25a and the second point 25b is set as L1, and the distance between the first point 25a and the sorting unit 90 is set as L2, in the case where the particle P performs uniform motion, the predicted arrival time is obtained as (L2 / L1)×(t2 - t1). In addition, there may be a time lag from when the sorting signal is generated until the particle P is sorted. When this time lag is set as B, in order to appropriately sort the particle P, it is necessary to emit the sorting signal B earlier, and the optimal sorting signal generation timing based on the predicted arrival time is obtained as (L2 / L1)×(t2 - t1) - B. Since L1 can be calculated based on the design value, the values of L2 and B are obtained through calibration here, so that the optimal delay time for generating the sorting signal can be obtained when t1 and t2 are given. As a specific method of calibration, the delay time calculation unit 80 commands the sorting signal control unit 86 to send sorting signals to each particle flowing in the flow path 10 while varying the coefficients L2 and the time lag B within a certain range. The sorting signal control unit 86 sends a signal to the sorting unit 90, and the sorting unit 90 attempts to sort. The second particle detector 72 calculates the sorting success rate for each of the coefficients L2 and the time lag B, and outputs the determination result to the delay time calculation unit 80. The delay time calculation unit 80 obtains the relationship between the coefficients L2 and the time lag B and the sorting success rate, and obtains the coefficients L2 and the time lag B with the maximum sorting success rate from there, thereby performing calibration.

Claims

1. An imaging flow cytometer, characterized in that, It has: a laser unit that irradiates a first laser and a second laser to a first point and a second point on the flow path of the particle flow respectively; a first imaging unit that captures an image of the particles flowing at the first point; a second imaging unit that captures an image of the particles flowing at the second point; a first detector that detects the particles passing through the first point; a second detector that detects the particles passing through the second point; a first particle detection unit that detects the particles based on the signals obtained by the first detector and the second detector, and issues an imaging timing indication signal for indicating the imaging timing to the first imaging unit and the second imaging unit; a system time management unit that issues system time to the first particle detection unit; an image storage unit that receives the images captured by the first imaging unit and the second imaging unit; and a sorting judgment unit that judges whether the particles are target particles by analyzing the images of the particles output from the image storage unit, the first imaging unit and the second imaging unit cut out the images of the captured particles based on the imaging timing indication signal.

2. The imaging flow cytometer according to claim 1, characterized in that the laser unit irradiates the first laser and the second laser with different wavelengths.

3. The imaging flow cytometer according to claim 1 or 2, characterized in that the first imaging unit and the second imaging unit respectively have line sensors.

4. The imaging flow cytometer according to claim 1 or 2, characterized in that the first imaging unit and the second imaging unit are arranged along the axial direction of the flow path.

5. The imaging flow cytometer according to claim 1 or 2, characterized in that the first particle detection unit sends an identification number assigned in units of the particles to the first imaging unit and the second imaging unit.

6. The imaging flow cytometer according to claim 5, characterized in that the first imaging unit and the second imaging unit respectively send the cut-out images of the particles and the identification numbers of the particles to the image storage unit.

7. The imaging flow cytometer according to claim 1 or 2, characterized in that, It further has: a delay time calculation unit that calculates the delay time of the particles; a sorting signal control unit that issues a sorting signal for sorting the particles in accordance with the timing calculated by the delay time calculation unit according to the judgment result of the sorting judgment unit; and a sorting unit that sorts the target particles based on the sorting signal, the delay time calculation unit calculates the delay time until the particles reach the sorting unit based on the system time of the timing when the first detector and the second detector obtain signals.

8. The imaging flow cytometer according to claim 7, characterized in that It also has a second particle detection unit that detects the particles passing through the detection position, and the detection position is located downstream of the first point and the second point on the flow path. The delay time calculation unit compares the predicted arrival time of the particle at the detection position calculated based on the system times when the particle passes through the first point and the second point with the arrival time of the particle detected by the second particle detection unit at the detection position. The sorting signal control unit adjusts the timing of sorting the target particles by the sorting unit based on the comparison result between the predicted arrival time and the arrival time.

9. The imaging flow cytometer according to claim 8, wherein the detection position is located downstream of the sorting unit, the second particle detection unit calculates the sorting success rate of the particles, the delay time calculation unit adjusts the timing of sorting the target particles by the sorting unit based on the predicted arrival time and the sorting success rate.

Images