Droplet real-time quality control method, equipment, storage medium and program product based on single-cell sequencing

Abstract

The present invention discloses a real-time droplet quality control method, device, storage medium and program product based on single-cell sequencing, wherein the method comprises: obtaining a droplet microfluidic channel image taken in real time during a droplet generation process, wherein the droplet microfluidic channel image comprises single microfluidic channel image information corresponding to each single microfluidic channel and / or main microfluidic channel image information corresponding to a main microfluidic channel formed after the single microfluidic channels merge, wherein solutions corresponding to each single microfluidic channel interact with each other to form oil-in-water droplets, which flow to the main microfluidic channel; and monitoring the droplet generation state based on the droplet microfluidic channel image.

Description

Technical Field

[0001] The present invention relates to the field of gene technology, and in particular to a droplet real-time quality control method, device, storage medium and program product based on single-cell sequencing. Background Art

[0002] At present, single-cell sequencing technology is mainly divided into two categories: droplet-based single-cell sequencing and plate-based single-cell sequencing. Both technologies are used to analyze the gene expression of single cells, but there are significant differences in operating principles and application scenarios.

[0003] Droplet-based single-cell sequencing is a high-throughput technology that is currently widely used in the industry. This technology processes single cells and reaction reagents by encapsulating them in tiny oil droplets. These droplets act independently as reaction chambers and can process thousands to tens of thousands of cells in parallel. Each droplet usually contains a unique barcoded DNA molecule that is used to identify RNA molecules derived from the same cell in subsequent genome sequencing. The main advantages of this technology are its high throughput and low sample input requirements.

[0004] Despite this, droplet-based single-cell sequencing has obvious limitations in actual operation. In the existing technology, it is impossible to directly observe the droplet generation and processing process through instruments, resulting in the inability to control the droplet generation process in real time and to control the quality of the generated droplets. In the existing droplet-based single-cell sequencing technology, once problems such as pore blockage, abnormal pressure, chip or instrument failure occur, researchers usually only know about it after the experiment is over, which may not only lead to waste of cell samples, but also affect the reliability of the test results. Summary of the invention

[0005] In order to solve the existing technical problems, the embodiments of the present invention provide a real-time droplet quality control method based on single-cell sequencing and a real-time droplet quality control device based on single-cell sequencing, which can monitor the abnormal state of the droplet generation process in real time, so as to facilitate maintenance personnel to promptly discover and eliminate abnormalities.

[0006] In a first aspect, a real-time droplet quality control method based on single-cell sequencing is provided, comprising: obtaining a droplet microfluidic channel image taken in real time during the droplet generation process, the droplet microfluidic channel image including single microfluidic channel image information corresponding to each single microfluidic channel and / or main microfluidic channel image information corresponding to the main microfluidic channel formed after the single microfluidic channels merge, wherein solutions corresponding to each single microfluidic channel interact with each other to form droplets and flow to the main microfluidic channel; based on the droplet microfluidic channel image, monitoring the droplet generation state.

[0007] In a second aspect, a droplet real-time quality control device based on single-cell sequencing is provided, comprising a memory and a controller, wherein the memory stores a computer program, and when the computer program is executed by the controller, the controller executes the steps of the droplet real-time quality control method based on single-cell sequencing provided in an embodiment of the present application.

[0008] In a third aspect, a single-cell sequencing library construction device is provided, including a droplet real-time quality control device based on single-cell sequencing provided in an embodiment of the present application.

[0009] In a fourth aspect, a computer-readable storage medium is provided, on which a computer program is stored, and the computer program is executed by a controller to perform the steps of a droplet real-time quality control method based on single-cell sequencing.

[0010] In a fifth aspect, a computer program product comprises a computer program, which, when executed by a controller, implements the steps of the real-time droplet quality control method based on single-cell sequencing as claimed in claim 1.

[0011] The embodiment of the present application can collect real-time droplet microfluidic channel images during the droplet generation process, and the droplet microfluidic channel images at least include the following single microfluidic channel image information and main microfluidic channel image information. By performing image analysis on the droplet microfluidic channel images, the flow state of each single microfluidic channel and the generated droplet conditions can be monitored, so that the abnormal state of the droplet generation process can be monitored in real time, which is convenient for maintenance personnel to discover abnormalities in time and eliminate abnormalities. BRIEF DESCRIPTION OF THE DRAWINGS

[0012] Figure 1 A schematic block diagram of a droplet real-time quality control device based on single-cell sequencing in one embodiment;

[0013] Figure 2 A schematic diagram of a microfluidic channel for generating droplets in a droplet real-time quality control device based on single-cell sequencing in one embodiment;

[0014] Figure 3 This is a flow chart of a droplet real-time quality control method based on single-cell sequencing in one embodiment;

[0015] Figure 4 A schematic diagram of uneven droplet sizes in one embodiment;

[0016] Figure 5 It is a flow chart of a droplet real-time quality control method based on single-cell sequencing in another embodiment;

[0017] Figure 6 is a schematic diagram of pressure anomaly in one embodiment;

[0018] Figure 7 is a schematic diagram of a foreign object existing in a single microchannel in one embodiment;

[0019] Figure 8 A schematic diagram of a process for identifying a removable marker in an embodiment;

[0020] Fig. 9 A schematic diagram of a droplet real-time quality control device based on single-cell sequencing in one embodiment;

[0021] Fig.10 Schematic diagram of the structure of a droplet real-time quality control device based on single-cell sequencing in one embodiment. DETAILED DESCRIPTION

[0022] The technical solution of the present invention is further described in detail below in conjunction with the accompanying drawings and specific embodiments of the specification.

[0023] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as those generally understood by those skilled in the art of the technical field of the present invention. The terms used herein in the specification of the present invention are only for the purpose of describing specific embodiments and are not intended to limit the scope of protection of the present invention. The term "and / or" used herein includes any and all combinations of one or more related listed items.

[0024] In the following description, reference is made to “some embodiments”, which describes a subset of all possible embodiments, but it should be understood that “some embodiments” may be the same subset or different subsets of all possible embodiments and may be combined with each other without conflict.

[0025] Microfluidics refers to a technology that manipulates fluids in micrometer-scale space. This technology can miniaturize the basic functions of chemistry, biology and other laboratories onto a chip of a few square centimeters, so it is also called a chip laboratory (Lab-on-a-chip). Microfluidic chip systems can manipulate the flow of fluids in tiny channels or components ranging from tens to hundreds of micrometers, and the volume of fluids manipulated can be as small as 10 -9 ~10 -18 L. It integrates the basic operations of biochemical experiments such as sample reaction, preparation, separation, and detection into a very small chip, and uses controllable fluids to form a network of microchannels throughout the microfluidic system. While realizing the functions of conventional biochemical laboratories, it reduces the cost of analysis and detection, speeds up the reaction speed, improves the reaction efficiency, and makes the experiment more controllable.

[0026] The generation of droplets on a microfluidic chip is a process in which one phase of fluid is dispersed in another phase of immiscible or partially miscible fluid. One of the two immiscible liquids is used as the continuous phase and the other as the dispersed phase. The dispersed phase is dispersed in a tiny volume (10 -15 ~10 -9 The L) unit is dispersed in the continuous phase to form droplets. At present, the methods for forming droplets can be divided into two categories: passive method and active method. The passive method refers to controlling the generation of droplets by controlling the microchannel structure and the two-phase flow rate ratio, while the active method generally drives and controls the generation of droplets by external force.

[0027] After droplet generation, the cells or nuclei in each droplet release their nucleic acids (DNA or RNA) in preparation for amplification, which is performed using specific amplification techniques such as multiple displacement amplification (MDA) or PCR amplification by droplets.

[0028] See also Figure 1, is a schematic block diagram of a droplet real-time quality control device based on single-cell sequencing in one embodiment. The droplet real-time quality control device based on single-cell sequencing includes but is not limited to the following devices or components: a controller 10, a gas generating device 11, a regulating device 12, an image acquisition device 13, a pressure detection device 14, a gas impurity removal device 15, a pressure output device 16 and a display device 17. The controller 10 is connected to each of the other devices respectively. As the core of the droplet real-time quality control device based on single-cell sequencing, the controller is responsible for coordinating and managing the functions of each component. The controller receives data from each device, processes and analyzes the data of each device, and generates different commands to control each device. The gas generating device 11 is used to provide stable and adjustable air pressure for use in the entire experimental process. The air pressure generating device 11 is a key device for maintaining the pressure environment required for the experiment to ensure that the experimental conditions meet the preset standards. The gas impurity removal device 15 is used to remove moisture from the gas to prevent water vapor from affecting the experimental accuracy and equipment performance. Dry gas helps maintain the stability of the system and extend the service life of the equipment. The air pressure generated by the air pressure generating device 11 is output to each single microchannel through the pressure output device 16 to maintain the pressure in each single microchannel. The pressure detection device 14 is used to detect the pressure level of each single microchannel in real time. The pressure detection device 14 can accurately measure and record the slight changes in the pressure in each single microchannel, providing key data for the experiment. The regulating device 12 is used to control the rate and amount of gas flowing into each single microchannel to ensure the uniformity and stability of the experimental conditions. The regulation of the regulating device 12 can be automatically adjusted according to the pressure data fed back by the pressure detection device 14 to respond to the changes in demand during the experiment. The regulating device 12 is connected to the air pressure generating device 11. The image acquisition device 13 is used to collect real-time droplet microchannel images during the droplet generation process, and transmit the captured images and / or videos to the display device 17 in real time. The image acquisition device 13 can capture detailed images of each moment in the droplet generation process, so as to facilitate the analysis of whether there is a fault in the droplet generation process or whether the generated droplets are qualified according to the images in the droplet generation process. The display device 17 may be a display component integrated in the droplet real-time quality control device based on single-cell sequencing, or an external display device connected to the droplet real-time quality control device based on single-cell sequencing, etc. The display device 17 is used to display the display data sent by the controller 10 for intuitive viewing by the user.

[0029] In some embodiments, the image acquisition device 13 may include but is not limited to the following components: a high frame rate camera, a collimating lens, a fill light, etc., wherein the high frame rate camera has high definition and high-speed image capture capabilities to capture rapidly changing experimental conditions. The collimating lens is used to focus and guide the light to accurately irradiate the photosensitive element to ensure image quality. The fill light is used to provide illumination in a weak light environment, so that a clear image can be obtained even in a weak environment. The image acquisition device 13 can be one or more. The adjustment device 12 can be one or more proportional valves.

[0030] Figure 2 This is a schematic diagram of a microfluidic channel used to generate droplets in a real-time droplet quality control device based on single-cell sequencing in one embodiment; the schematic diagram shown in this embodiment includes four single microfluidic channels, a first single microfluidic channel, a second single microfluidic channel, a third single microfluidic channel and a fourth single microfluidic channel. The air pressure output by the pressure output device 16 maintains the pressure of the four single microfluidic channels. Under the action of pressure, different solutions in each single microfluidic channel flow in their respective channels. When two single microfluidic channels have a confluence, the respective solutions in the two single microfluidic channels mix at the intersection, and finally droplets are generated at the intersection structure of the chip flow channel, and flow out to the droplet storage chamber through the main microfluidic channel. Figure 2 As shown, for example, the solution flowing in the first single microfluidic channel is a cell marker solution, the solution flowing in the second single microfluidic channel is a cell solution, the solution flowing in the third single microfluidic channel is a reaction reagent solution, and the solution flowing in the fourth single microfluidic channel is an oil phase solution. The cell marker solution (such as microspheres with nucleic acid sequences) and the cell solution are placed in one place of the chip flow channel (such as Figure 2 The mixed solution A and the reaction reagent solution are mixed at another location of the chip flow channel (such as Figure 2 The oil phase solution is mixed at the intersection structure of the chip flow channel (for example Figure 2 After the mixed solution B is cut to form oil-in-water droplets at the third position shown in the figure, droplets are generated, and the droplets wrap the cells, cell markers (such as microspheres with nucleic acid sequences) and reaction reagents, and flow into the droplet storage chamber to complete the sample processing before the single-cell sequencing library is built. It can be understood that the first position, the second position and the third position are just a position for the convenience of understanding and illustration, and of course, it can also be illustrated in Figure 2 Other reasonable positions in the reaction. The reaction reagents include but are not limited to cell lysate, primers, enzymes, and nucleoside triphosphate (NTP).

[0031] In the prior art, foreign matter sometimes exists in a single microchannel during the droplet generation process. These foreign matter may cause blockage, or although they do not block, they will affect the flow rate. The presence of foreign matter will affect the flow rate of the solution in a single microchannel, which can easily lead to uneven droplet sizes or discontinuous droplet spacing due to too slow a flow rate. Sometimes, there will be pressure instability in a single microchannel, which will affect the flow rate of the solution, causing the generated droplets to be uneven in size and discontinuous in spacing. The prior art is unable to monitor and control these abnormal phenomena in real time, and is unable to control the quality of the generated droplets in real time, resulting in abnormal droplets entering the droplet storage chamber and mixing with normal droplets. If the number of abnormal droplets is large, they can be found under a microscope by size differences, but even if the abnormality is found, it cannot be remedied and the experiment can only be repeated, resulting in waste of specimens, especially precious trace samples; at the same time, reagents and consumables for single-cell sequencing are also wasted, resulting in cost and time waste. If the number of abnormal droplets is small or the operator does not observe carefully, it is difficult to distinguish and count the abnormal droplets under the microscope. These droplets will continue to the next step of amplification and library construction, resulting in erroneous results, which seriously affects scientific research and clinical diagnosis and treatment. In the above two cases, it is only possible to observe and discover them under a microscope after the experiment is completed and the machine is shut down. The existing technology cannot perform real-time quality control on the quality of the droplets before the machine is shut down, and it is even more impossible to intervene in the quality of abnormal droplets during the experiment.

[0032] See also Figure 3 , is a flow chart of a droplet real-time quality control method based on single-cell sequencing provided in an embodiment of the present application. The droplet real-time quality control method based on single-cell sequencing is applied to a gene sequencer, and the droplet real-time quality control method based on single-cell sequencing includes the following steps:

[0033] S11. Acquire a droplet microfluidic channel image captured in real time during the droplet generation process.

[0034] In this embodiment, the droplet microfluidic channel image includes the single microfluidic channel image information corresponding to each single microfluidic channel and / or the main microfluidic channel image information corresponding to the main microfluidic channel formed after the single microfluidic channels merge, wherein the solutions corresponding to each single microfluidic channel interact with each other to form droplets and flow to the main microfluidic channel.

[0035] In some embodiments, the droplet microfluidic channel image may be an image captured by one image acquisition device 13, or may be an image captured by multiple image acquisition devices 13. Due to the limitation of the field of view of the image acquisition device 13, the image acquisition devices 13 may be respectively arranged at different positions, so that the real-time droplet microfluidic channel image during the droplet generation process can be captured. In some implementations, the image acquisition device 13 may also be an image acquisition device 13 with a motion mechanism, and the controller 10 drives the motion mechanism to rotate, so that the image acquisition device 13 can capture real-time images of the droplet generation process at different angles. Therefore, the droplet microfluidic channel image may include one or more images, and each image may include the main microfluidic channel image information and / or one or more single microfluidic channel image information.

[0036] In some optional methods, the image acquisition parameters of the image acquisition device 13 can be configured on the user interface, where the image acquisition parameters include but are not limited to frame rate, resolution, exposure time, focal length, storage image format, storage video mode, etc., and the image acquisition device 13 is controlled by the image acquisition parameters.

[0037] S12. Monitoring the droplet generation state based on the droplet microfluidic channel image.

[0038] In this embodiment, the image acquisition device 13 can acquire real-time droplet microfluidic channel images during the droplet generation process, and can perform image analysis on the acquired droplet microfluidic channel images to monitor the droplet generation state. The droplet generation state includes at least one of the following: the flow state of each single microfluidic channel during the droplet generation process, and the droplet generated in the main microfluidic channel. The flow state includes but is not limited to the presence of foreign matter, the presence of foreign matter and the foreign matter is a removable marker, the presence of foreign matter and the foreign matter is an unremovable marker, a normal flow state, and the like. By extracting the droplet image from the main microfluidic channel image information and analyzing one or more droplets in the droplet image, it is possible to determine whether the generated droplets meet the preset conditions, thereby monitoring whether abnormal problems occur during the droplet generation process.

[0039] In some embodiments, when the flow state of each single microchannel is abnormal based on the droplet microchannel image monitoring, an early warning prompt is executed; when the generated droplet condition is abnormal based on the droplet microchannel image monitoring, an early warning prompt is executed. In an optional implementation, the method of executing the early warning prompt includes but is not limited to one or more of the following combinations: executing an abnormal prompt on the user interface, sending an abnormal prompt to the terminal device of the relevant personnel. When executing the abnormal prompt on the user interface, the image, pressure data, etc. corresponding to the abnormal situation can be displayed.

[0040] In the above embodiment, the real-time droplet microfluidic channel image during the droplet generation process is collected in real time, and the droplet microfluidic channel image at least includes the following single microfluidic channel image information and main microfluidic channel image information. By performing image analysis on the droplet microfluidic channel image, the flow state of each single microfluidic channel and the generated droplet state can be monitored, so that the abnormal state in the droplet generation process can be monitored in real time, which is convenient for maintenance personnel to discover and eliminate abnormalities in time.

[0041] In some embodiments, S12 further includes:

[0042] When abnormal droplet information is identified in the main microchannel image information, an abnormal warning is performed, wherein the abnormal droplet information includes at least one of the following: uneven droplet size, a single droplet size does not meet the standard droplet size, and the intervals between droplets are uneven.

[0043] The standard droplets are droplets collected by the device under normal operating conditions. In an optional manner, multiple droplet microfluidic channel images can be collected under normal operating conditions of the device, and then multiple droplet images can be extracted from the multiple droplet microfluidic channel images using an image recognition algorithm, and the standard droplet contour is extracted based on the multiple droplet images, and then the size of the standard droplet contour is calculated using a size calculation method, so that the size of the standard droplet can be calculated. In an optional manner, multiple standard droplets are identified from multiple droplet microfluidic channel images using an image recognition algorithm; then the distance between each two adjacent droplets among the multiple standard droplets is calculated to obtain multiple spacing distances; the average of the multiple spacing distances is calculated to obtain the standard droplet spacing.

[0044] Optionally, an image recognition algorithm is used to extract the current droplet image monitored from the main microfluidic channel image information, and based on the current droplet image, the size of the current droplet is calculated; the size of the current droplet is compared with the size of the standard droplet to determine whether the current droplet is abnormal; in some embodiments, in order to reduce errors, the sizes of multiple current droplets are calculated, and the size average of multiple current droplets is calculated, and the size average is compared with the size of the standard droplet to determine whether the current droplet is abnormal; when the size of the current droplet is not within the error range of the size of the standard droplet, or the size average is not within the error range of the size of the standard droplet, it means that the droplet size is uneven and the size of a single droplet does not meet the standard droplet size, and it is determined that the currently generated droplet is abnormal, that is, abnormal droplet information is identified in the main microfluidic channel image information; when the size of the current droplet is within the error range of the size of the standard droplet, or the size average is within the error range of the size of the standard droplet, it means that the droplet size is uniform and the size of a single droplet meets the standard droplet size, and it is determined that the currently generated droplet is not abnormal, that is, no abnormal droplet information is identified in the main microfluidic channel image information. Figure 4 FIG. 1 is a schematic diagram showing uneven droplet sizes in one embodiment.

[0045] Optionally, an image recognition algorithm is used to extract the monitored current droplet image from the main microchannel image information, and multiple current droplet contours are identified from the current droplet image; the interval between every two adjacent droplet contours in the multiple current droplet contours is calculated to obtain multiple current droplet intervals; the multiple current droplet intervals are accumulated and averaged to obtain the current droplet interval mean; when the current droplet interval mean and the standard droplet interval are not within a preset interval error range, that is, the intervals between the droplets are uneven, it is determined that the currently generated droplets are abnormal, that is, abnormal droplet information is identified in the main microchannel image information; when the current droplet interval mean and the standard droplet interval are within the preset interval error range, that is, the intervals between the droplets are uniform, it is determined that the currently generated droplets are not abnormal, that is, no abnormal droplet information is identified in the main microchannel image information.

[0046] In the above embodiment, the droplet conditions are analyzed in real time from the main microchannel image information monitored in real time, the size of the currently generated droplets and the intervals between the droplets are monitored, and the conditions of the currently generated droplets are monitored from multiple angles of the droplets, so as to timely discover abnormal conditions occurring during the droplet generation process and perform abnormal warnings.

[0047] In some embodiments, in order to more accurately determine the abnormal type in the droplet generation process, a more accurate early warning prompt is provided, such as Figure 5 As shown, Figure 5 It is a flow chart of a droplet real-time quality control method based on single-cell sequencing in another embodiment;

[0048] S11, acquiring a droplet microfluidic channel image captured in real time during the droplet generation process.

[0049] S13, obtaining pressure data of each single microfluidic channel.

[0050] S14, when it is identified that there is abnormal droplet information in the main microchannel image information, and the pressure data of each single microchannel meets the corresponding preset pressure conditions, a first warning prompt is executed.

[0051] S15, when it is identified that there is abnormal droplet information in the main microchannel image information, and unqualified pressure data that does not meet the corresponding preset pressure conditions exists in the pressure data of each single microchannel, a second early warning prompt is executed.

[0052] In one embodiment, in the process of generating droplets, steps S11 and S13 are executed in no particular order. In the process of generating droplets, the image acquisition device 13 acquires the droplet microfluidic channel image in real time, and the pressure detection device 14 monitors the pressure data of each single microfluidic channel in real time. The controller 10 obtains the real-time droplet microfluidic channel image through step S11, and the controller 10 obtains the pressure data of each single microfluidic channel through step S13. The implementation method of identifying whether there is abnormal droplet information in the main microfluidic channel image information has been described in detail in the above embodiment, and will not be repeated here. The pressure data of each single microfluidic channel indicates the pressure data within the time period of collecting the droplet microfluidic channel image containing abnormal droplet information. Specifically, according to the acquisition time of the droplet microfluidic channel image containing abnormal droplet information, the pressure data of each single microfluidic channel within the target time period including the acquisition time is acquired. The target time period may indicate a preset time period before and after the acquisition time, or a preset time period before the acquisition time.

[0053] In an optional embodiment, after identifying that there is abnormal droplet information in the main microchannel image information, and that there is unqualified pressure data that does not meet the corresponding preset pressure conditions in the pressure data of each single microchannel, before executing the second early warning prompt, the method further includes:

[0054] Control the backup gas generating device to generate gas pressure and output it to each single microchannel, and continue to obtain the pressure data of each single microchannel. If abnormal droplet information continues to be identified in the main microchannel image information, and if unqualified pressure data that does not meet the corresponding preset pressure conditions exists in the pressure data of each single microchannel, execute the second early warning prompt.

[0055] The controller 10 controls the gas generating device 11 to generate gas pressure, collects the pressure data of each single microchannel through the pressure detection device 14, and when it is identified that there is abnormal droplet information in the main microchannel image information, and there is unqualified pressure data that does not meet the corresponding preset pressure conditions in the pressure data of each single microchannel collected by the pressure detection device 14, the controller 10 switches the gas generating device 11 to the backup gas generating device, and continues to execute the above steps S13-S15. If it is still identified that there is abnormal droplet information in the main microchannel image information, and there is unqualified pressure data that does not meet the corresponding preset pressure conditions in the pressure data of each single microchannel, then the second early warning prompt is executed, indicating that a gas leakage phenomenon occurs; if after switching to the backup gas generating device, no abnormal droplet information is identified in the main microchannel image information, or no unqualified pressure data that does not meet the corresponding preset pressure conditions is detected in the pressure data of each single microchannel, it means that the fault is eliminated by activating the backup gas generating device, then the air pump is faulty, and the backup gas generating device is activated.

[0056] When abnormal droplet information is identified in the main microchannel image information, there may be foreign matter in the single microchannel or the pressure of the single microchannel is unstable. In order to more accurately determine the type of abnormality, the real-time data collected by multiple sensors are combined to locate the abnormal problem. When abnormal droplet information is identified in the main microchannel image information, and the pressure data of each single microchannel meets the corresponding preset pressure conditions, it is determined that the pressure of each single microchannel is stable, and there may be foreign matter in the single microchannel, and the first early warning prompt is executed; when abnormal droplet information is identified in the main microchannel image information, and unqualified pressure data that does not meet the corresponding preset pressure conditions exists in the pressure data of each single microchannel, it is determined that the pressure data in at least one single microchannel is unstable, and the second early warning prompt is executed. Figure 6 As shown, Figure 6 FIG. 1 is a schematic diagram of pressure anomaly in an embodiment. The pressure data of four single microfluidic channels are displayed on the user interface, and the number 1 indicates the first single microfluidic channel, the number 2 indicates the second single microfluidic channel, the number 3 indicates the third single microfluidic channel, and the number 4 indicates the fourth single microfluidic channel. The pressure anomaly is displayed in the abnormal single microfluidic channel area, such as Figure 6 In the example, the pressure data of the first single microfluidic channel is abnormal, and the pressure abnormality is displayed in the area corresponding to the first single microfluidic channel.

[0057] Optionally, the abnormality type corresponding to the first warning prompt is different from the abnormality type corresponding to the second warning prompt. The abnormality type corresponding to the first warning prompt includes the presence of an abnormality in a single microchannel, and the abnormality type corresponding to the second warning prompt includes at least one of the following: unstable pressure. There is an abnormality in a single microchannel, such as foreign matter in a single microchannel, which may be bubbles, fiber filaments, and the self-depression of the microchannel glue product in the device, etc. For example, a single microchannel is made of glue products, which may cause the microchannel to deform and sag after long-term use, thereby blocking the microchannel, etc. The unstable pressure may be caused by air leakage or leakage in a single microchannel, or failure of the pressure output device 16, etc. According to the identified abnormality type, different warning prompts are displayed on the user interface to guide maintenance personnel to locate the problem in time, and the abnormal image corresponding to the warning prompt or the video where the abnormal image is located is displayed on the user interface, or the pressure data of each single microchannel is displayed on the user interface as a curve change trend, so that the user can determine the automatically identified abnormality type based on the monitored real-time image data or real-time pressure data, and help maintenance personnel quickly find abnormalities in the droplet generation process.

[0058] In the above embodiment, the real-time collected image information of the main microchannel and the real-time collected pressure data of each single microchannel are combined to automatically evaluate the abnormal type and perform different early warning prompts. Combined with the real-time data collected by multiple sensors, the abnormal problem can be located more accurately, helping maintenance personnel to quickly find the abnormalities in the droplet generation process and solve the problems.

[0059] In some embodiments, S12 further includes:

[0060] When abnormal droplet information is identified in the main microchannel image information and foreign matter is identified in at least one single microchannel, or when foreign matter is identified in at least one single microchannel, a warning is given that foreign matter exists in the single microchannel.

[0061] In this embodiment, since the single microchannel and the main microchannel have different morphologies, the morphological differences include but are not limited to different diameters or the shape of multiple single microchannels combined together is different from the shape of the main microchannel. Therefore, the difference in morphology can be used to obtain the single microchannel image information from the droplet microchannel image, and the internal area of ​​the single microchannel can be segmented using an image segmentation method. The search window area is used to find whether there is an area with a pixel average value greater than a preset pixel value in the internal area of ​​the single microchannel. If there is an area with a pixel average value greater than the preset pixel value in the internal area of ​​the single microchannel, it is determined that there is a foreign body; if there is no area with a pixel average value greater than the preset pixel value in the internal area of ​​the single microchannel, it is determined that there is no foreign body. Generally, the pixel value of foreign matter will be greater than the pixel value inside a normal single microchannel. Therefore, searching the internal area of ​​the single microchannel with a small area window can more accurately determine whether there is a foreign body in the single microchannel. Of course, other image processing methods can also be used to identify foreign matter.

[0062] Optionally, in order to make the recognition more accurate, the main microchannel image information can be combined with the single microchannel image to identify abnormalities. When there is foreign matter in a single microchannel, the flow rate in the single microchannel will be affected, and therefore, the droplets generated in the main microchannel will also be affected. Therefore, combining the main microchannel image information with the single microchannel image to identify abnormalities can improve the accuracy of identifying the abnormal type in the droplet generation process. Figure 7 FIG. 1 is a schematic diagram of a foreign object existing in a single microchannel in one embodiment; Figure 7 Schematic diagram of bubbles in a single microchannel, and schematic diagram of foreign matter blocking a single microchannel.

[0063] Optionally, a warning method for indicating the presence of foreign matter in a single microfluidic channel includes, but is not limited to: displaying an abnormal image indicating the presence of foreign matter or a dynamic video indicating the presence of foreign matter on a user interface.

[0064] In the above embodiment, the real-time situation in the microfluidic channel during the droplet generation process can be monitored in real time, and whether there are foreign objects in a single microfluidic channel can be identified. When the presence of foreign objects in a single microfluidic channel is identified, an early warning prompt can be given, and abnormal problems in the droplet generation process can be located more accurately, helping maintenance personnel to quickly find abnormalities in the droplet generation process and solve the problems.

[0065] In some embodiments, S12 further includes:

[0066] When it is recognized that a removable marker that can be removed by a pressurizing operation exists in the single micro-channel image, a pressurizing operation is performed on the single micro-channel within a target time period to remove the removable marker.

[0067] The target time period indicates the time period during which the pressurization operation can be performed, and the target time period can be a time period configured by the user based on the user interface. The removable markers include but are not limited to bubbles, fiber filaments, cell debris, cell clusters, and the like.

[0068] Optionally, before performing the pressurization operation, the user may be prompted whether to prepare for the preparatory work before the pressurization operation, such as changing the output position of the main microchannel to avoid the discharge of foreign matter from contaminating normal droplets, etc. After obtaining the confirmation data of the user's accurate pressurization operation, the pressurization operation is performed. Optionally, after performing the pressurization operation for a preset number of times, it is recognized that a removable marker exists in the single microchannel image, the pressurization operation is stopped, and the user is prompted that the automatic removal of foreign matter has failed. In an optional implementation, after prompting the user that the automatic removal of foreign matter has failed, the operation of the droplet real-time quality control device based on single-cell sequencing is automatically stopped, and an alarm prompt is issued.

[0069] In the above embodiment, it is possible to identify whether there are removable markers in the real-time captured droplet microfluidic channel image. When there are removable markers, a pressurization operation is performed to automatically clean the removable markers, thereby automatically locating the specific abnormality type, automatically processing the abnormality type, and automatically controlling the abnormalities in the droplet generation process.

[0070] In some embodiments, the method further comprises:

[0071] During the pressurization process, a pressure change curve in the single microchannel where the removable marker is located is displayed on the user interface, and a real-time dynamic image of the single microchannel where the removable marker is located is obtained and displayed;

[0072] Based on the pressurization adjustment operation of the pressurization configuration item obtained from the user interface, a pressurization adjustment instruction is generated, and the pressurization adjustment instruction adjusts the pressurization operation. The pressurization configuration item includes at least one of the following: maximum pressure, minimum pressure, number of pressurizations, pressurization duration, and single microchannel identification item.

[0073] The user interface includes a pressurization configuration interface, which is provided with pressurization configuration items, and the pressurization configuration items include at least one of the following: maximum pressure, minimum pressure, number of pressurizations, pressurization duration, and single microchannel identification item. The single microchannel that needs to be pressurized can be selected through the single microchannel identification item, and the parameter items in the pressurization operation can be adjusted in real time according to the foreign matter situation in the single microchannel by configuring the maximum pressure, minimum pressure, number of pressurizations, and pressurization duration to quickly remove foreign matter. At the same time, the pressure change curve and dynamic image can be displayed on the pressurization configuration interface, which is convenient for maintenance personnel to observe the pressure changes during the pressurization operation and whether the pressurization operation can remove foreign matter.

[0074] In the above embodiment, the pressurization operation can be adjusted in real time through the pressurization configuration items on the user interface, and the pressure change curve and dynamic images are displayed on the interface, which is convenient for maintenance personnel to observe the pressure changes during the pressurization operation and whether the pressurization operation can eliminate foreign matter, thereby improving the efficiency of foreign matter removal.

[0075] In some embodiments, the method further includes: identifying whether there is a removable marker in each single microfluidic channel image.

[0076] like Figure 8 As shown, Figure 8 A schematic diagram of a process for identifying a removable marker in an embodiment, the process diagram includes:

[0077] S81, acquiring morphological data of the main microchannel and the single microchannel, and extracting a single microchannel image from the droplet microchannel image based on the morphological data of the main microchannel and the single microchannel.

[0078] In this embodiment, since the single microchannel and the main microchannel have different morphologies, including but not limited to the diameter or the shape of multiple single microchannels combined together, the difference in morphology can be used to obtain single microchannel image information from the droplet microchannel image, and the internal area of ​​the single microchannel can be segmented using image segmentation methods.

[0079] S82. Using an image recognition algorithm, identify whether there is foreign matter in the single microfluidic channel image.

[0080] Optionally, the identifying whether there is a foreign object in the single microfluidic channel image by using an image recognition algorithm includes at least one of the following:

[0081] The inner area of ​​the single microchannel is traversed in sequence with the search window area, and the average pixel value in the search window area during each traversal process is calculated respectively; whether there is an area with an average pixel value greater than a preset pixel value is found, and if there is an area with an average pixel value greater than the preset pixel value in the inner area of ​​the single microchannel, it is determined that there is a foreign object; if there is no area with an average pixel value greater than the preset pixel value in the inner area of ​​the single microchannel, it is determined that there is no foreign object;

[0082] A standard single microfluidic channel image without foreign matter is obtained, and the internal area of ​​the standard single microfluidic channel is segmented; the pixel value of each pixel point in the internal area of ​​the standard single microfluidic channel is obtained, and the standard pixel range in the internal area of ​​the standard single microfluidic channel is determined; the target pixel point whose pixel value in the internal area of ​​the single microfluidic channel is greater than the standard pixel range is searched, and if the target pixel point is a non-isolated point and the number of target pixel points is greater than a preset number, it is determined that a foreign matter exists.

[0083] The search window area is generally set to be small, such as a 4*4 window, and the internal area of ​​a single microchannel is traversed in sequence through the search window area, so that the internal area of ​​the single microchannel can be searched in sequence to see if there are foreign objects. In another implementation, because foreign objects are generally small areas, not isolated points, when the target pixel point is an isolated point or the number of target pixel points is less than a preset number, it means that these target pixel points are noise points, not foreign objects.

[0084] S83. When a foreign object is identified, input data of a pre-trained single microfluidic channel foreign object recognition model is formed based on the single microfluidic channel image, the foreign object features in the single microfluidic channel image are extracted using a feature extraction network in the pre-trained single microfluidic channel foreign object recognition model, and the foreign object features are identified using a classification network in the pre-trained single microfluidic channel foreign object recognition model, and an identification result is output, wherein the identification result can indicate whether the foreign object is a removable marker.

[0085] In this embodiment, the pre-trained single microfluidic channel foreign body recognition model is a model obtained by training based on a training data set, wherein each training sample in the training data set includes a single microfluidic channel image containing foreign matter and label data, wherein the label of the training sample containing a removable marker is a removable marker label, and the label of the training sample containing a non-removable marker is a non-removable marker label.

[0086] Optionally, the method further comprises:

[0087] Get the training dataset;

[0088] Construct an initial single microfluidic channel foreign body recognition model;

[0089] The initial single microfluidic channel foreign body recognition model is iteratively trained through the training data set, and training samples are obtained from the training data set to form input sample data. The single microfluidic channel foreign body recognition model in the iterative process outputs output data corresponding to the input sample data. Based on the loss function, the label data corresponding to the input sample data and the output data corresponding to the input sample data, the loss value in each iteration process is calculated until the iteration termination condition is reached, thereby obtaining the pre-trained single microfluidic channel foreign body recognition model.

[0090] When preparing the training data set, a large number of images of foreign matter in a real single microfluidic channel are obtained by actual shooting or other methods as training samples, and the types of foreign matter are marked therein, that is, clearable marker labels or non-clearable marker labels. A large number of such training samples constitute the training data set.

[0091] The single microchannel foreign body recognition model includes a feature extraction network and a classification network, and the feature extraction network is used to extract foreign body features from a single microchannel image. Specifically, the feature extraction network may include a convolution layer, an activation layer, a normalization process, a pooling layer, a fully connected layer, and the like, wherein the convolution layer is used to perform a convolution operation and extract a feature map of a single microchannel image; the feature map of a single microchannel image is input into the activation layer, and nonlinear mapping is performed using an activation function. The normalization operation avoids a sudden increase or decrease in the value during the processing process, and the pooling layer is used to downsample the output of the convolution layer, reduce the data dimension, and reduce the model complexity and the amount of calculation. The fully connected layer can expand the output obtained after processing by the convolution layer, the activation layer, and the pooling layer, and connect the expanded content to the output layer in a fully connected manner to obtain abnormal feature data. The classification network layer can be a classification network in the form of a Unet-like structure. The classification network layer can identify whether the foreign body in a single microchannel is a removable foreign body or an unremovable foreign body.

[0092] The loss function includes but is not limited to the mean square error function, the cross entropy function, and the like. The single microfluidic channel foreign body recognition model is supervisedly trained by taking the label data corresponding to the sample data as the training target. During the training process, the single microfluidic channel foreign body recognition model continuously learns the characteristics of the clearable markers and the characteristics of the non-clearable markers, and optimizes the parameters in the single microfluidic channel foreign body recognition model. In each iteration, the loss value between the output data corresponding to the input sample data and the label data corresponding to the input sample data is calculated based on the loss function, and whether the current iteration meets the iteration termination condition is determined based on the loss value. When the current iteration does not meet the iteration termination condition, continue to obtain the input sample data from the training data set to continue training, and update the parameters in the single microfluidic channel foreign body recognition model until the iteration termination condition is met. The single microfluidic channel foreign body recognition model after stopping the iteration is used as the trained single microfluidic channel foreign body recognition model.

[0093] In the above embodiment, it is possible to automatically identify whether foreign matter exists in each single microchannel in the real-time captured droplet microchannel image. When foreign matter exists, it is possible to automatically identify whether the foreign matter is a removable marker through a pre-trained single microchannel foreign matter recognition model. If the foreign matter is identified as a removable marker, the foreign matter can be automatically cleaned. Therefore, it is possible to monitor and control abnormalities in the droplet generation process in real time, and handle the abnormalities in a timely manner to ensure the generation of normal droplets.

[0094] On the other hand, the present application provides a computer program product, including a computer program, which, when executed by a processor, implements the droplet real-time quality control method based on single-cell sequencing described in any embodiment of the present application.

[0095] Among them, in the computer program product, the optional implementation form of the program module architecture of the computer program that implements each step of the method can be a droplet real-time quality control device based on single-cell sequencing. Fig. 9 An embodiment of the present application provides a real-time droplet quality control device based on single-cell sequencing, including: an acquisition module 90, used to acquire a droplet microfluidic channel image taken in real time during the droplet generation process, the droplet microfluidic channel image including single microfluidic channel image information corresponding to each single microfluidic channel and / or main microfluidic channel image information corresponding to the main microfluidic channel formed after the single microfluidic channels merge, wherein the solutions corresponding to each single microfluidic channel interact with each other to form droplets and flow to the main microfluidic channel; a monitoring module 91, used to monitor the droplet generation state based on the droplet microfluidic channel image.

[0096] Optionally, the monitoring module 91 is further used for:

[0097] When abnormal droplet information is identified in the main microchannel image information, an abnormal warning is performed, wherein the abnormal droplet information includes at least one of the following: uneven droplet size, a single droplet size does not meet the standard droplet size, and the intervals between droplets are uneven.

[0098] Optionally, the monitoring module 91 is further used for:

[0099] Obtain pressure data of each single microfluidic channel;

[0100] When it is identified that there is abnormal droplet information in the main microchannel image information and the pressure data of each single microchannel meets the corresponding preset pressure conditions, a first early warning prompt is executed.

[0101] When it is identified that there is abnormal droplet information in the main microchannel image information, and unqualified pressure data that does not meet the corresponding preset pressure conditions exists in the pressure data of each single microchannel, a second early warning prompt is executed;

[0102] Optionally, the monitoring module 91 is also used for: after identifying that there is abnormal droplet information in the main microchannel image information, and that the pressure data of each single microchannel contains unqualified pressure data that does not meet the corresponding preset pressure conditions, before executing the second early warning prompt, the method also includes: controlling the backup gas generating device to generate air pressure, and outputting it to each single microchannel, and continuing to obtain the pressure data of each single microchannel. If it continues to be identified that there is abnormal droplet information in the main microchannel image information, and that there is unqualified pressure data that does not meet the corresponding preset pressure conditions in the pressure data of each single microchannel, the second early warning prompt is executed.

[0103] Optionally, the abnormality type corresponding to the first warning prompt is different from the abnormality type corresponding to the second warning prompt, the abnormality type corresponding to the first warning prompt indicates that there is an abnormality in the single microchannel, and the abnormality type corresponding to the second warning prompt indicates unstable pressure.

[0104] Optionally, the monitoring module 91 is further used for:

[0105] When abnormal droplet information is identified in the main microchannel image information and foreign matter is identified in at least one single microchannel, or when foreign matter is identified in at least one single microchannel, a warning is given that foreign matter exists in the single microchannel.

[0106] Optionally, the monitoring module 91 is further used for:

[0107] When it is recognized that a removable marker that can be removed by a pressurizing operation exists in the single micro-channel image, a pressurizing operation is performed on the single micro-channel within a target time period to remove the removable marker.

[0108] Optionally, the monitoring module 91 is further used for:

[0109] During the pressurization process, a pressure change curve in the single microchannel where the removable marker is located is displayed on the user interface, and a real-time dynamic image of the single microchannel where the removable marker is located is obtained and displayed;

[0110] Based on the pressurization adjustment operation of the pressurization configuration item obtained from the user interface, a pressurization adjustment instruction is generated, and the pressurization adjustment instruction adjusts the pressurization operation. The pressurization configuration item includes at least one of the following: maximum pressure, minimum pressure, number of pressurizations, pressurization duration, and single microchannel identification item.

[0111] Optionally, the monitoring module 91 is further used for:

[0112] identifying whether there is a removable marker in each single microfluidic channel image;

[0113] The step of identifying whether there is a removable marker in each single microfluidic channel image comprises:

[0114] Acquire the morphological data of the main microchannel and the single microchannel, and extract the single microchannel image from the droplet microchannel image based on the morphological data of the main microchannel and the single microchannel;

[0115] Use image recognition algorithms to identify whether there are foreign objects in a single microfluidic channel image;

[0116] When a foreign object is identified, input data of a pre-trained single microfluidic channel foreign object recognition model is formed based on the single microfluidic channel image, the feature extraction network in the pre-trained single microfluidic channel foreign object recognition model is used to extract the foreign object features in the single microfluidic channel image, and the classification network in the pre-trained single microfluidic channel foreign object recognition model is used to identify the foreign object features, and outputs whether the foreign object is a removable marker, wherein the pre-trained single microfluidic channel foreign object recognition model is a model trained based on a training data set, and each training sample in the training data set includes a single microfluidic channel image containing a foreign object and label data, wherein the label of the training sample containing a removable marker is a removable marker label, and the label of the training sample containing a non-removable marker is a non-removable marker label.

[0117] Optionally, the monitoring module 91 is further used for:

[0118] Traversing the inner area of ​​the single microfluidic channel in sequence with the search window area, and calculating the pixel average value in the search window area during each traversal process; finding whether there is an area with a pixel average value greater than a preset pixel value, if there is an area with a pixel average value greater than the preset pixel value in the inner area of ​​the single microfluidic channel, it is determined that there is a foreign object; if there is no area with a pixel average value greater than the preset pixel value in the inner area of ​​the single microfluidic channel, it is determined that there is no foreign object; or

[0119] A standard single microfluidic channel image without foreign matter is obtained, and the internal area of ​​the standard single microfluidic channel is segmented; the pixel value of each pixel point in the internal area of ​​the standard single microfluidic channel is obtained, and the standard pixel range in the internal area of ​​the standard single microfluidic channel is determined; the target pixel point whose pixel value in the internal area of ​​the single microfluidic channel is greater than the standard pixel range is searched, and if the target pixel point is a non-isolated point and the number of target pixel points is greater than a preset number, it is determined that a foreign matter exists.

[0120] It can be understood by those skilled in the art that Fig. 9 The structure of the droplet real-time quality control device based on single-cell sequencing does not constitute a limitation on the droplet real-time quality control device based on single-cell sequencing, and each module can be implemented in whole or in part by software, hardware, and a combination thereof. The above modules can be embedded in or independent of the controller in the computer device in the form of hardware, or can be stored in the memory in the computer device in the form of software, so that the controller can call and execute the operations corresponding to the above modules. In other embodiments, the droplet real-time quality control device based on single-cell sequencing may include more or fewer modules than shown in the figure.

[0121] See also Fig.10 In another aspect of the embodiments of the present application, a real-time droplet quality control device 200 based on single-cell sequencing is provided, comprising a controller 10 and a memory 18, wherein the memory 18 stores a computer program. When the computer program is executed by the controller, the controller 10 executes the steps of the real-time droplet quality control method based on single-cell sequencing provided in any of the above embodiments of the present application.

[0122] The controller 10 is a control center, which uses various interfaces and lines to connect various parts of the entire computer device, and executes various functions of the computer device and processes data by running or executing software programs and / or modules stored in the memory 18, and calling data stored in the memory 18. Optionally, the controller 10 may include one or more processing cores; preferably, the controller 10 may integrate an application controller and a modem controller, wherein the application controller mainly processes the operating system, user pages and application programs, etc., and the modem controller mainly processes wireless communications. It is understandable that the above-mentioned modem controller may not be integrated into the controller 10.

[0123] The memory 18 can be used to store software programs and modules. The controller 10 executes various functional applications and data processing by running the software programs and modules stored in the memory 18. The memory 18 can mainly include a program storage area and a data storage area, wherein the program storage area can store an operating system, an application required for at least one function (such as a sound playback function, an image playback function, etc.), etc.; the data storage area can store data created according to the use of the computer device, etc. In addition, the memory 18 can include a high-speed random access memory, and can also include a non-volatile memory, such as at least one disk storage device, a flash memory device, or other volatile solid-state storage devices. Accordingly, the memory 18 can also include a memory controller to provide the controller 10 with access to the memory 18.

[0124] On the other hand, an embodiment of the present application further provides a storage medium storing a computer program. When the computer program is executed by a controller, the controller executes the steps of the droplet real-time quality control method based on single-cell sequencing provided in any of the above embodiments of the present application.

[0125] On the other hand, an embodiment of the present application further provides a single-cell sequencing device, including a droplet real-time quality control device based on single-cell sequencing provided by any of the above embodiments of the present application.

[0126] On the other hand, an embodiment of the present application further provides a single-cell sequencing library construction device, including a droplet real-time quality control device based on single-cell sequencing provided by any of the above embodiments of the present application.

[0127] Those skilled in the art can understand that all or part of the processes in the methods provided in the above embodiments can be completed by instructing the relevant hardware through a computer program, and the program can be stored in a non-volatile computer-readable storage medium. When the program is executed, it can include the processes of the embodiments of the above methods. Among them, any reference to memory, storage, database or other media used in the embodiments provided in the present application can include non-volatile and / or volatile memory. Non-volatile memory may include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM) or flash memory. Volatile memory may include random access memory (RAM) or external cache memory. As an illustration and not limitation, RAM is available in many forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous link (Synchlink) DRAM (SLDRAM), memory bus (Rambus) direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM).

[0128] The above is only a specific embodiment of the present invention, but the protection scope of the present invention is not limited thereto. Any person skilled in the art can easily think of changes or substitutions within the technical scope disclosed by the present invention, which should be included in the protection scope of the present invention. The protection scope of the present invention should be based on the protection scope of the claims.

Claims

1. A droplet real-time quality control method based on single-cell sequencing, characterized in that: include: Acquire a droplet microfluidic channel image captured in real time during the droplet generation process, wherein the droplet microfluidic channel image includes single microfluidic channel image information corresponding to each single microfluidic channel captured, or the single microfluidic channel image information and the main microfluidic channel image information corresponding to the main microfluidic channel formed after the single microfluidic channels merge, wherein the solutions corresponding to each single microfluidic channel interact with each other to form droplets and flow to the main microfluidic channel; Based on the droplet microfluidic channel image, monitoring the droplet generation state; The monitoring of the droplet generation state based on the droplet microfluidic channel image further comprises: identifying whether there is a removable marker in each single microfluidic channel image; When it is identified that there is a removable marker that can be removed by a pressurizing operation in the single microfluidic channel image, a pressurizing operation is performed on the single microfluidic channel within a target time period to remove the removable marker; The step of identifying whether there is a removable marker in each single microfluidic channel image comprises: Acquire the morphological data of the main microchannel and the single microchannel, and extract the single microchannel image from the droplet microchannel image based on the morphological data of the main microchannel and the single microchannel; Use image recognition algorithms to identify whether there are foreign objects in a single microfluidic channel image; When a foreign object is identified, input data of a pre-trained single microfluidic channel foreign object recognition model is formed based on the single microfluidic channel image, the feature extraction network in the pre-trained single microfluidic channel foreign object recognition model is used to extract the foreign object features in the single microfluidic channel image, and the classification network in the pre-trained single microfluidic channel foreign object recognition model is used to identify the foreign object features, and outputs whether the foreign object is a removable marker, wherein the pre-trained single microfluidic channel foreign object recognition model is a model trained based on a training data set, and each training sample in the training data set includes a single microfluidic channel image containing a foreign object and label data, wherein the label of the training sample containing a removable marker is a removable marker label, and the label of the training sample containing a non-removable marker is a non-removable marker label.

2. The droplet real-time quality control method based on single-cell sequencing according to claim 1, characterized in that: The monitoring of the droplet generation state based on the droplet microfluidic channel image comprises: When abnormal droplet information is identified in the main microchannel image information, an abnormal warning is performed, wherein the abnormal droplet information includes at least one of the following: uneven droplet size, a single droplet size does not meet the standard droplet size, and the intervals between droplets are uneven.

3. The droplet real-time quality control method based on single-cell sequencing according to claim 1, characterized in that: The monitoring of the droplet generation state based on the droplet microfluidic channel image further comprises: Obtain pressure data of each single microfluidic channel; When it is identified that there is abnormal droplet information in the main microchannel image information and the pressure data of each single microchannel meets the corresponding preset pressure conditions, a first warning prompt is executed; When it is identified that there is abnormal droplet information in the main microchannel image information, and unqualified pressure data that does not meet the corresponding preset pressure conditions exists in the pressure data of each single microchannel, a second early warning prompt is executed.

4. The method for real-time droplet quality control based on single-cell sequencing according to claim 3, characterized in that: After identifying that there is abnormal droplet information in the main microchannel image information, and that the pressure data of each single microchannel contains unqualified pressure data that does not meet the corresponding preset pressure conditions, and before executing the second early warning prompt, the method also includes: controlling the backup gas generating device to generate gas pressure, and outputting it to each single microchannel, and continuing to obtain the pressure data of each single microchannel. If it continues to be identified that there is abnormal droplet information in the main microchannel image information, and that there is unqualified pressure data that does not meet the corresponding preset pressure conditions in the pressure data of each single microchannel, the second early warning prompt is executed.

5. The droplet real-time quality control method based on single-cell sequencing according to claim 3 or 4, characterized in that: The abnormality type corresponding to the first warning prompt is different from the abnormality type corresponding to the second warning prompt. The abnormality type corresponding to the first warning prompt indicates that there is an abnormality in the single microchannel, and the abnormality type corresponding to the second warning prompt indicates unstable pressure.

6. The method for real-time droplet quality control based on single-cell sequencing according to claim 1, characterized in that: The monitoring of the droplet generation state based on the droplet microfluidic channel image further comprises: When abnormal droplet information is identified in the main microchannel image information and foreign matter is identified in at least one single microchannel, or when foreign matter is identified in at least one single microchannel, a warning is given that foreign matter exists in the single microchannel.

7. The method for real-time droplet quality control based on single-cell sequencing according to claim 1, characterized in that: The method further comprises: During the pressurization process, a pressure change curve in the single microchannel where the removable marker is located is displayed on the user interface, and a real-time dynamic image of the single microchannel where the removable marker is located is obtained and displayed; Based on the pressurization adjustment operation of the pressurization configuration item obtained from the user interface, a pressurization adjustment instruction is generated, and the pressurization adjustment instruction adjusts the pressurization operation. The pressurization configuration item includes at least one of the following: maximum pressure, minimum pressure, number of pressurizations, pressurization duration, and single microchannel identification item.

8. The method for real-time droplet quality control based on single-cell sequencing according to claim 7, characterized in that: The method of using an image recognition algorithm to identify whether there is a foreign object in a single microfluidic channel image includes at least one of the following: Traversing the inner area of ​​the single microfluidic channel in sequence with the search window area, and calculating the pixel average value in the search window area during each traversal process; finding whether there is an area with a pixel average value greater than a preset pixel value, if there is an area with a pixel average value greater than the preset pixel value in the inner area of ​​the single microfluidic channel, it is determined that there is a foreign object; if there is no area with a pixel average value greater than the preset pixel value in the inner area of ​​the single microfluidic channel, it is determined that there is no foreign object; or A standard single microfluidic channel image without foreign matter is obtained, and the internal area of ​​the standard single microfluidic channel is segmented; the pixel value of each pixel point in the internal area of ​​the standard single microfluidic channel is obtained, and the standard pixel range in the internal area of ​​the standard single microfluidic channel is determined; the target pixel point whose pixel value in the internal area of ​​the single microfluidic channel is greater than the standard pixel range is searched, and if the target pixel point is a non-isolated point and the number of target pixel points is greater than a preset number, it is determined that a foreign matter exists.

9. A droplet real-time quality control device based on single-cell sequencing, characterized in that: The method comprises a memory and a controller, wherein the memory stores a computer program, and when the computer program is executed by the controller, the controller executes the steps of the method according to any one of claims 1 to 8.

10. A single-cell sequencing library construction device, characterized in that: Including a droplet real-time quality control device based on single-cell sequencing as described in claim 9.

11. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores a computer program, and when the computer program is executed by the controller, the real-time droplet quality control method based on single-cell sequencing according to any one of claims 1 to 8 is implemented.

12. A computer program product, comprising a computer program, characterized in that When the computer program is executed by the controller, the real-time droplet quality control method based on single-cell sequencing as described in any one of claims 1 to 8 is implemented.

Images