A method and device for generating droplets

Abstract

The present invention discloses a method and device for droplet generation, the method comprising: S1. sample preparation; S2. chip loading; S3. start the test, start the droplet generation device, so that the droplets required for the reaction are generated in the microfluidic chip, and the droplet generation process inside the microfluidic chip is collected in real time by the image acquisition unit in the device and provided to the control unit; the control unit can analyze and calculate the parameters about the size, quantity, proportion and the like of different types of droplets according to the image acquisition information in the droplet generation process; S4. reaction liquid collection and experimental process quality control result generation; the present invention can simultaneously perform quality control on a variety of droplet morphological parameters such as droplet size, empty droplet microbead package, single droplet microbead package, multiple droplet microbead packages and total droplet amount. Droplets encapsulating different types of cells can be quality controlled without affecting the experimental process or the reaction collection after the experiment is completed. The recognition accuracy is high and the stability is good. The quality control statistical data can be real-time statistical data.

Description

Technical Field

[0001] The present invention relates to the technical field of microfluidic chips, and in particular to a method and device for generating droplets. Background Art

[0002] In the field of biochemistry and medical testing, droplet microfluidics technology has been widely used in recent years due to its advantages such as easy fluid manipulation, miniaturization, low cost, high sensitivity and high throughput.

[0003] Microfluidic chips use external force to drive cell suspensions, reaction reagents, microbeads, separation oils, etc. to flow in a directional manner in their corresponding microchannels, thereby generating droplets. The generation of microdroplets is based on the coupling effect formed by the surface tension and fluid shear force between different phases, and requires precise control of the parameters of each phase fluid. The physicochemical properties of each phase fluid, the roughness of the fluid channel surface and the surface treatment process, the geometric structure design of the flow channel, and the precise control of the driving pressure during the experiment will affect the quality of droplet generation, and then affect the quality of the reaction collection product. In traditional microfluidic experiments, the quality of droplet generation is usually judged by the manual experience of the experimenter, which is not only time-consuming and laborious, but the quality result judgment is heavily dependent on the experience and subjective judgment criteria of the operator, and has hysteresis. Therefore, it is necessary to develop new droplet generation systems and quality control methods to solve the above problems.

[0004] Existing technologies, such as Figure 1 It is a microfluidic channel structure commonly used in microfluidic chips for generating droplets. Figure 1 a is a cross-shaped microfluidic channel. Figure 1 b is a T-type microfluidic channel. Figure 1 c is a concentric focusing microfluidic channel. In the figure, 1 is a continuous phase fluid, 2 is a discrete phase fluid, and 3 is a droplet formed in the microfluidic channel. The direction indicated by the arrow is the flow direction of each phase fluid. In scientific research and detection technology, microdroplets, as the basic unit for reaction, provide a carrier for the biochemical action of various reactants in a small space. Usually, the droplets contain various reactants required for biochemical reactions. For example, in digital PCR technology, the microdroplets contain nucleic acid templates, primers, MasterMix and other reactants required for PCR reactions; in single-cell sequencing technology, the microdroplets contain cells or cell nuclei, reaction reagents, encoding microbeads and other reactants required for the reaction. Since various reactants are randomly distributed to each microdroplet during the droplet formation process, following the basic principle of Poisson distribution, the reaction collection obtained in the microfluidic chip will also contain a variety of different types of droplets according to the different types of reactants and their components contained in the droplets. Figure 2Schematic diagram of the different droplet types obtained in single-cell sequencing experiments. Figure 2 In the figure, 10 is a schematic diagram of a microfluidic channel structure for single-cell analysis, wherein 1, 2, and 3 are respectively a continuous phase fluid, a discrete phase fluid, and microdroplets generated by the reaction. The continuous phase fluid 1 contains at least one continuous phase reactant 11, and the discrete phase fluid 2 contains a plurality of discrete phases, namely a first reactant 21, a second reactant 22, and a third reactant 23. For example, in single-cell sequencing analysis, discrete phase reactants are such as samples to be tested, coded microspheres, reaction reagents, etc. Since each reactant is randomly distributed inside the droplet during the droplet generation process, the droplets generated by the reaction will also be of different types according to the different components and quantities of the reaction liquid contained in each droplet. Figure 2 31-36 are schematic diagrams of different droplet types containing different reaction substances.

[0005] Accurate statistical analysis of the number of various types of droplets and their proportions can evaluate and control the quality of the reaction collection, thereby providing accurate quality control information for experimental operators. Since the formation and generation of droplets in microfluidic chips is a fast dynamic process, in traditional experimental methods, the above information is usually obtained after the experiment is completed by sampling the reaction collection, and the experimenter will perform microscopic analysis on a small part of the sampled droplet group under a microscope to obtain quality control results. The disadvantages of this are:

[0006] 1) It is impossible to perform real-time quality analysis of droplet generation during the experiment. It can only be performed after the experiment is completed. The analysis results are lagging. Once the reaction collection does not meet the quality standards, it can only be discarded and re-experimented;

[0007] 2) Since the reaction collection is sampled for microscopic examination, the test results have a certain deviation from the actual situation. Often, the droplet generation quality in the sampled test cannot fully represent the actual situation in the reaction collection;

[0008] 3) Sampling microscopy after the reaction requires one or more extractions of part of the reaction collection, which will result in a reduction in the expected reaction collection. If the sampling amount is too small, the accuracy of the results will also be greatly affected. This is especially significant for some relatively scarce or precious initial reactants;

[0009] 4) In some experiments, in order to ensure the accuracy of sampling quality detection, the reaction collection in the reaction collection chamber will be sampled at multiple points. The process of using the pipette to sample the reaction multiple times may also damage the normal droplets in the reaction collection chamber, thereby affecting the experimental results;

[0010] 5) The method of manual sampling and microscopic examination after the experiment adds additional manual operations, which increases the workload;

[0011] 6) The traditional method based on manual sampling microscopy is completely dependent on the experimental operator to complete, and its test results depend on the experience of manual interpretation, and the uniformity of the interpretation results cannot be guaranteed;

[0012] In order to solve the above problems, the present invention proposes a new device and method for generating droplets. Summary of the invention

[0013] A technical problem to be solved by the present invention is to provide a simple and reliable droplet generation quality control system and method, which overcomes the problem that the current manual quality control of droplet generation is unstable, time-consuming and labor-intensive.

[0014] The present invention first provides a method for generating droplets, including the steps of generating droplets inside a microfluidic chip and performing quality control at the same time.

[0015] In one embodiment according to the present invention, the quality control is to collect and calculate droplet parameters;

[0016] In one embodiment according to the present invention, the droplet parameters include one or more parameters such as size, quantity, ratio, etc. of different types of droplet generation.

[0017] In one embodiment according to the present invention, the method comprises the following steps:

[0018] S1. Sample preparation;

[0019] S2. Add samples to the chip;

[0020] S3. Start testing;

[0021] S4. Collection of reaction solution and generation of quality control results during the experiment;

[0022] In one embodiment according to the present invention, the sample preparation includes preparing a cell suspension, preparing reagents required for the reaction, and preparing a microfluidic chip for droplet generation;

[0023] In one embodiment of the present invention, the chip loading includes adding samples and various reagents required for the reaction into the microfluidic chip, and adding the microfluidic chip into a droplet generation device with image acquisition and recognition functions;

[0024] In one embodiment of the present invention, the starting test includes starting a droplet generation device to generate droplets required for the reaction in the microfluidic chip, and the image acquisition unit in the droplet generation device collects the droplet generation process in the microfluidic chip in real time and provides it to the control unit; the control unit can analyze and calculate the parameters of the size, quantity, proportion, etc. of different types of droplets according to the image acquisition information in the droplet generation process;

[0025] In one embodiment of the present invention, the reaction liquid collection and experimental process quality control result generation include collecting the reaction liquid containing droplets from the reaction collection chamber of the microfluidic chip after the reaction is completed, and generating overall experimental process quality control parameters for reference by experimenters.

[0026] In one embodiment of the present invention, the droplet parameters include droplet size, droplet generation rate, droplet volume, number of droplets, number and proportion of different types of droplets distinguished by containing different reactants, or statistical distribution of one or more of the above parameters and combinations thereof;

[0027] In one embodiment of the present invention, the presentation form of the data recording the parameters includes one or more of a table, an array, text, a chart, a video or a combination thereof.

[0028] In one embodiment according to the present invention, the control unit can accurately calculate the real-time droplet generation rate according to the characteristic droplet sequence during the droplet generation process in the microfluidic chip.

[0029] In one embodiment of the present invention, the image acquisition information of the droplet production process can be used for quality control of the reaction collection, or as a basis for real-time feedback, and the control unit adjusts the driving pressure of each reactant to optimize the experimental process and achieve the expected experimental results.

[0030] The present invention also provides a droplet generating device, which comprises: a precision pressure controller, an image acquisition unit, an image recognition and processing unit and a processor.

[0031] In one embodiment according to the present invention, the image acquisition unit focuses on an image acquisition area on the microfluidic chip to acquire and analyze experimental phenomena in this area during the experiment;

[0032] In one embodiment of the present invention, the image acquisition area is a flow channel and its surrounding area on a microfluidic chip for generating droplets.

[0033] In one embodiment according to the present invention, it further comprises a switching device disposed at the collection end of the reactants;

[0034] In one embodiment according to the present invention, the image recognition and processing unit analyzes and processes the data obtained by the image acquisition unit to obtain the morphological data of the micro-droplets;

[0035] In one embodiment of the present invention, the recognition algorithm used by the image recognition and processing unit is any one of the machine learning target detection algorithms;

[0036] In one embodiment according to the present invention, the micro-droplet morphology that can be identified by the image recognition and processing unit includes at least the size of the micro-droplet, the number of droplets, the volume of the droplets, the number of different droplet types and their ratios;

[0037] In one embodiment according to the present invention, the image recognition and processing unit can identify the three main stages of micro-droplet generation: initial, stable, and final;

[0038] In one embodiment according to the present invention, the image recognition and processing unit can identify the number of different types of reactants contained in a single micro-droplet;

[0039] In one embodiment of the present invention, the image recognition and processing unit can accurately adjust the driving pressure of the pressure controller on each phase of reactants according to the calculation results of the image acquisition information.

[0040] In one embodiment of the present invention, the camera frame rate of the image acquisition unit can be estimated and set before the experiment starts, and after the experiment starts, it can be adjusted according to the real-time droplet generation rate calculated by the control unit and match the droplet generation rate, so that the image acquisition unit can capture every newly generated droplet during the experiment inside the microfluidic chip; and the above-mentioned matching relationship can be adjusted in real time by the control unit during the entire experiment and always keep matching.

[0041] In one embodiment of the present invention, the control unit can reduce the acquisition frame rate of the camera according to a certain relationship, thereby reducing memory space and computing resources.

[0042] Finally, the present invention also provides an application of a droplet generation method or a droplet generation device, which includes one or more of the following applications: a droplet generation; b droplet detection; c guiding microfluidic chip preparation; d detection, and optionally the detection is a medical detection for non-diagnostic purposes.

[0043] In order to facilitate the image acquisition unit and the image processor to accurately identify the various reactants in the droplet, the reactants can be pre-processed. For example, without affecting the downstream reaction and mathematical analysis, the cells or cell nuclei involved in the reaction can be dyed, or pre-treated with chemical reagents to increase their contrast under the camera field of view. For the reaction reagents, one or more additives can be added according to the light source type of the image acquisition unit, so that the cells, cell nuclei, microbeads and other reactants inside the droplet can be better identified.

[0044] The beneficial effects of the above technical solution of the present invention are as follows:

[0045] A droplet generation quality control system and method of the present invention can (1) simultaneously perform quality control on a variety of droplet morphology parameters such as droplet size, empty droplet microbead package, single droplet microbead package, multiple droplet microbead packages, and total droplet amount.

[0046] (2) The droplets encapsulating different types of cells can be quality controlled without affecting the experimental process or the reaction collection after the experiment is completed.

[0047] (3) It mainly classifies data through image AI recognition methods, without any processing of the droplets and will not cause physical damage to the droplets.

[0048] (4) It has high recognition accuracy and good stability.

[0049] (5) Its quality control statistical data is generated completely automatically without the need for human intervention.

[0050] (6) The quality control statistical data can be real-time statistical data.

[0051] (7) Its quality control statistical data is presented in interactive forms such as charts and dynamic graphs, which is simple and clear.

[0052] (8) Its quality control data can be saved, archived and downloaded for subsequent research. BRIEF DESCRIPTION OF THE DRAWINGS

[0053] Figure 1 Schematic diagram of a microfluidic channel structure commonly used in microfluidic chips for generating droplets; Figure 1 (a) is a cross-shaped microfluidic channel. Figure 1 (b) is a T-type microfluidic channel. Figure 1 (c) is a concentric focusing microfluidic channel. In the figure, 1 is the continuous phase fluid, 2 is the discrete phase fluid, and 3 is the droplet formed in the microfluidic channel. The direction indicated by the arrow is the flow direction of each phase fluid.

[0054] Figure 210 is a schematic diagram of different droplet types obtained in a single-cell sequencing experiment; 11 is a schematic diagram of a microfluidic channel structure for single-cell analysis, wherein 1, 2, and 3 are continuous phase fluid, discrete phase fluid, and microdroplets generated by the reaction, respectively. Continuous phase fluid 1 contains at least one continuous phase reactant 11, and discrete phase fluid 2 contains multiple discrete phases, namely, a first reactant 21, a second reactant 22, and a third reactant 23. 31-36 are schematic diagrams of different droplet types containing different reactants.

[0055] Figure 3 A droplet generating device proposed in the present invention; the accompanying drawings are marked with 20, generating device; 40, microfluidic chip; 50, control unit 50; 41, 42a, 42b are all liquid adding chambers; 43, reaction collection chamber; 45, image acquisition unit; 44, image acquisition area.

[0056] Figure 4 (a) to (c) are schematic diagrams of different stages of droplet generation during the entire experimental process in the image acquisition area of ​​the microfluidic chip in one embodiment.

[0057] Figure 5 This is a photograph of a stable stage of micro-droplet generation captured by a camera and a labeled schematic diagram thereof in an embodiment of the present invention.

[0058] Figure 6 (a) to (e) are schematic diagrams of the dynamic description of the droplet generation process within a droplet production cycle in the stable stage.

[0059] Figure 7 (a) to (c) are schematic diagrams of the droplet sequence at the moment of transition from an unstable state to a stable state.

[0060] Figure 8 (a) to (c) are schematic diagrams for determining characteristic droplet sequences based on different reactants within the droplets.

[0061] Fig. 9 (a) to (c) are schematic diagrams of characteristic droplet sequences when the camera's capture frame rate is 1 / 2 of the droplet generation rate.

[0062] Fig.10 The invention relates to a droplet generating device with a switching device.

[0063] Fig.11 It is a droplet generation quality control report formed after the experimental process of the droplet generation device and method proposed by the present invention is completed. DETAILED DESCRIPTION

[0064] In order to make the technical problems, technical solutions and advantages to be solved by the present invention clearer, the following will be described in detail in conjunction with the accompanying drawings and specific embodiments. Regardless of the microfluidic channel structure, its basic principle is a physical phenomenon formed by the interaction between the interfacial tension between the mobile phase and the discrete phase liquid and the shear force generated during the fluid flow process. Therefore, here and below, the cross-shaped microfluidic channel is used as the main embodiment to be explained, which is intended to clearly express the technical features proposed by the present invention without limiting the scope of the present invention in any way.

[0065] Embodiment 1 Device for generating droplets

[0066] A device and method for droplet generation. In addition to being able to generate microdroplets inside a microfluidic chip, the device also integrates an image acquisition unit, an image recognition unit, and a processor for the droplet generation area on the chip. The device can be used to perform real-time analysis and statistics on the droplet generation results while the droplets are being generated, and automatically generate quality control output information after the experiment is completed for reference by experimental operators.

[0067] Figure 3 A droplet generation device proposed by the present invention. A droplet generation device 20 includes a microfluidic chip 40 for preparing droplets, and a control unit 50 for controlling the microfluidic chip to conduct experiments. The microfluidic chip 40 has a plurality of reaction liquid adding chambers 41, 42a, and 42b, which are respectively used to add the reactants required to generate droplets before the experiment begins. 43 is a reaction collection chamber of the microfluidic chip. 45 is an image acquisition unit, including various components for image acquisition functions, such as cameras, optical elements, magnifying lenses, light sources, etc. The image acquisition unit can capture a section of a microfluidic channel on the microfluidic chip for generating microdroplets and its surrounding area, such as the image acquisition area shown in the dotted box 44. The control unit 50 includes a pressure controller for driving the flow of liquid in each liquid adding chamber in the microfluidic chip, an image recognition processor, and a system processor.

[0068] Figure 4 Schematic diagram of different stages of droplet generation during the entire experimental process in the image acquisition area 44 within the microfluidic chip. Figure 4(a) is a schematic diagram of the droplet generation inside the microfluidic channel in the image acquisition area at the initial stage of the experiment. At this time, a stable droplet sequence has not yet been formed in the microfluidic chip. Each reactant is added to the reactant liquid addition chamber by the experimenter before the experiment begins, and then the microfluidic chip is placed in the droplet generation device. The experimenter inputs the command and starts it, and then turns on the pressure controller to push each reactant from the liquid addition chamber into the microfluidic channel. Since there are no reactants inside the microfluidic channel in the initial state, and there is a time sequence for each reactant to arrive at the droplet generation area, it takes a certain amount of time to reach a stable experimental state. This time is affected by factors such as the flow channel structure of the microfluidic chip, the physicochemical properties of the reactants, the surface treatment process of the microfluidic chip, and the size of the liquid driving pressure, and is usually a few seconds to tens of seconds or longer. Figure 4 (b) is a schematic diagram of droplet generation when the experimental process reaches a stable state. At this time, a stable droplet sequence is formed in the microfluidic channel, and each reactant is encapsulated in the generated microdroplets. Figure 4 (c) is the final stage of the reaction. At this time, as the reactants in the liquid addition chamber tend to be exhausted, the droplets in the microfluidic channel begin to become unstable. With the exhaustion of one or all of the reactants, the experimental process ends. In some cases, in order to avoid introducing an unstable droplet sequence at the end of the reaction into the collected product, the experimenter will terminate the experiment before reaching the end of the reaction, thereby improving the quality of the reaction collection product. In some cases, such as when the number of reaction collections is relatively small, or when the samples used in the experiment are relatively rare or precious, the experimenter will try to allow the reaction to proceed as fully as possible in order to obtain more reaction collections.

[0069] Example 2 Automatic Identification of Droplet Generation Device

[0070] A droplet generation device and method proposed in the present invention can automatically identify the initial stage, stable stage and final stage of the droplet generation process, and detect and calculate the total number of droplets, the number of each reaction substance in the droplets and their statistical proportions starting from the stable droplet generation stage in the experimental process.

[0071] Figure 5 It is a photo of the stable stage of micro-droplet generation in the image acquisition area 44 taken by the camera during the experiment. The above information is processed by the image recognition processor of the control unit 50 to form a real-time recognition and annotation map, and the image recognition processor includes a micro-droplet morphology data model and a real-time recognition and annotation algorithm, wherein the micro-droplet morphology model is obtained by training a pre-collected micro-droplet morphology image information set through a deep learning algorithm, and the real-time recognition and annotation algorithm is an algorithm that combines the real-time image information with the micro-droplet morphology model to calculate the micro-droplet morphology. Figure 5 The real-time annotation graph contains the number of reactants and their confidence.

[0072] In microfluidic chips, the generation of microdroplets is a periodic physical process, the basic principle of which is based on the interaction between the interfacial tension between the mobile phase and the discrete phase liquid and the shear force generated during the fluid flow. Figure 6 (a) to (e) in FIG. 4 are schematic diagrams for describing the dynamic droplet generation process in a droplet production cycle in the image acquisition area 44 during the stable stage. Since the droplet generation process is a continuously changing process, Figure 6 Only a few representative droplet generation dynamic feature nodes are selected for illustration purposes in the droplet generation nodes described in (a) to (e). Depending on the reactant components, the structural parameters of the microfluidic channel, the magnitude of the driving pressure of each phase, the target droplet size and other factors, the droplet generation cycle is different during the experiment. Taking the cross-shaped microfluidic channel as an example, in the usual experimental process, the cycle to complete a droplet generation is usually sub-millisecond, millisecond, tens or hundreds of milliseconds, or even seconds.

[0073] The image acquisition unit 45 uses a camera to detect the droplet generation in the image acquisition area 44 in the microfluidic chip and calculate and analyze the relevant information. Since the camera is based on a certain frame rate setting to shoot during the image capture process, if the frame rate setting of the camera is consistent with the frequency of droplet generation, that is, the time between two adjacent frames of the camera during the shooting process is consistent with the time to complete a droplet generation cycle, then each frame of the camera can just capture a newly generated droplet in turn, ensuring that the detection results of the droplet generation process are accurate and effective.

[0074] For example, assuming that the frequency of droplet generation is 200 per second, that is, a new droplet is generated every 5 milliseconds in the microfluidic channel, and the camera frame rate is set to 800 frames, that is, 800 shots per second, then the information captured by the first four frames of the camera in sequence is as follows: Figure 6 As shown in (a) to (d) (assuming Figure 6 (a) to (d) correspond to the 0, 1 / 4, 1 / 2, and 3 / 4 cycles of a droplet generation, respectively. The information captured by the camera in the fifth frame is as follows: Figure 6(e) is shown, where 301-305 are the droplet sequences that have been generated before the current droplet generation cycle, and 310 is the newly generated droplet in the current droplet generation cycle. Therefore, when the stable stage of droplet generation begins, a new droplet will be generated at the cross-shaped channel mouth position in the picture corresponding to each 4n+1th frame (n is an integer greater than or equal to 1) of the camera. If the frame rate of the camera is set to 200 frames in the above case, the camera's shooting frame rate is consistent with the frequency of droplet generation, and the next frame shot after each acquisition can just correspond to a newly generated droplet 310. It can be seen that in the latter case, not only can the generation of each new droplet be fully captured, but also the computing resources and storage space of the image processing unit are saved. In addition, when the same industrial camera is set at different frame rates, usually a high frame rate corresponds to a relatively poor imaging resolution. Selecting a camera frame rate consistent with the droplet generation frequency can improve the imaging quality while fully capturing the droplet production situation, facilitating a more accurate analysis of the reactants and their quantity inside the droplet.

[0075] If the camera frame rate is not set to an integer multiple of the droplet production frequency, for example, assuming that the droplet generation frequency is 200 per second, and the camera frame rate is set to 330 frames, that is, a new droplet is generated every 5 milliseconds in the microfluidic channel, and the time interval between each camera shot is about 3.03 milliseconds. At this time, if the experimental process is recorded with this camera setting, the actual number of qualified droplets generated cannot be accurately calculated. Assume that the droplet generation captured by the camera at the nth frame is as follows Figure 6 As shown in (a), the droplet generation captured in the n+1th frame is close to Figure 6 In the case shown in (c), no new droplets are generated, and the dispersed phase liquid at the cross-shaped channel position is stretched (new droplets have not been cut off), resulting in a decrease in the number of droplets that can be observed in the camera window. In this case, the number of qualified droplets finally counted has a certain deviation from the number of droplets actually generated in the experiment.

[0076] In some applications, if the time between two adjacent frames of the camera is much shorter than the cycle of completing a droplet generation, such as using a specific high-speed camera for shooting, the above situation will not occur. However, the price of a high-speed camera is much higher than that of an ordinary industrial camera, and it requires more advanced hardware equipment to support and analyze and calculate the information it captures. Its own volume and weight are much larger than ordinary industrial cameras, which greatly increases the cost and complexity of the system.

[0077] Example 3 Correcting the camera's acquisition frame rate

[0078] 1) In the sample addition stage before the experiment, estimate the sample addition amount of each reactant, calculate the total volume of droplets to be collected during the experiment based on the expected reaction volume of the discrete phase reactants, and then estimate the rate of droplet generation during the experiment based on the expected droplet size and reaction time. For example, in an experiment, the total amount of droplets to be collected is 150μL, the diameter of the microdroplets generated in the reaction is 90μm, and the reaction process lasts for 15 minutes. It can be calculated that the total amount of microdroplets generated by the reaction is theoretically about 3,937,000, and the average number of microdroplets generated per second is about 438.

[0079] 2) Under the condition that the image acquisition area 44 has the image capture range and clarity that meet the requirements, if the camera's settable frame rate is not less than the droplet generation frequency, the camera frame rate is set to n times (n is an integer equal to or greater than 1) the droplet generation frequency calculated above; if the camera's settable frame rate is less than the droplet generation frequency, the camera frame rate is set so that the droplet generation frequency is n times (n is an integer greater than 1) the camera frame rate. For example, taking the above experimental estimation as an example, if the camera's settable maximum frame rate in the above situation is 1000, the camera frame rate is set to 438 or 876; if the camera's settable maximum frame rate is 300, the camera frame rate can be set to 219, that is, the droplet generation frequency at this time is twice the camera frame rate setting.

[0080] 3) Start the experiment, and the reactants in each phase of the microfluidic chip are pushed into the microfluidic channel under the action of the pressure controller, and then enter the stable droplet generation stage after the initial stage of the experiment. After the experiment starts, the image acquisition unit 45 and the control unit 50 collect and analyze the droplet generation process in the microfluidic chip in real time, and calculate the actual droplet generation speed based on the characteristic droplet sequence and the initially set camera frame number;

[0081] 4) According to the actual droplet generation speed obtained by the above calculation and analysis, the frame rate setting of the camera is adjusted. At this time, the control unit 50 can also fine-tune the driving pressure of each phase of the reactant through the pressure controller at the same time, so that the speed of the above droplet generation is completely matched with the frame rate of the camera while ensuring that the experimental results meet the requirements, that is, the frequency of droplet generation is n times, 1 times or 1 / n (n is an integer greater than 1) of the camera frame rate setting. For example, according to the above analysis, the control unit calculates that the actual droplet generation rate in the microfluidic chip is 411 per second. At this time, the controller can adjust the frame rate setting of the camera to 411, which is completely consistent with the droplet generation rate; or, if the initial camera can set the frame rate to 300, the camera frame rate setting can be adjusted to 205, and the pressure controller can fine-tune the driving pressure of each phase of the reactant at the same time, so that the droplet generation rate is adjusted to 410, that is, the camera frame rate is just 1 / 2 of the droplet generation rate.

[0082] 5) The above-mentioned camera frame rate setting and the adjustment of the droplet generation rate during the experiment can be dynamically completed by the control unit 50 according to the real-time droplet generation situation, ensuring that the camera frame rate setting always maintains a certain relationship with the real-time droplet generation rate during the entire experiment;

[0083] 6) When the droplet generation reaches a stable state, the control unit can reduce the real-time acquisition frame rate of the camera to save memory and controller computing resources;

[0084] Since the frequency of the size of the droplets generated in the microfluidic channel is stable and unchanged during the stable stage of droplet generation, the image information captured by the two adjacent frames of the image acquisition unit 45 is almost overlapping, so the control unit needs to judge the different positions of the droplets at the two moments before and after according to the characteristic droplet sequence, and then accurately calculate the real-time droplet generation rate. The characteristic droplet sequence is a sequence composed of one or more droplets, which can provide a judgment basis for the control unit to calculate the real-time droplet generation rate. All sequences composed of droplets with the above characteristics can be used as characteristic droplet sequences.

[0085] The characteristic droplet sequence is the droplet sequence at the moment of transition from unstable state to stable state, such as Figure 7 shown. Figure 7 (a) to (c) in the figure are the nth, n+1th, and n+2th frames captured by the image acquisition unit, respectively, where 401 is a droplet generated in the unstable stage. A stable droplet sequence has not yet been formed in the image captured in the nth frame. Starting from the n+1th frame, a stable droplet 310 begins to form in the microfluidic chip. At this time, the unstable droplet 401 and the droplet 310 can be combined as a characteristic droplet sequence. Starting from the n+2th frame, the control unit can test the above-mentioned characteristic droplet sequence. If new droplets can be stably generated in each of the consecutive frames from the n+2th frame, the above-mentioned characteristic droplet sequence is established, and the droplet generation rate can be accurately calculated based on the time relationship between the different positions of the characteristic droplet sequence in the microfluidic channel and the frame rate. On the contrary, if no new droplets are generated after the n+3th frame, the above-mentioned characteristic droplet sequence is not established, and the control unit needs to redetermine the characteristic droplet sequence.

[0086] If the camera capture frame rate is set to n times the droplet generation rate estimated by the experiment, one or more droplets in the microfluidic chip closest to the droplet generation area, that is, the cross-shaped microfluidic channel structure, can be used as the characteristic droplet sequence. For example, if the camera capture frame rate is set to 2 times the droplet generation rate estimated by the experiment, if the image information captured by the nth frame is as follows Figure 6 As shown in (a), the image information captured by the n+1th frame is as follows Figure 6At this time, because the time interval between two adjacent frames of the camera is less than the period of a droplet generation, there is no new droplet generated in the n+1th frame image. Figure 6 The generated droplets 301, 302, etc. can be used as a characteristic droplet sequence. According to the interval time between the previous and next two frames and the droplet positions at different times, the droplet generation rate at this time can be accurately calculated.

[0087] The number and type of different reactants contained in the droplets can be used to determine the characteristic droplet sequence. For example, in single-cell omics experiments, the droplets usually contain different reactants such as cells, cell nuclei, and coded microspheres. If a droplet type is marked with an array (0,1), it means that the droplet contains zero first reactants and one second reactant. Assume that the frame rate of the camera is set to the same as the droplet generation rate estimated by the experiment. Figure 8 As shown, Figure 8 (a) to (c) are the nth, n+1th, and n+2th frames captured by the image acquisition unit. Figure 8 In a, there are three consecutive droplets 501, 502, and 503, and their droplet types are {(1,1), (0,2), (0,1)}, respectively. 511 and 512 are new droplets generated in the n+1th frame and the n+2th frame, respectively. Since the types and quantities of different reaction substances contained in the droplets are based on random distribution, the probability that the types of consecutive droplets are exactly the same and appear again in the image information captured in adjacent frames is almost zero, that is, starting from the nth frame image, the probability that three consecutive droplets appear again in the subsequent consecutive frames of images, and their types are {(1,1), (0,2), (0,1)} is almost zero. Therefore, at this time, the droplet sequence composed of three consecutive droplets 501, 502, and 503 can be used as a characteristic droplet sequence to calculate the real-time droplet generation rate.

[0088] The above calculation process is completed by the image processor and controller of the control unit, so the current droplet generation rate can be calculated in real time according to the above characteristic droplet sequence, and the camera capture frame rate and droplet generation rate can be adjusted to make the two completely match. After the calculation is completed, the control unit can also verify the above calculation results based on the image information collected at the next moment. If the calculation results are accurate and valid, the new droplet position obtained in each next droplet generation cycle should completely overlap with the position of the new droplet produced in the previous cycle, thereby verifying the accuracy of the control unit's adjustment results.

[0089] The above method for determining the characteristic droplet sequence is also applicable to the case where the camera capture frame rate is less than the droplet generation rate. For example, if the camera capture frame rate is 1 / 2 of the droplet generation rate, the above method for calculating the droplet generation rate based on the characteristic droplet sequence is still valid. The difference is that, in this case, in the image information collected by the n+1th frame, there are 2 newly generated droplets compared to the image information collected by the nth frame. Fig. 9 (a) to (c) in the figure are the nth, n+1th, and n+2th frame images captured by the image acquisition unit, respectively. Since the droplet generation rate is twice the camera capture frame rate at this time, there are two newly generated droplets 511 and 512 in the image captured in the n+1th frame compared to the nth frame, and there are two newly generated droplets 513 and 514 in the image captured in the n+2th frame compared to the n+1th frame. Therefore, when calculating the droplet generation rate, the droplet position movement at this time will also be greater, and the result does not affect the accuracy of the droplet generation rate calculation.

[0090] It should be noted that, subject to the hardware parameter settings of the camera, since the camera's capture area is limited, when using a low frame rate for droplet generation image acquisition (such as the above example, the camera is set to 1 / 2 of the droplet generation rate), the camera's capture range needs to be considered, that is, Fig. 9 The size parameters of the image acquisition area 44 shown. Otherwise, if the camera frame rate is set too low, the information containing the characteristic droplet sequence collected in the nth frame may be partially or completely lost in the image collected in the n+1 frame, resulting in an inaccurate calculation. In addition, the method of using a low frame rate for image acquisition and recognition can obtain the droplet rate calculation result with the same accuracy, and occupy less memory space and computing resources. Therefore, if the frame rate of the camera is set to a high level in the initial stage, for example, n times the droplet generation rate, the correction calculation and adjustment of the real-time droplet speed are performed, and after the adjustment is completed and the camera frame rate matches the droplet generation rate, the camera frame rate can be reduced to 1 / n of the real-time droplet generation rate, so that more storage space and computing resources can be saved while ensuring the concentration of the calculation results.

[0091] By adding a selective switching device downstream of the microfluidic chip, the system can selectively discard unqualified droplet sequences based on the information captured by the image acquisition unit, which can further increase the proportion of qualified droplets in the reaction collection. Fig.10As shown, 60 is a droplet generation device with a switching device. 61 is a reaction product collection tube, the downstream of which can be connected to a reactant collection container 63 or a waste liquid collection container 64, and switched by a switching device 62. The control unit can control the switching device 62 according to the information collected by the image, and then decide whether to collect or discard the reaction products. For example, the control unit can discard the unstable droplet sequence to the waste liquid collection container 64 in the initial stage of the reaction, and then switch the collection tube 61 to the reactant collection container 63 after the droplet generation enters the stable stage. In this way, the effective droplet ratio in the reactant collection container can be improved.

[0092] After the experiment is over, the control unit can summarize the results and generate a quality control report based on the results of the analysis and calculation during the experiment, and then the experiment operator can interpret it. The quality control report includes but is not limited to the total amount of droplets generated during the experiment, the total amount of effective droplets, the total amount of various types of droplets containing different reaction substances, and the proportion of various types of droplets and the statistical results of the proportion of effective reaction substances calculated based on the above data. Fig.11 This is an example of an experimental process quality control report generated after the experimental process is completed. The first reactant is a cell or a cell nucleus, and the second reactant is a coded microbead. According to the expected results of the experiment, a droplet containing one cell / cell nucleus and one coded microbead is considered to be a valid droplet, that is, a droplet of type (1,1). A single-reactant droplet is a droplet containing only one cell / cell nucleus or a droplet containing only one coded microbead or an empty droplet, that is, a droplet of type (0,0), (0,1), (1,0). The droplets generated in other experimental processes are multi-reactant droplets, that is, the droplet contains two different reactants and their quantities are both greater than one.

[0093] Since the droplet generation method of the present invention can capture the image information of each droplet generation in real time and perform real-time analysis and calculation of the results, while in the traditional experimental method, the quality of droplet generation can only be tested by the experimenter after the experiment is completed, which has a lag. Therefore, in addition to performing quality control analysis on droplet generation during the experiment, the droplet generation device and method proposed in the present invention can also be used for feedback control of the experimental process, thereby optimizing the quality of the collection obtained from the experiment. For example, in the reaction collection expected in the experiment, it is necessary to ensure that the number and proportion of (1,1) and (2,1) type droplets are within an expected range. After the experiment has been carried out for a period of time, the droplet generation process reaches a stable stage, and the frame rate setting of the camera is matched with the droplet generation rate after being calculated and debugged by the processor. At this time, the control unit can analyze the different types of droplets currently generated and their proportions based on the real-time experimental data. If the experimental process generates too many droplets such as (0,1), (0,0), (0,2), and the droplets of (1,1) and (2,1) types have not reached the expected experimental ratio, the control unit can fine-tune the driving pressure of each phase reactant and change its output flow rate (assuming that the reactants added before the experiment start are sufficient), that is, appropriately increase the driving pressure of the first reactant in the discrete phase, so that the volume flow rate of the first reactant output per unit time increases. In this way, during the droplet generation process, the probability of the first reactant being included in the droplet will also increase accordingly, thereby optimizing and improving the reaction process in the direction expected by the experiment. On the contrary, if the control unit calculates that the number of droplets containing the first reactant is too large based on the image acquisition information at the beginning, the control unit can appropriately reduce the driving pressure of the first reactant by adjusting the experimental parameters to meet the experimental expectations. After the control unit adjusts the driving pressure of each phase, the rate of droplet generation will also change slightly. The real-time droplet generation rate is recalculated using the above-mentioned characteristic droplet sequence and matched with the corresponding camera capture frame rate. The control unit can calculate and form new experimental output parameters to ensure the accuracy of the experimental results.

[0094] The above is a preferred embodiment of the present invention. It should be pointed out that for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the principles of the present invention. These improvements and modifications should also be regarded as the scope of protection of the present invention.

Claims

1. A method for generating droplets, characterized in that: This includes steps to perform quality control while droplets are being generated inside a microfluidic chip. The quality control is to collect and calculate the droplet parameters; The droplet parameters include one or more of the size, quantity, and ratio parameters of different types of droplet generation; S1. Sample preparation; S2. Add samples to the chip; S3. Start testing; S4. Collection of reaction solution and generation of quality control results during the experiment; The sample preparation includes preparing a cell suspension, preparing reagents required for the reaction, and preparing a microfluidic chip for droplet generation; The chip loading includes adding samples and various reagents required for the reaction into the microfluidic chip, and adding the microfluidic chip into a droplet generation device with image acquisition and recognition functions; The starting test includes starting the droplet generation device to generate droplets required for the reaction in the microfluidic chip, and using an image acquisition unit in the droplet generation device to capture the droplet generation process inside the microfluidic chip in real time and provide it to the control unit; the control unit can analyze and calculate the parameters of the size, quantity, and proportion of different types of droplets according to the image acquisition information during the droplet generation process; The control unit reduces the acquisition frame rate of the camera according to the parameter relationship, thereby reducing memory space and computing resources; The reaction liquid collection and the generation of the experimental process quality control results include collecting the reaction liquid containing the droplets from the reaction collection chamber of the microfluidic chip after the reaction is completed, and generating the overall experimental process quality control parameters for reference by the experimenter; The camera frame rate of the image acquisition unit is estimated and set before the experiment starts, and is adjusted according to the real-time droplet generation rate calculated by the control unit after the experiment starts, and matches the droplet generation rate, so that the image acquisition unit can capture every newly generated droplet during the experiment inside the microfluidic chip; and the above matching relationship can be adjusted in real time by the control unit during the entire experiment and always keep matching; the steps of the estimation and setting are as follows: 1) At the sample addition stage before the experiment, estimate the sample addition amount of each reactant, calculate the total volume of droplets to be collected during the experiment based on the expected reaction volume of the discrete phase reactants, and then estimate the rate of droplet generation during the experiment based on the expected droplet size and reaction time of the experiment; 2) If the settable frame rate of the camera is not less than the droplet generation frequency, the camera frame rate is set to n times the droplet generation frequency calculated above, where n is an integer equal to or greater than 1; if the settable frame rate of the camera is less than the droplet generation frequency, the camera frame rate is set so that the droplet generation frequency is n times the camera frame rate, where n is an integer greater than 1.

2. The method for droplet generation according to claim 1, characterized in that: The droplet parameters include droplet size, droplet generation rate, droplet volume, number of droplets, number and ratio of different types of droplets distinguished by containing different reactants, or statistical distribution of one or more of the above parameters and combinations thereof; The presentation form of the data recording the parameters includes one or more of a table, an array, text, a chart, a video or a combination thereof.

3. The method for droplet generation according to claim 1, characterized in that: The control unit can accurately calculate the real-time droplet generation rate according to the characteristic droplet sequence in the droplet generation process in the microfluidic chip.

4. The method for generating droplets according to claim 1, characterized in that: The image acquisition information during the droplet generation process can be used for quality control of the reaction collection, or as a basis for real-time feedback, and the control unit adjusts the driving pressure of each reactant to optimize the experimental process and achieve the expected experimental results.

5. A droplet generation device, comprising a microfluidic chip having a reaction collection chamber; characterized in that: Also includes: Precision pressure controller, image acquisition unit, image recognition and processing unit and processor; The image acquisition unit focuses on the image acquisition area on the microfluidic chip to collect and analyze experimental phenomena in this area during the experiment; The image acquisition area is the flow channel and its surrounding area on the microfluidic chip for generating droplets; The camera frame rate of the image acquisition unit is estimated and set before the experiment starts, and is adjusted according to the real-time droplet generation rate calculated by the control unit after the experiment starts, and matches the droplet generation rate, so that the image acquisition unit can capture every newly generated droplet during the experiment inside the microfluidic chip; and the above matching relationship can be adjusted in real time by the control unit during the entire experiment and always keep matching; the steps of the estimation and setting are as follows: 1) At the sample addition stage before the experiment, estimate the sample addition amount of each reactant, calculate the total volume of droplets to be collected during the experiment based on the expected reaction volume of the discrete phase reactants, and then estimate the rate of droplet generation during the experiment based on the expected droplet size and reaction time of the experiment; 2) If the settable frame rate of the camera is not less than the droplet generation frequency, the camera frame rate is set to n times the droplet generation frequency calculated above, where n is an integer equal to or greater than 1; if the settable frame rate of the camera is less than the droplet generation frequency, the camera frame rate is set so that the droplet generation frequency is n times the camera frame rate, where n is an integer greater than 1.

6. The droplet generating device according to claim 5, characterized in that: It also includes a switching device arranged at the end of the reaction collection chamber.

7. The droplet generating device according to claim 5, characterized in that: The image recognition and processing unit analyzes and processes the data obtained by the image acquisition unit to obtain the morphological data of the micro-droplets; the recognition algorithm adopted by the image recognition and processing unit is any one of the machine learning target detection algorithms; the micro-droplet morphology that can be recognized by the image recognition and processing unit at least includes the size of the micro-droplets, the number of droplets, the volume of droplets, the number of different droplet types and their proportions; The image recognition and processing unit can identify the three main stages of micro-droplet generation: initial, stable, and final; the image recognition and processing unit can identify the number of different types of reactants contained in a single micro-droplet; The image recognition and processing unit can accurately adjust the driving pressure of the pressure controller on each phase of reactants according to the calculation results of the image acquisition information.

8. A method for generating droplets according to any one of claims 1 to 4 or a use of a droplet generating device according to any one of claims 5 to 7, characterized in that: The application includes one or more of the following applications: a. droplet generation; b. droplet detection; c. guiding microfluidic chip preparation; d. detection, wherein the detection is a medical detection for non-diagnostic purposes.

Images