Multiplex high-throughput analysis system and method based on microfluidics, and microfluidic chip

Abstract

The invention relates to a three-dimensional multiple high-throughput analysis system and method based on microfluidics, and the system comprises a microfluidic chip integrating a temperature control reaction area and a detection area, an X-ray plate scanning module, a lens-free imaging unit, a synchronous control system and a data processing module. The device breaks through the bottleneck of plane limitation and spectrum crosstalk in the traditional technology, realizes high flux, three-dimensional space resolution and multi-parameter comprehensive analysis (multi-channel fluorescence + three-dimensional form + reaction kinetics), and has the characteristics of compact equipment, low cost, easiness in maintenance and the like; and an efficient and flexible analysis tool is provided for the fields of biomedical research, clinical diagnosis, drug screening and the like.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of biomedical engineering and analytical chemistry, in particular to a microfluidic-based multi-high-throughput analysis system, method and microfluidic chip. BACKGROUND

[0002] In the field of modern life science research, clinical diagnosis and drug development, high-throughput, multi-dimensional, three-dimensional dynamic analysis of biological samples has become a core requirement. For example, single-cell omics research requires simultaneous analysis of gene expression, protein localization and spatial structure of cells; drug screening requires real-time monitoring of the dynamic effects of compounds on cell morphology and function; and clinical diagnosis requires rapid detection of the spatial and temporal distribution of multiple markers. However, the existing mainstream analysis techniques (such as traditional flow cytometry and mature flow cytometry) still have significant limitations in information dimension, throughput, device integration and functional expandability, making it difficult to meet the growing research needs.

[0003] Specifically, the existing technology has the following shortcomings: 1. Single information dimension, lack of three-dimensional and dynamic analysis capability. Traditional flow cytometry mainly relies on fluorescence / scattering intensity to provide two-dimensional information, and cannot capture the three-dimensional morphology of biological samples (such as the spatial structure of cell clusters) and the dynamic changes during the reaction process (such as the real-time progress of PCR amplification); mature flow cytometry has added two-dimensional morphology + fluorescence detection, but is still limited to the plane level and cannot analyze three-dimensional spatial structure, resulting in incomplete understanding of complex biological systems; 2. Difficulty in balancing throughput and multi-parameter. Traditional flow cytometry has extremely high single-column throughput (up to 100,000 levels / second), but is limited in multi-parameter parallel detection; mature flow cytometry is limited by single-column structure and has low throughput (thousands of levels / second), making it difficult to meet the needs of high-throughput screening scenarios. Both of them have not achieved effective coordination of "high throughput" and "multi-parameter"; 3. Large device size, high cost, difficult to integrate. Traditional flow cytometry is equipped with large optical systems, photomultiplier tubes (PMT) and other complex components, which are bulky and expensive; mature flow cytometry is slightly smaller in size, but still requires expensive PMT and optical components, with moderate cost. This design hinders its integration on small platforms such as microfluidic chips; 4. Lack of dynamic reaction monitoring capability. Traditional and mature flow cytometry do not support real-time monitoring of biochemical reactions (such as PCR amplification, cell incubation, membrane potential changes, etc.), and cannot capture dynamic information during the reaction process, limiting their application in real-time analysis scenarios; 5. Simple data presentation, insufficient analysis depth. Traditional flow cytometers mainly output results in numerical form (such as FCS files), lacking intuitive images and three-dimensional models; mature flow cytometers can only provide two-dimensional images, unable to generate three-dimensional reconstruction or reaction kinetics curves, resulting in insufficient comprehensive and in-depth data analysis; 6. Poor application expansion and maintenance convenience. The channel expansion of traditional flow cytometers is limited, making it difficult to flexibly add detection parameters (such as adding a fluorescence channel); the application scenarios of mature flow cytometers are also relatively fixed. In addition, the maintenance process of both is cumbersome (such as PMT calibration, optical system cleaning), and the operation is complex, increasing the use cost and time.

[0004] In summary, the existing flow cytometer technology has obvious limitations in information dimension, throughput, integration, and functional expansion, making it difficult to meet the needs of high-throughput, multi-dimensional, and three-dimensional dynamic analysis in modern life science research. Therefore, it is urgent to develop a new type of analysis system that can integrate microfluidic, three-dimensional imaging, and multi-dimensional detection technology to achieve high-throughput, low-cost, and automated three-dimensional multi-dimensional analysis to fill the current technical gap. SUMMARY

[0005] The purpose of the present application is to provide a multi-dimensional high-throughput analysis system and method based on microfluidics to solve the problem of the lack of a multi-dimensional high-throughput analysis system and method based on microfluidics.

[0006] To achieve the above-mentioned purpose, the present application provides the following technical solutions: The multi-dimensional high-throughput analysis system based on microfluidics comprises: A microfluidic chip integrates a temperature-controlled reaction area and a detection area, the reaction area is used for biochemical reactions of biological samples, and the detection area is provided with a single-layer or multi-layer microchannel structure; A light sheet scanning module comprising a multi-wavelength laser source, a light sheet shaping assembly, and a high-speed galvanometer is used to form a scanning light sheet with a thickness of 1-5 μm in the detection area and perform tomographic scanning along the Z-axis; A lens-free imaging unit comprising a high-resolution sCMOS sensor and a mosaic filter array, the mosaic filter array is periodically arranged on the surface of the sensor, and is used for multi-spectral spatial encoding of excitation signals; A synchronous control system realizes hardware-level synchronization of light sheet scanning, image acquisition, and temperature control through FPGA or high-performance SoC; A data processing module is used for multi-channel decoding, three-dimensional reconstruction, and reaction kinetics analysis of the mosaic encoded image.

[0007] Further, the microfluidic chip integrates micro-heating electrodes or Peltier elements in the temperature-controlled reaction area, and realizes partition temperature control with an accuracy of ±0.1℃ through temperature sensor feedback, supporting nucleic acid amplification, protein incubation or enzyme reaction.

[0008] Further, in the light sheet scanning module: The light sheet shaping assembly is a cylindrical lens or a Powell prism; the high-speed galvanometer has a response frequency of >1 kHz, a scanning range of 50-200 μm, and 50-200 Z layers are collected in a single scanning; The light sheet scanning and sCMOS exposure are strictly synchronized through a TTL signal to avoid interlayer signal aliasing.

[0009] Further, in the lens-free imaging unit: The sCMOS sensor has a resolution of ≥2048×2048 pixels, a pixel size of <6 μm, and a frame rate of >500 fps; The distance between the bottom surface of the detection area and the sensor is <100 μm; The mosaic filter array is periodically arranged in 2×2 or 4×4 super-pixel units, supports ≥10 fluorescence channels, and supports full-channel morphological acquisition.

[0010] Further, the super-pixel unit of the mosaic filter array comprises FITC / PE / APC three-color fluorescence filters and a full-channel filter, each filter has a bandwidth of 20-30 nm, and a cutoff depth OD of >4.

[0011] Further, the data processing module performs the following operations: According to the arrangement rule of the mosaic filter array, the multi-channel signals are separated; Through interpolation or deep learning super-resolution algorithm, two-dimensional images of each channel are reconstructed; According to the Z-axis tomographic sequence, a three-dimensional multi-channel volume data is synthesized, and the voxel resolution is 2.4 μm in X / Y direction and 2 μm in Z direction; The reaction kinetics curve and three-dimensional morphological parameters of the biological sample are generated.

[0012] The three-dimensional multi-reagent high-throughput analysis method based on microfluidics comprises the following steps: S1: introducing a biological sample into a reaction area of a microfluidic chip to perform a temperature-controlled biochemical reaction; S2: when the sample flows through the detection area, Z-axis tomographic excitation is performed through the light sheet scanning module; S3: using the mosaic filter array to encode the excitation signal in multiple spectra, and synchronously collecting the encoded images through the lens-free sCMOS sensor; S4: multi-channel decoding and three-dimensional reconstruction are performed on the encoded images to obtain multi-parameter dynamic data of the sample.

[0013] Further, the synchronization control of the light sheet tomography excitation and image acquisition includes: The high-speed galvanometer moves to the target Z layer and triggers the sCMOS exposure after being stabilized; After the exposure of each layer is completed, the galvanometer steps to the next layer, and the cycle is repeated until all tomographic acquisitions are completed; The galvanometer settling time and the exposure window are strictly overlapped, and the motion blur is less than 1 μm.

[0014] Further, the multi-channel decoding in step S4 includes: According to the periodic arrangement rule of the mosaic filter array, the original image pixels are allocated to the corresponding fluorescence / white light channels; The spatial interpolation reconstruction is performed on the sparse sampling points of each channel to form a continuous two-dimensional image; All multi-channel images of the tomographic layers are fused to generate three-dimensional volume data.

[0015] A microfluidic chip applied to a microfluidic multi-high-throughput analysis system, comprising: The substrate is PDMS, glass or silicon, which is prepared by micro-nano processing, soft lithography or 3D printing; the microchannel has a width of 50-200 μm and a height of 50-200 μm, and supports single-layer or multi-layer flow structure; the reaction area is integrated with a thin film heater and a temperature sensor to realize closed-loop temperature control with an accuracy of ±0.1 ℃.

[0016] Compared with the prior art, the microfluidic chip has the following beneficial effects: (1) Information dimension upgrade: breaking through the two-dimensional limitation of the traditional flow cytometer, through light sheet tomography scanning and three-dimensional reconstruction technology, the comprehensive information acquisition of "multi-channel fluorescence + three-dimensional morphology + reaction kinetics" is realized, which can analyze the spatial structure (such as three-dimensional distribution of cell clusters) and reaction dynamics (such as PCR amplification process) of biological samples, and provides complete cognition for complex biological research; (2) High-throughput and multi-parameter cooperation: adopting a "single column / multi-column parallel" architecture, combining the integrated advantages of the microfluidic chip, realizing high (single column / multi-column parallel) throughput (10 million level / second), and supporting multi-channel fluorescence detection (≥10), which meets the dual requirements of "high throughput" and "multi-parameter", and is suitable for large-scale analysis scenes such as single-cell omics and drug screening; (3) Equipment miniaturization and low cost: abandoning the large optical system and PMT of the traditional flow cytometer, adopting a "lens-free, PMT-free" design, directly coupling the mosaic filter array and the sCMOS sensor, greatly reducing the volume (chip-level integration) and the manufacturing cost (about 1 / 3-1 / 2 of the traditional equipment), facilitating the combination with the microfluidic chip and the portable platform, and promoting the popularization of the technology; (4) Dynamic reaction monitoring capability: through the microfluidic chip temperature control reaction zone and the synchronous control system, the biochemical reaction process (such as PCR amplification curve, cell morphology change after drug treatment) is tracked in real time, the "reaction-detection" integration is realized, and the real-time monitoring scene is suitable for molecular diagnosis, drug screening and the like. BRIEF DESCRIPTION OF DRAWINGS

[0017] Figure 1 It is the first perspective structural diagram of the microfluidic-based multiplex high-throughput analysis system of the application. Figure 2 It is the second perspective structural diagram of the microfluidic-based multiplex high-throughput analysis system of the application. Figure 3 It is the structural diagram of the mosaic filter array of the microfluidic-based multiplex high-throughput analysis system of the application. Figure 4 It is the flow chart of the microfluidic-based three-dimensional multiplex high-throughput analysis method of the application. Figure 5 It is the basic parameter comparison diagram of the microfluidic-based multiplex high-throughput analysis system of the application and the traditional flow cytometer. Figure 6 It is the comparison diagram of the microfluidic-based multiplex high-throughput analysis system of the application and the traditional flow cytometer in the field of multiplex protein detection.

[0018] In the figure: 1, monochromatic light source (excitation light); 2, cylindrical lens; 3, sCMOS sensor; 4, mosaic filter array; 5, PDMS chip; 6, ITO glass. DETAILED DESCRIPTION

[0019] The technical solutions in the embodiments of the application will be clearly and completely described below with reference to the drawings in the embodiments of the application. Obviously, the described embodiments are only part of the embodiments of the application, rather than all the embodiments of the application. Based on the embodiments in the application, all other embodiments obtained by those skilled in the art without creative labor fall within the protection scope of the application.

[0020] Please refer to Figure 1 , Figure 2 and Figure 3 The microfluidic-based three-dimensional multiplex high-throughput analysis system of the application mainly comprises a microfluidic chip, an optical sheet scanning module, a lens-free imaging unit, a synchronous control system and a data processing module.

[0021] The microfluidic chip is made of PDMS chip 5, glass or silicon as a substrate, and is prepared by micro-nano processing technology, and integrates a temperature control reaction area and a detection area. The reaction area is internally provided with a micro-heating electrode or a Peltier element, and cooperates with a temperature sensor to realize partition temperature control with an accuracy of ±0.1 ℃, and supports biochemical reactions such as PCR amplification and protein incubation. The detection area is provided with a single-layer or multi-layer micro-channel structure, the channel width and height are both 50-200 μm, and is used for sample transmission and detection. The light sheet scanning module comprises a multi-wavelength laser source (such as 405 nm, 488 nm, 561 nm, 638 nm, etc.), a light sheet shaping assembly (cylindrical lens 2 or Powell prism), and a high-speed galvanometer (response frequency > 1 kHz, scanning range 50-200 μm), which can convert the laser beam into a thin light sheet with a thickness of 1-5 μm, and perform tomographic scanning on the sample in the detection area along the Z-axis direction. The lens-free imaging unit adopts a high-resolution sCMOS sensor 3 (resolution ≥ 2048 × 2048 pixels, frame rate > 500 fps), and a mosaic filter array 4 is periodically arranged on the surface of the sCMOS sensor 3, which is taken as a unit of 2 × 2 or 4 × 4 super-pixel units, and comprises FITC, PE, APC three-color fluorescence filters and an all-pass filter (ITO glass 6), each filter has a bandwidth of 20-30 nm and a cutoff depth OD > 4, and realizes direct multi-spectral spatial coding of the excitation signal. The synchronous control system is designed based on FPGA or high-performance SoC, and the working pace of the light sheet scanning, image acquisition and temperature control modules is coordinated through hardware-level instructions, so that the galvanometer settling time and the sCMOS sensor 3 exposure window are strictly overlapped (motion blur < 1 μm). The data processing module is equipped with a special algorithm library, supports multi-channel decoding, interpolation super-resolution reconstruction and three-dimensional body data synthesis, and finally outputs numerical values, images, three-dimensional models and reaction kinetics curves.

[0022] Working principle: refer to Figure 4 When the system works, first, the biological sample (such as cell suspension, nucleic acid mixture) to be analyzed is injected into the reaction area of the microfluidic chip, the temperature control module is started to maintain a preset temperature (such as 37 ℃), and the sample completes biochemical reactions (such as PCR amplification, antigen-antibody combination) in the reaction area. Subsequently, the sample enters the detection area with the fluid flow, at this time, the multi-wavelength laser source of the light sheet scanning module emits monochromatic light source (excitation light) 1, forms a thin light sheet through the light sheet shaping assembly, and the high-speed galvanometer drives the light sheet to scan the detection area layer by layer along the Z-axis direction. After the galvanometer is stabilized before each layer scanning, the sCMOS sensor 3 is triggered to expose, so as to ensure that the light sheet and the sample fully act. The sCMOS sensor 3 of the lens-free imaging unit is closely attached to the bottom surface of the detection area (spacing < 100 μm), receives the fluorescence signal emitted by the sample, the mosaic filter array 4 encodes the fluorescence signals of different wavelengths into a spatially distributed pixel pattern, and the sCMOS sensor 3 synchronously collects the encoded images at a frame rate of > 500 fps. The synchronous control system strictly controls the galvanometer step and the exposure timing through TTL signals, and ensures the integrity of each layer scanning.

[0023] After receiving the encoded images, the data processing module first allocates the original image pixels to the corresponding fluorescence / white light channels (excitation light 1) according to the periodic arrangement rules of the mosaic filter array 4, then performs spatial interpolation reconstruction on the sparse sampling points of each channel to form continuous two-dimensional images; then, all the two-dimensional images of the layers are fused in the Z-axis tomographic sequence to generate three-dimensional multi-channel volume data (voxel resolution X / Y direction 2.4 μm, Z direction 2 μm); finally, through the time series analysis of the volume data, the fluorescence intensity change, three-dimensional morphological parameters and reaction kinetics curve (such as PCR amplification curve, cell morphological change rate) of the sample are extracted, and multi-parameter dynamic evaluation is realized.

[0024] Core algorithm principle: 1. High-speed galvanometer and light sheet three-dimensional tomographic scanning principle and process Principle: The high-speed galvanometer (Galvo Scanner) controls the rapid scanning of the laser light sheet in the vertical direction of the microfluidic channel. By gradually changing the position of the light sheet, the sample is excited at different layers to obtain three-dimensional tomographic information.

[0025] Process: The galvanometer controller adjusts the position of the light sheet in the Z-axis periodically according to the set step or continuous trajectory. Each position corresponds to one excitation and imaging, and the sCMOS synchronously collects the encoded images of the current layer. The galvanometer scanning period and the sCMOS exposure period are strictly synchronized to ensure that each exposure corresponds to only one light sheet layer and avoid layer signal aliasing. The multi-layer data is collected in time sequence, and the three-dimensional spatial structure is reconstructed by the rear-end algorithm; 2. Timing control logic of galvanometer and sCMOS exposure

[0026] Timing control points: each time the galvanometer moves to the specified Z layer, the sCMOS exposure signal is triggered; after exposure, the galvanometer moves to the next layer and triggers the next exposure; after all layers are scanned, the three-dimensional tomographic data of the particles / cells are synthesized.

[0027] Core timing process:

[0028] (1) Initialization: set the initial position of the galvanometer, exposure parameters, number of scanning layers and step.

[0029] (2) Loop scanning: the galvanometer moves to the specified Z layer and waits for stabilization (settling time). Send the sCMOS sensor 3 exposure trigger signal, and the sCMOS sensor 3 collects the encoded images of the layer. After exposure, read the image data. The galvanometer moves to the next layer and repeats until all layers are collected.

[0030] (3) After each sample tomographic collection is completed, three-dimensional reconstruction is performed.

[0031] 3. Synchronization mechanism

[0032] Hardware trigger synchronization: galvanometer controller and CMOS are connected by TTL / level signal line to ensure accurate timing.

[0033] FPGA / microcontroller coordination: use FPGA or high-performance microcontroller as the main control to achieve nanosecond-level synchronization.

[0034] Galvanometer steps to Z1, and triggers CMOS exposure EXP1 after stabilization; move to Z2 after EXP1 ends, and cycle in turn.

[0035] Exposure window and galvanometer settling window overlap strictly to avoid motion blur.

[0036] 4. Mosaic filter array arrangement rules

[0037] Arrangement: cover the sCMOS sensor surface with a periodic micro-filter array, each period containing all channels (e.g. 3-color fluorescence + 1 white light).

[0038] Unit array: each period contains 4 filters (e.g. R, G, B, W), arranged in 2x2 or linear 4-pixel groups.

[0039] Array period: the array period matches the sensor pixel size to ensure uniform spatial sampling.

[0040] Design principles: high transmittance, strong band selectivity, and uniform arrangement.

[0041] 5. Multi-channel signal decoding algorithm principle

[0042] Decoding process: (1) Raw acquisition: sCMOS sensor 3 acquires the original image encoded by the mosaic filter array 4.

[0043] (2) Channel separation: according to the filter arrangement rules, assign each pixel or pixel group to the corresponding fluorescence / white light channel.

[0044] (3) Spatial recombination: use interpolation or super-resolution reconstruction algorithms to reconstruct discrete pixels in each channel into continuous images.

[0045] (4) Three-dimensional reconstruction: for each layer, reconstruct multi-channel images, and then synthesize three-dimensional multi-channel volume data according to the sequence of sections.

[0046] Common algorithms: nearest neighbor interpolation, bilinear interpolation, deep learning super-resolution reconstruction; combined with the sequence of sections, realize multi-channel fusion of three-dimensional volume data.

[0047] 6. Three-dimensional reconstruction resolution calculation

[0048] (1) Parameter assumptions

[0049] Z-axis scan step = 2 pm (galvanometer + optical slice stepping / continuous scanning)

[0050] Number of scanning layers N = 50 (covering 100 pm thickness)

[0051] (2) Three-dimensional voxel resolution: X x Y x Z = 1548 x 1040 x 50

[0052] Physical size of each voxel: XY plane: 2.4 pm x 2.4 pm; Z direction: 2 pm

[0053] (3) Three-dimensional resolution limit. Limited by lensless reconstruction algorithm, system PSF, etc., the actual effective resolution is generally slightly lower than the physical pixel (e.g., 3-4 pm in-plane, 2-3 pm axial, which needs to be calibrated).

[0054] 7. Multi-spectral crosstalk rate estimation

[0055] Crosstalk sources: bandwidth and cutoff depth of the mosaic filter array 4 (i.e., spectral selectivity of the filter); Leakage of signals from adjacent pixels (e.g., optical diffusion, PSF broadening); spectral overlap of the fluorescent probes themselves Typical parameters: high-quality interference filter, bandwidth 20-30 nm, cutoff depth OD > 4 (transmission ratio < 0.01%), laboratory measured multi-color crosstalk rate generally < 2% (i.e., the signal leakage of adjacent channels accounts for < 2% of the main channel signal) Using orthogonal fluorescent probes + deep learning color separation algorithm, the crosstalk can be reduced to < 1%. Estimation formula: Crosstalk rate

[0056] Where, is the signal introduced by the filter transmission / fluorescent overlap for non-main channel pixels, is the effective signal of the main channel pixel.

[0057] To verify the performance advantages of the system, a comparison experiment was conducted between the traditional flow cytometer (representing existing mature technology) and the system. Figure 5

[0058] (I) Basic parameter comparison

[0059] ​In terms of information dimension, the system can simultaneously obtain multi-channel fluorescence signals, three-dimensional spatial structure and dynamic changes of fluorescence intensity during the reaction process (such as real-time curve of PCR amplification), while the traditional flow cytometer can only provide two-dimensional fluorescence / scattering intensity data. Although the mature flow cytometer increases two-dimensional shape detection, it cannot analyze three-dimensional structure.

[0060] In terms of flux performance, the system realizes high (single / multi-column parallel) flux (10 million level / second) through the "single / multi-column parallel" architecture, which balances the needs of high flux and multi-parameter detection. Although the traditional flow cytometer has extremely high single-column flux (10 million level / second), the efficiency is limited when multiple parameters are detected in parallel. The mature flow cytometer has a flux of only several thousand levels / second. In terms of equipment integration, the system adopts a "lens-free, PMT-free" design, which is compact in size (such as chip-level integration) and low in cost (about 1 / 3 of the traditional equipment), and is convenient to combine with microfluidic chips. Due to the large optical system and PMT, the traditional flow cytometer is bulky and expensive. Although the mature flow cytometer is slightly smaller in size, it still needs expensive components. In terms of dynamic reaction monitoring, the system supports real-time tracking of PCR, incubation and other biochemical reaction processes, while the traditional and mature flow cytometers do not have this function.

[0061] In terms of three-dimensional spatial resolution, the system realizes high (new layer scanning, three-dimensional reconstruction) spatial resolution through light sheet scanning and new layer scanning technology. The traditional and mature flow cytometers cannot perform three-dimensional imaging.

[0062] In terms of data presentation richness, the system outputs numerical values, images, three-dimensional models and reaction kinetics curves, making the analysis results more intuitive and comprehensive. The traditional flow cytometer mainly outputs numerical values (FCS files), and the mature flow cytometer only provides two-dimensional images.

[0063] In terms of application expansion, the system is designed flexibly, and can easily expand the number of fluorescence channels and imaging functions by replacing the mosaic filter array 4 filter array or adjusting the microfluidic chip structure. The channel expansion of the traditional flow cytometer is limited, and the application scenario of the mature flow cytometer is fixed.

[0064] In terms of maintenance convenience, the system adopts an "easy-to-maintain, chip-dismantlable" structure, and the core components can be quickly replaced, making daily maintenance simple. The maintenance of the traditional flow cytometer is complicated, and the maintenance of the mature flow cytometer is relatively complex.

[0065] The experimental results show that the system is significantly superior to the prior art in terms of information dimension, flux, integration, functional expansion, etc., and can meet the efficient analysis needs in the fields of biomedical research, clinical diagnosis and drug screening.

[0066] (II) Multiple protein detection

[0067] ReferenceFigure 6 In the multiple protein detection project, compared with the performance of the existing traditional flow cytometer, the microfluidic-based multiple high-throughput analysis system of the application shows significant advantages.

[0068] In terms of the number of multiple detection channels, although the traditional instrument can reach 8-20 channels (up to 30 channels for high-end models, but limited by the number of filters), the system can realize 4-8 channels (theoretically can be expanded to more than 10) through the design of the mosaic filter array 4, and the channel expansion flexibility is stronger.

[0069] In terms of multiple detection types, in addition to multi-color fluorescence, the system can also realize white light pattern analysis and three-dimensional structure reconstruction, and the information dimension is richer; in terms of light crosstalk rate, the traditional instrument causes crosstalk to increase (2-8%, which needs to be strictly compensated) due to the increase in channels, and the system controls the crosstalk to be <2% through the algorithm optimization of the high-quality mosaic filter array 4, which significantly improves the detection specificity.

[0070] In terms of sample consumption and throughput, although the traditional instrument has high throughput (tens of millions of cells per second), the system maintains low consumption while achieving a throughput level of tens of millions through optimization of 3D scanning speed; in terms of quantitative sensitivity, the traditional instrument has reached the single molecule level, and the system also maintains high sensitivity due to higher imaging sensitivity and algorithm improvement; in terms of specificity (cross reaction), the traditional instrument is easily affected by light density and stray light (single cell level), and the system significantly improves the specificity through spectral-spatial separation technology.

[0071] In terms of data dimension, the traditional instrument is mostly 1D / 2D point data (single cell level), and the system can obtain multi-parameter three-dimensional spatial data (voxel level / subcell level), and the analysis depth is deeper; in terms of protein / nucleic acid multi-parameter detection, although the traditional instrument can combine the number of channels and fluorescent probes, the system supports more comprehensive joint analysis; in terms of self-identification classification ability, although the traditional instrument is strong, it needs complex compensation at high throughput, and the system has better performance due to stronger segmentation ability and spatial collaborative analysis.

[0072] In terms of automation and portability, the traditional instrument is a desktop, large-scale equipment with high automation but poor portability, and the system adopts chip-based and integrated design, which is easy to realize automation and convenient to carry. These comparison results fully prove that the measurement system of the patent is significantly superior to the traditional flow cytometer in multiple detection, three-dimensional imaging, high throughput and portability, and can meet the complex analysis needs of modern biomedical research.

[0073] (Three) Multiple nucleic acid detection

[0074] In multiple nucleic acid detection projects, traditional flow cytometry can be combined with in situ hybridization (Flow-FISH), but the actual multiple capacity is limited by the number of fluorescence channels and probe design. Under high channel number, signal crosstalk and sensitivity loss are obvious.

[0075] The method of the present application supports multiple nucleic acid probes (such as multi-color FISH or molecular barcodes), and can distinguish different signals in three-dimensional space, realizing complex analysis such as spatial transcriptome. Multi-color decoding is completed by the joint action of the mosaic filter array 4 and the algorithm, and the theoretical expansion of the channel is higher.

[0076] Multiple detection accuracy and specificity: the specificity and accuracy of traditional flow multiple detection depend on spectral compensation and instrument calibration, the more the number of channels, the greater the risk of error. Multiple data is a "single cell parameter table", and the spatial information is missing. The multiple detection accuracy of the method of the present application is high, and the three-dimensional space can assist in removing crosstalk and misjudgment (such as the typing of different types of cell clusters). The output is "single cell / cluster multi-channel three-dimensional body data", which is rich and intuitive.

[0077] In summary, through technical innovation, the present application effectively solves the limitations of existing flow cytometers in information dimension, throughput, integration, functional expansion, etc., and provides an efficient, flexible and low-cost analysis tool for biomedical research, clinical diagnosis and drug screening, etc. It has important academic value and application prospect.

[0078] Although embodiments of the present application have been shown and described, it will be understood by those having ordinary skill in the art that various changes, modifications, substitutions and alterations can be made thereto without departing from the principles and spirit of the present application, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A microfluidic-based high-throughput multivariate analysis system, characterized in that, include: A microfluidic chip integrates a temperature-controlled reaction zone and a detection zone. The reaction zone is used for biochemical reactions of biological samples, and the detection zone has a single-layer or multi-layer microchannel structure. The optical sheet scanning module includes a multi-wavelength laser source, an optical sheet shaping component, and a high-speed galvanometer, used to form a scanning optical sheet with a thickness of 1~5μm in the detection area and perform tomographic scanning along the Z-axis; The lensless imaging unit includes a high-resolution sCMOS sensor (3) and a mosaic filter array (4), wherein the mosaic filter array (4) is periodically arranged on the sensor surface for multispectral spatial encoding of the excitation signal; The synchronous control system achieves hardware-level synchronization of optical sheet scanning, image acquisition, and temperature control through FPGA or high-performance SoC; The data processing module is used for multi-channel decoding, 3D reconstruction, and reaction kinetic analysis of mosaic-encoded images.

2. The microfluidic-based high-throughput analysis system according to claim 1, characterized in that, The microfluidic chip integrates micro-heating electrodes or Peltier elements in its temperature-controlled reaction zone, and achieves zoned temperature control with an accuracy of ±0.1℃ through temperature sensor feedback, supporting nucleic acid amplification, protein incubation, or enzyme reactions.

3. The microfluidic-based high-throughput analysis system according to claim 1, characterized in that, In the optical sheet scanning module: The light sheet shaping component is a cylindrical lens (2) or a Powell prism; the high-speed galvanometer response frequency is >1kHz, the scanning range is 50~200μm, and 50~200 Z layers are acquired in a single scan; Light film scanning and sCMOS exposure are strictly synchronized via TTL signals to avoid signal aliasing between layers.

4. The microfluidic-based high-throughput analysis system according to claim 1, characterized in that, In the lensless imaging unit: sCMOS sensor (3) resolution ≥ 2048×2048 pixels, pixel size < 6μm, frame rate > 500fps; The minimum distance between the bottom surface of the detection area and the sensor is <100μm; The mosaic filter array (4) is arranged periodically in 2×2 or 4×4 superpixel units, supporting ≥10 fluorescence channels and full-channel morphological acquisition.

5. The microfluidic-based high-throughput analysis system according to claim 1, characterized in that, The superpixel unit of the mosaic filter array (4) includes a FITC / PE / APC three-color fluorescent filter and an all-pass filter. Each filter has a bandwidth of 20~30nm and a cutoff depth OD>4.

6. The microfluidic-based high-throughput analysis system according to claim 1, characterized in that, The data processing module performs the following operations: Separate multi-channel signals according to the arrangement rules of the mosaic filter array (4); Reconstruct two-dimensional images of each channel using interpolation or deep learning super-resolution algorithms; Three-dimensional multi-channel volume data were synthesized based on the Z-axis tomographic sequence, with voxel resolution of 2.4 μm in the X / Y direction and 2 μm in the Z direction; The reaction kinetic curves and three-dimensional morphological parameters of the generated biological samples.

7. A three-dimensional multivariate high-throughput analysis method based on microfluidics, characterized in that, Includes the following steps: S1: Introduce biological samples into the reaction zone of the microfluidic chip to perform temperature-controlled biochemical reactions; S2: When the sample flows through the detection area, Z-axis tomography is excited by the optical sheet scanning module; S3: The excitation signal is multispectral encoded using a mosaic filter array (4), and the encoded image is synchronously acquired using a lensless sCMOS sensor (3); S4: Perform multi-channel decoding and 3D reconstruction on the encoded image to obtain multi-parameter dynamic data of the sample.

8. The microfluidic-based three-dimensional multiple high-throughput analysis method according to claim 7, characterized in that, Synchronous control of optical tomography excitation and image acquisition includes: After the high-speed galvanometer moves to and stabilizes the target Z layer, sCMOS exposure is triggered. After each layer of exposure is completed, the galvanometer steps to the next layer, and the cycle continues until all tomographic acquisitions are completed. The galvanometer settling time strictly overlaps with the exposure window, resulting in motion blur of <1μm.

9. The microfluidic-based three-dimensional multiple high-throughput analysis method according to claim 7, characterized in that, The multi-channel decoding in step S4 includes: According to the periodic arrangement rules of the mosaic filter array (4), the original image pixels are assigned to the corresponding fluorescence / white light channels; Spatial interpolation reconstruction is performed on sparse sampling points in each channel to form a continuous two-dimensional image; Multi-channel images from all fault planes are fused to generate three-dimensional volume data.

10. A microfluidic chip, characterized in that, Applied to the system according to any one of claims 1 to 6, comprising: The substrate is PDMS, glass or silicon, and it is prepared by micro-nano processing, soft lithography or 3D printing. The microchannels have a width of 50~200μm and a height of 50~200μm, supporting single-layer or multi-layer flow structures; The reaction zone integrates a thin-film heater and a temperature sensor to achieve closed-loop temperature control with an accuracy of ±0.1℃.

Images