A scale-adaptive dynamic inversion analysis method for cellular mechanical-electrical impedance properties
By designing symmetrical deformation channels and multi-position electrode detection in a microfluidic chip, and combining simulation and experimental data matching, the mechanical and electrical properties of cells are dynamically inverted, solving the problem of insufficient accuracy in cell typing in existing technologies and achieving high-precision cell analysis.
Patent Information
- Application Number
- CN202610482338.8
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2026-04-13
- Publication Date
- 2026-11-17
- Estimated Expiration
- 2046-04-13
Smart Images

Figure CN122471906B_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of biological cell characteristic detection and analysis technology, and more specifically, to a method for joint dynamic inversion and classification of cell mechanical and electrical characteristics based on microfluidic technology and multiphysics simulation. Background Technology
[0002] The mechanical properties (such as elastic modulus and viscosity) and electrical properties (such as membrane capacitance and cytoplasmic conductivity) of cells are key biophysical markers, closely related to cell type, state, and physiological and pathological functions. Accurately obtaining these multidimensional physical properties of single cells is of great value for disease diagnosis (such as identification of circulating tumor cells), drug screening, and quality control of cell therapy.
[0003] Currently, mainstream single-cell physical property detection technologies are developing along two independent directions. The first is microfluidic-based mechanical property detection: by designing microchannels with specific geometries (such as constriction channels), cells are deformed under fluid-driven conditions. High-speed microscopic imaging captures the deformation process, and mechanical parameters such as the cell's elastic modulus are inverted based on cell mechanical models (such as elastic solid models or viscoelastic models). This method can effectively reflect the mechanical stiffness of cells, but existing schemes typically only provide a single, static force process, leading to inaccuracies in the estimation of cell mechanical parameters due to noise and load uncertainties. The inversion results usually depend on the inversion calculation of a single deformation process, and the results are easily affected by accidental factors such as initial cell orientation and instantaneous flow field fluctuations in the channel, resulting in a need to improve the stability and repeatability of the inversion parameters. The second is microimpedance sensing-based electrical property detection: microelectrodes are integrated within microchannels, and the changes in impedance (or dielectric properties) caused by the cell flowing through the electrode region are measured to characterize the cell's size, membrane integrity, and internal electrical properties. This method offers a fast response and is easy to integrate, but it primarily reflects the electrical properties of cells and is difficult to directly reveal their mechanical characteristics. Furthermore, the signal is easily affected by the precise location of the cell within the electrode region.
[0004] In recent years, some studies have attempted to combine deformation observation and electrical detection on microfluidic platforms. However, most existing methods only acquire and analyze the two signals in parallel, failing to establish a quantitative correlation model between mechanical property parameters, dynamic deformation processes, and multi-location impedance signals. This leaves the analysis in an "information silo" state, unable to utilize the inherent coupling relationships between multi-dimensional information for mutual verification and supplementation, thus limiting the accuracy and reliability of cell typing. Furthermore, traditional methods typically design fixed-size detection areas for cells of specific sizes, resulting in insufficient adaptability to cell populations of different sizes; and single-path measurements cannot identify and utilize the type-discriminating memory effects or viscoelastic response differences that cells may exhibit during loading-unloading.
[0005] Therefore, how to develop a high-precision dynamic inversion and classification method that can synchronously and correlatedly acquire cellular mechanical and electrical properties, realize dynamic mechanical property inversion, and adapt to different cellular scales has become a technical problem that urgently needs to be solved in this field. Summary of the Invention
[0006] To address the shortcomings of the aforementioned traditional techniques, this application proposes a scale-adaptive dynamic inversion analysis method for cellular mechanical-electrical impedance properties.
[0007] This application provides an intelligent analysis method for cell mechanical properties driven by cell cohort flow mechanism, including:
[0008] S100. Simulated flow characteristic data and simulated electrical impedance signals of each single cell sample are obtained through simulation. Based on the simulated flow characteristic data and simulated electrical impedance signals, a simulation database containing the correspondence between simulated mechanical property parameters, simulated cell deformation data and simulated electrical impedance signals is established.
[0009] S200: Acquire experimental flow characteristic data and experimental impedance signals of the cell sample to be tested;
[0010] S300. Establish a theoretical model of cell mechanical properties. Based on the experimental flow characteristic data of the cell sample to be tested, obtain the theoretical deformation data of the symmetrical contraction and expansion segments of the cell in the microchannel. Based on the different theoretical deformation data of the cell in the contraction and expansion segments, dynamically deduce at least one set of theoretical cell mechanical property parameters. Based on the difference in the response of the contraction and expansion segments to the theoretical cell mechanical property parameters, classify the cell sample to be tested into cells and obtain the theoretical classification results.
[0011] S400, Match the experimental flow characteristic data of the cell sample to be tested with the simulated flow characteristic data in the simulation database, and match the experimental electrical impedance signal of the cell sample to be tested with the simulated electrical impedance signal in the simulation database. Based on the matching results, perform mechanical property inversion and classification on the cell sample to be tested to obtain the simulation matching classification results.
[0012] S500. Compare the theoretical typing results with the simulation matching typing results, and verify the accuracy of the cell typing of the cell sample to be tested based on the comparison results.
[0013] Furthermore, step S200 includes:
[0014] The cell sample to be tested is injected into the microfluidic chip and the cell flow is driven, while the image sequence of the cell flow and the electrical impedance signal are acquired simultaneously.
[0015] Image analysis was performed on the image sequence to obtain experimental flow characteristic data of the cells.
[0016] Furthermore, the microchannel has a hyperbolic symmetrical structure, comprising, from left to right, a first straight tube section, a constriction section, a slit section, an expansion section, and a second straight tube section; the diameter gradually decreases from the first straight tube section to the constriction section, the diameter continues to decrease from the constriction section to the slit section, the diameter gradually increases from the slit section to the expansion section, and the diameter continues to increase from the expansion section to the second straight tube section; the center point of the slit section is the point where the diameter of the entire microchannel is the smallest.
[0017] Furthermore, a differential electrode pair is installed in each of the first straight tube section, the contraction section, the slit section, the expansion section, and the second straight tube section. A constant current is passed through the differential electrode pairs, and the voltage changes generated by the cell flowing through different regions of each section are monitored, and the voltage is converted into resistance to obtain the impedance signal.
[0018] Further, step S100 includes:
[0019] Set different combinations of simulated cell mechanical property parameters;
[0020] Simulated cell deformation data corresponding to different single cell types were obtained through cell fluid-structure interaction model simulation.
[0021] The simulated cell deformation data is input into the cell electrical signal simulation model, and the simulated electrical impedance signal corresponding to different single cell samples is obtained through the simulation of the cell electrical signal simulation model.
[0022] Establish and store the corresponding data relationships among the simulated mechanical property parameters, simulated cell deformation data, and simulated electrical impedance signals.
[0023] Further, step S300 includes:
[0024] S301. Based on the experimental flow characteristic data and first-order constitutive equation of the contraction section, calculate the first theoretical mechanical property parameters of the cell when it is in the contraction section.
[0025] S302. Based on the experimental flow characteristic data and first-order constitutive equation of the expansion section, calculate the second theoretical mechanical property parameters of the cell when it is in the expansion section.
[0026] S303. Based on the first theoretical mechanical property parameter and / or the second theoretical mechanical property parameter, perform cell typing to obtain the theoretical typing result.
[0027] Furthermore, step S301 includes:
[0028] Cellular strain was constructed based on experimental flow characteristic data of the contraction segment;
[0029] Based on the geometric functions and flow parameters of the microchannel, the equivalent shear stress experienced by the cell is constructed;
[0030] Construct the first-order viscoelastic constitutive equation;
[0031] Based on the cell strain recovery curve under the corresponding first-order viscoelastic constitutive equation, the cell relaxation time is obtained by fitting; the cell strain, cell relaxation time and the equivalent shear stress on the cell are substituted into the first-order viscoelastic constitutive equation to calculate the cell viscosity by inversion.
[0032] The cell elastic modulus is calculated based on the cell viscosity and the cell relaxation time to obtain the first theoretical mechanical property parameter that includes cell viscosity and cell elastic modulus.
[0033] Furthermore, the contraction segment is the part where the microchannel width gradually changes from 2.5A to 0.85A, and the expansion segment is the part where the microchannel width gradually changes from 0.85A to 2.5A, where A is the equivalent diameter of the cell.
[0034] Furthermore, step S400 includes:
[0035] The experimental flow characteristic data of the cell sample to be tested are matched with the simulated cell deformation data corresponding to different simulated mechanical property parameters in the simulation database, and the first matching result between the two is obtained.
[0036] The experimental electrical impedance signal of the cell sample to be tested is matched with the simulated electrical impedance signal corresponding to different simulated mechanical property parameters in the simulation database to obtain the second matching result of the successful match between the two.
[0037] When the simulated mechanical property parameters pointed to by the first matching result are consistent with those pointed to by the second matching result, the simulated mechanical property parameters of the cell sample to be tested are determined and the typing is completed, thus obtaining the simulated matching typing result.
[0038] Furthermore, the experimental flow characteristic data includes individual cell geometric features;
[0039] The individual geometric features of the cell include at least one of the following: cell minor axis length, cell major axis length, cell equivalent area, cell equivalent perimeter, cell centroid coordinates, cell roundness, and cell Taylor coefficient.
[0040] The beneficial effects of this invention are as follows: This application relates to a scale-adaptive dynamic inversion analysis method for cell mechanical-electrical impedance properties. On the one hand, addressing the problem that traditional schemes typically only provide a single, static force process, leading to inaccuracies in the estimation of cell mechanical parameters due to noise and load uncertainty, or insufficient precision to distinguish small differences in static cell mechanical properties, this invention utilizes a symmetrical deformation channel to perform multiple calculations on the contraction and expansion segments of the same cell at symmetrical positions, dynamically inverting cell mechanical property parameters to improve the accuracy and robustness of mechanical property parameter inversion. On the other hand, by introducing cell electrical property detection, this invention uses multi-position electrode detection and synchronous imaging pairing to obtain corresponding electrical characteristics under different deformation states, establishing the relationship between mechanical properties, electrical properties, and deformation, thereby achieving joint characterization and classification of cell mechanical and electrical properties. Attached Figure Description
[0041] Figure 1 A flowchart illustrating a scale-adaptive dynamic inversion analysis method for cellular mechanical-electrical impedance properties provided in an embodiment of this application.
[0042] Figure 2 This is a structural block diagram of a microfluidic experimental system provided in an embodiment of this application.
[0043] Figure 3 This is a schematic diagram of the structure of a microfluidic experimental system provided in an embodiment of this application.
[0044] Figure 4 This is a schematic diagram of a microfluidic chip in a microfluidic experimental system provided in an embodiment of this application.
[0045] Figure 5 A sequence of continuous images of the cell sample to be tested, acquired using the scale-adaptive dynamic inversion analysis method for cell mechanical-electrical impedance properties provided in an embodiment of this application.
[0046] Figure 6 To Figure 5 Several key image frames are obtained after extracting key image frames from a continuous image sequence.
[0047] Figure 7 These are several image frames generated during the simulation process after the differential electrode pair is placed.
[0048] Figure 8 The diagram shows the correspondence between the first straight section, the contraction section, the slit section, the expansion section, the second straight section, and each part of the electrical impedance curve.
[0049] Figure 9 The impedance curve generated in the scale-adaptive dynamic inversion analysis method for cellular mechanical-electrical impedance characteristics provided in an embodiment of this application.
[0050] Figure 10 This is a comparison diagram of the contraction segment and expansion segment of the theoretical classification results of an embodiment of this application.
[0051] Figure 11 It is a schematic diagram of the contraction segment where the width of the microchannel gradually changes from 2.5A to 0.85A, and the expansion segment where the width of the microchannel gradually changes from 0.85A to 2.5A.
[0052] Reference numerals: 10-Pressure driving device; 20-Flow rate sensor; 30-Microfluidic chip; 310-Inlet region; 320-Filtration region; 330-Cell detection region; 331-Differential electrode pair; 332-Electrode plate; 333-First straight tube section; 334-Contraction section; 335-Slit section; 336-Expansion section; 337-Second straight tube section; 340-Outlet region; 40-Cell impedance meter; 50-High-speed imaging device; 510-CCD high-speed camera; 520-Microscope; 60-Data processing device; 70-Recovery device; 80-Cell. Detailed Implementation
[0053] To make the objectives, technical solutions, and advantages of this application clearer, the following detailed description is provided in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the scope of this application.
[0054] On the one hand, this application provides a scale-adaptive dynamic inversion analysis method for cellular mechanical-electrical impedance properties.
[0055] like Figure 1 As shown in one embodiment of this application, the scale-adaptive dynamic inversion analysis method for cellular mechanical-electrical impedance properties includes:
[0056] S100 acquires simulated flow characteristic data and simulated electrical impedance signals for each single cell sample through simulation, and establishes a simulation database based on the simulated flow characteristic data and simulated electrical impedance signals, which includes the correspondence between simulated mechanical property parameters, simulated cell deformation data and simulated electrical impedance signals.
[0057] S200 acquires experimental flow characteristic data and experimental impedance signals of the cell sample to be tested.
[0058] S300: Establish a theoretical model of cell mechanical properties. Based on the experimental flow characteristic data of the cell sample to be tested, obtain the theoretical deformation data of the symmetrical contraction and expansion segments of the cell in the microchannel. Based on the different theoretical deformation data of the cell in the contraction and expansion segments, dynamically deduce at least one set of theoretical cell mechanical property parameters. Based on the difference in the response of the contraction and expansion segments to the theoretical cell mechanical property parameters, classify the cell sample to be tested and obtain the theoretical classification results.
[0059] S400 matches the experimental flow characteristic data of the cell sample to be tested with the simulated flow characteristic data in the simulation database, and matches the experimental electrical impedance signal of the cell sample to be tested with the simulated electrical impedance signal in the simulation database. Based on the matching results, the mechanical properties of the cell sample to be tested are inverted and classified to obtain the simulation matching classification results.
[0060] The S500 compares the theoretical typing results with the simulation matching typing results, and verifies the accuracy of the cell typing of the cell sample to be tested based on the comparison results.
[0061] Specifically, the experimental flow characteristic data in S200 is acquired in the microfluidic experimental system, and after acquisition, the experimental flow characteristic data is stored in a pre-created cell flow experimental database.
[0062] Optionally, prior to S100, scale-adaptive dynamic inversion analysis methods for cellular mechanical-electrical impedance properties also include:
[0063] S101, Construct a microfluidic experimental system.
[0064] S102 sets the boundary conditions for single-cell sample delivery.
[0065] S103 sets the initial conditions for single-cell sample delivery.
[0066] Specifically, in S101, the microfluidic experimental system (such as...) Figure 2 , Figure 3 (As shown) It includes a pressure drive device 10, a flow rate sensor 20, a microfluidic chip 30, a cell impedance meter 40, a high-speed imaging device 50, a data processing device 60, and a recovery device 70.
[0067] In S102, the boundaries of the microchannel walls and cell membrane surfaces are all subject to no-slip boundary conditions.
[0068] The no-slip boundary condition refers to the situation where, at the fluid-solid interface, fluid particles adhering to the solid surface have zero velocity relative to the solid surface. Therefore, under the combined effect of local stress equilibrium and the no-slip condition, the fluid velocity is equal to the velocity of the particles moving on the cell membrane.
[0069] The initial conditions in S103, such as Figure 4 As shown, the following can be implemented: a cell release point is set at a preset distance of 310 from the entrance area, and cells are released sequentially according to the migration distance trigger mechanism, ensuring that the initial distance between adjacent cells is not less than 1.2 times the average cell radius. Subsequently, a stable cell flow can be formed by driving the flow through pressure difference. This initial condition setting can effectively avoid non-physical cell collisions and ensure the reasonable development of the lubrication layer.
[0070] A stable pressure difference is established between the pressure-driven device 10 (which may be an injection pump in this application) connected to the inlet region 310 and the outlet region 340, driving the cell suspension to generate background flow, thereby forming a stable cell flow. This pressure difference is provided by the pressure-driven device 10. For example, if the pressure-driven device 10 is an injection pump, the injection pump pushes the piston to apply a positive pressure (higher than atmospheric pressure) to the fluid at the inlet end, while the outlet end is usually connected to the atmosphere (at atmospheric pressure). This establishes a pressure difference between the inlet and outlet ends of the microchannel. The microchannel is the cell detection region 330. This pressure difference is the driving force of the background flow, which determines the average flow velocity of the fluid in the channel. This average flow velocity in the channel is the average background flow velocity V at the inlet region 310 of the microchannel, and it is this background flow that carries the cells forward.
[0071] Optionally, prior to S100, the following are also included:
[0072] S104, Construct a single-cell sample library for storing cell samples.
[0073] Specifically, S104 includes:
[0074] S104a: Take a cell suspension containing only a single type of cell, fix the cell concentration, keep the total volume constant, and add the cell suspension as a single-type cell sample to the single-type cell sample library.
[0075] S104b: Change the cell type. Perform S104a once for each cell type to obtain single cell samples of different cell types and store them in the single cell sample library.
[0076] Optionally, a single-cell sample may contain only one cell.
[0077] The S100 establishes a multiphysics simulation model of cell fluid-structure interaction and electrical signal simulation, outputting flow characteristic data and impedance signal data consistent with experiments. This model is used to construct a quantitative relationship between electrical properties, mechanical properties, and cell deformation, and is compared and verified with the inversion results and classification results of two loading paths, thereby improving the reliability and applicability of the model and system.
[0078] The S400 uses both mechanical inversion parameters and impedance characteristics for classification. Compared to single deformation or single impedance indicators, it is better able to distinguish heterogeneity within cell populations (such as differences in membrane tension / shear modulus, differences in membrane capacitance / conductivity), thereby improving classification accuracy and anti-interference ability.
[0079] In this embodiment, a closed-loop verification system is constructed by introducing a dual-load path theory inversion based on symmetrical contraction and expansion segments, combined with dual matching and cross-validation of simulation and experimental data. This system enables collaborative analysis of mechanical and electrical properties and mutual verification of results. This method effectively overcomes the random errors of single deformation inversion results and solves the information silo problem caused by isolated analysis of mechanical and electrical information. Furthermore, by comparing theoretical and simulation models, the overall accuracy, reliability, and robustness of cell mechanical property inversion and classification are significantly improved. Moreover, this embodiment emphasizes multidimensional features based on deformation and impedance, as well as model inversion, which does not rely on fixed size thresholds and has better compatibility and scalability for cells of different sizes and types. Dual-path measurement and symmetrical position comparison of the same cell naturally provide an internal reference, which can partially offset systematic errors caused by external factors such as flow fluctuations, imaging noise, and electrode drift, improving long-term operational stability and making it less sensitive to environmental disturbances and system drift.
[0080] In one embodiment of this application, S200 includes acquiring experimental flow characteristic data and experimental impedance signals of the cell sample to be tested, including:
[0081] S210 injects the cell sample to be tested into the microfluidic chip and drives the cell flow, while simultaneously acquiring image sequences and electrical impedance signals of the cell flow;
[0082] S220 performs image analysis on the image sequence to obtain experimental flow characteristic data of cells.
[0083] Specifically, the cell sample to be tested is injected into the microfluidic chip 30 via the pressure-driven device 10, causing the cells to flow within the microfluidic chip 30. Image sequences of the cell flow are acquired using a high-speed imaging device 50. Flow parameters of the cell flow are acquired using a flow rate sensor 20.
[0084] Optionally, the experimental flow characteristic data includes individual cell geometric features.
[0085] S220 includes: S221, constructing a Cartesian coordinate system and coupling the image sequence of cell flow with the Cartesian coordinate system.
[0086] Specifically, this allows us to determine several characteristics of an individual cell, such as the centroid coordinates of that cell. The origin of the Cartesian coordinate system can be set arbitrarily.
[0087] The Cartesian coordinate system is defined with the geometric center O of the microchannel as its origin. The microchannel can be a narrow channel or a channel with a non-uniform cross-section. The Cartesian coordinate system is defined with the background flow direction as the positive x-axis, the increasing depth direction of the narrow or non-uniform cross-section channel as the positive y-axis, and the increasing width direction of the narrow or non-uniform cross-section channel as the positive z-axis. The microchannel is the cell detection zone 330.
[0088] Figure 5 A sequence of continuous images of the cell sample to be tested, acquired using the scale-adaptive dynamic inversion analysis method for cell mechanical-electrical impedance properties provided in an embodiment of this application. Figure 6 To Figure 5 The S220 extracts key image frames from a continuous image sequence, resulting in several key image frames. The S220 can extract key frames from a portion of a continuous image sequence, or it can process all key frames one by one.
[0089] In one embodiment of this application, the individual geometric features of a cell include at least one of the following: cell minor axis length, cell major axis length, cell outline shape, cell equivalent area, cell equivalent perimeter, cell centroid coordinates, the ratio between cell minor axis length and cell major axis length, cell axial length, cell roundness, and cell Taylor coefficient.
[0090] Specifically, the core method for obtaining the geometric features of individual cells is to perform edge detection and extraction of cell contours on the image. This is a common existing technology, and this application will not explain or elaborate on it in detail. Any technology for recognizing and extracting cell contour shapes, as long as it can achieve the extraction of cell contour shapes, can be used in this application.
[0091] In this embodiment, by explicitly defining the specific individual cell geometric features (such as major and minor axes, area, centroid, etc.) included in the experimental flow characteristic data, the key data dimensions used for inversion and matching are defined. These geometric features are direct inputs for calculating strain and deformation parameters and for simulation. Their explicit definition ensures data consistency and result repeatability in the implementation of the method, providing a standardized feature input specification for the entire analysis process.
[0092] like Figure 8 As shown, in one embodiment of this application, the microchannel has a hyperbolic symmetrical structure and includes, from left to right, a first straight tube section 333, a constriction section 334, a slit section 335, an expansion section 336, and a second straight tube section 337. The diameter gradually decreases from the first straight tube section 333 to the constriction section 334, continues to decrease from the constriction section 334 to the slit section 335, gradually increases from the slit section 335 to the expansion section 336, and continues to increase from the expansion section 336 to the second straight tube section 337. The center point of the slit section 335 is the point where the diameter of the entire microchannel is at its minimum.
[0093] In this embodiment, a specific microchannel with a hyperbolic symmetric structure is employed, comprising a first straight tube section 333, a constriction section 334, a slit section 335, an expansion section 336, and a second straight tube section 337, arranged from left to right, providing a geometrically symmetrical loading and unloading environment for the cell. This structural design ensures that the same cell undergoes mechanically comparable but opposite deformations when flowing through symmetrical positions, providing a stable and repeatable physical basis for dynamic inversion and analysis based on differences in symmetrical segments.
[0094] like Figure 7 As shown in one embodiment of this application, a differential electrode pair 331 is installed in each of the first straight tube section 333, the contraction section 334, the slit section 335, the expansion section 336, and the second straight tube section 337. A current of constant magnitude is passed through the differential electrode pair 331, and the voltage change generated by the cell 80 flowing through different regions of each section is monitored, and the voltage is converted into resistance to obtain the impedance signal.
[0095] Specifically, cells 80 are only labeled in this embodiment; for brevity, they are not labeled in other embodiments. The differential electrode pair 331 is used to acquire impedance signals. Therefore, we actually set up the differential electrode pair 331 not only in S200 to acquire the experimental flow characteristic data and experimental impedance signals of the cell sample under test, but also in the simulation process of S100 to simulate the impedance signals. Each differential electrode pair 331 includes three electrode plates 332.
[0096] Each electrode plate 332 is attached to the substrate of the microfluidic chip 30. The leftmost electrode plate serves as the positive electrode, the middle electrode plate as ground, and the rightmost electrode plate as the positive electrode. The voltage between the leftmost and middle electrode plates is measured as U1, and the voltage between the rightmost and middle electrode plates is measured as U2. Under normal conditions, U1 equals U2. When cells flow into the area between the leftmost and middle electrode plates, U1 decreases, and the difference between U1 and U2 becomes negative; when cells flow into the area between the rightmost and middle electrode plates, U2 decreases, and the difference between U1 and U2 becomes positive. We can detect and calculate this difference between U1 and U2 using a cell impedance meter 40. We calculate the absolute value of the difference between U1 and U2, then divide this absolute value by a constant current, and we get a resistance value, which is the impedance. Figure 9 The ordinate.
[0097] Each cell takes approximately 0.06 seconds to flow from the leftmost first straight tube segment to the rightmost second straight tube segment. We sample the difference between U1 and U2 at each sampling time point to obtain a corresponding resistance value, which is used as the impedance value. This allows us to obtain a time-impedance curve, as shown below. Figure 8 As shown, this curve will show 5 regions, corresponding to the first straight tube section 333, the contraction section 334, the slit section 335, the expansion section 336, and the second straight tube section 337 respectively. Each region has a distinct "peak", which corresponds to the sampling time node when the impedance value of each section is the largest.
[0098] Both S100 and S300 generate this curve. The data source for this curve in the S300 simulation process comes from the simulation results of cell mechanical properties. Because the peaks in the curve are significantly affected by the dielectric constant, which can be understood as a constant in the experimental environment, other influencing factors are the cell mechanical property parameters. Since the cell mechanical property parameters are fixed, the cell's geometry and deformation trend are determined, and the curve is roughly formed. Unless the experimental environment constant changes drastically, causing a significant change in the dielectric constant, the curve remains relatively constant.
[0099] Therefore, different cell types will result in different mechanical properties, and different mechanical effects will lead to different electrical impedance curves.
[0100] The detection unit with multiple locations and multiple impedances can obtain the impedance change curves and spatial characteristics of cells moving along the channel, enhance the ability to analyze signal changes caused by cell membrane electrical properties and size / deformation, and improve the risk of misjudgment from single-point electrical measurement.
[0101] In this embodiment, by setting differential electrode pairs 331 in all five functional sections of the microchannel, multi-point, distributed impedance monitoring of the entire cell flow process is achieved. This design can capture the electrical response of the cell at different deformation stages (such as compression and recovery), thereby generating an impedance change curve containing spatial location information, providing high spatiotemporal resolution electrical data for correlating electrical signals with specific mechanical deformation processes of the cell.
[0102] This embodiment simultaneously acquires morphological / dynamic features (contour, deformation, size, velocity, etc.) of cell flow deformation and multi-point electrode impedance signals at the single-cell scale, realizing the synchronous detection and correlation analysis of mechanical and electrical properties, improving the identification dimension and information content of cell state differences, achieving synchronous coupling of mechanical and electrical properties, and demonstrating strong multidimensional characterization capabilities.
[0103] In one embodiment of this application, S100 includes obtaining simulated flow characteristic data and simulated electrical impedance signals for each single cell sample through simulation, and establishing a simulation database based on the simulated flow characteristic data and simulated electrical impedance signals, which includes the correspondence between simulated mechanical property parameters, simulated cell deformation data, and simulated electrical impedance signals, including:
[0104] S110 allows for setting different combinations of simulated cell mechanical property parameters.
[0105] S120 uses a cell fluid-structure interaction model to simulate and obtain simulated cell deformation data for different single cell types.
[0106] S130: Input the simulated cell deformation data into the cell electrical signal simulation model, and obtain the simulated electrical impedance signal corresponding to different single cell samples through the cell electrical signal simulation model.
[0107] S140, establish and store the corresponding data relationships between the simulated mechanical property parameters, simulated cell deformation data, and simulated electrical impedance signals.
[0108] Specifically, in S120, the cell-structure interaction model uses a microcapsule model to represent the cell as a whole. The Skalak constitutive model can be used as the constitutive equation for the cell-structure interaction model, which describes the in-membrane strain energy. The cell-structure interaction model uses a viscous fluid to represent the cell contents. During numerical simulation in S120, the Helfrich model is used to describe the membrane bending energy in the cell-structure interaction model.
[0109] S120 may include:
[0110] S121, Establish a cell-fluid-structure interaction model. The cell-fluid-structure interaction model is used to simulate the flow of cells in microchannels.
[0111] Optionally, S121 includes:
[0112] S121a uses a microcapsule model to physically model the initial morphology of a single cell.
[0113] S121b models cell flow processes.
[0114] S121c, adjust the cell size within the microchannel and set the simulation mechanical property parameters of the cells in the cell fluid-structure interaction model.
[0115] Specifically, in S121a, a microcapsule model is used to characterize cell dynamics.
[0116] S121b constructs a system containing N units with an average radius of... A cell suspension system model. N is a positive integer, and N is greater than 1.
[0117] When setting the simulated mechanical property parameters of cells in the cell fluid-structure interaction model in S121c, the various simulated mechanical property parameters of the cell membrane are determined according to the following principles:
[0118] (a) Intra-membrane deformation behavior: The Skalak model was used to characterize the intra-membrane deformation behavior. The intra-membrane deformation behavior was mainly quantified by two parameters: the strain energy density within the cell membrane and the area expansion modulus.
[0119] The quantitative formula for strain energy density within the cell membrane is shown in formula (1).
[0120] (1).
[0121] in, This represents the strain energy density within the cell membrane. This is the in-film shear modulus. As the first strain invariant, This is the second strain invariant. It is the compressibility coefficient, specifically the Skalak coefficient which controls the degree of incompressibility of the area.
[0122] The quantitative formula for strain energy density within the cell membrane is shown in formula (2).
[0123] The area expansion modulus is determined by it as follows:
[0124] (2).
[0125] Among them, It is the modulus of area expansion. The compressibility coefficient. This is the in-film shear modulus.
[0126] (b) Membrane bending behavior: The Helfrich model is used to describe the membrane bending energy.
[0127] The formula for characterizing membrane bending energy is shown in formula (3).
[0128] (3).
[0129] in, This refers to the membrane bending energy. This represents the energy stored in the entire membrane due to bending. For membrane flexural modulus, The larger the membrane, the stiffer it is, and the less flexible it is. The mean curvature of the membrane describes how drastically the membrane bends at a specific point in time. The spontaneous curvature of the membrane can be understood as the degree of bending that the membrane "prefers" when it is not subjected to any external force and without considering the effects of other molecules. It is the integral symbol. This is the total reference area of the entire membrane. Formula (3) means that the total torsional energy of the entire cell membrane... This is equivalent to summing up the bending energies of every tiny region on the membrane and calculating it through integration. The Helfrich model can accurately describe the bending, folding, and local bulging behaviors of cells during flow.
[0130] (c) Cell membrane surface viscosity to the ratio of intracellular to extracellular viscosity
[0131] Cell membrane surface viscosity is expressed as .
[0132] Both intracellular and extracellular fluids are considered viscous fluids, and the intracellular and extracellular viscosity ratio is calculated according to formula (4).
[0133] (4).
[0134] in, The ratio of intracellular to extracellular viscosity. The viscosity of the fluid inside the cell. This refers to the viscosity of the fluid outside the cell.
[0135] In S101c, cell flow behavior is regulated through the synergistic effect of the following dimensionless parameters:
[0136] The dimensionless parameter I is the number of capillaries. , The formula for calculating the ratio of pressure-driven fluid inertial force to cell membrane elastic force is shown in formula (5).
[0137] (5).
[0138] in, This represents the number of capillaries. This refers to the viscosity of the fluid outside the cell. This is the in-film shear modulus. The background fluid flow velocity is at the inlet end of the microchannel.
[0139] The dimensionless parameter II is the initial spacing d between cells. d is the average distance between the centroids of two adjacent cells under the initial conditions, and it is used to characterize the spatial distribution of cells.
[0140] Dimensionless parameter III is the scale ratio between cell size and channel size. , The calculation formula is shown in formula (6), which is used to quantify the channel geometric constraint effect.
[0141] (6).
[0142] in, This represents the scale ratio between cell size and channel size. denoted as the average radius of the cell. This is the width at the narrowest point of the microchannel.
[0143] Following S121 are:
[0144] S122, Construct a microchannel model.
[0145] S123 couples the cell fluid-structure interaction model with the Cartesian coordinate system and the microchannel model with the Cartesian coordinate system.
[0146] S124, construct the driving force conditions for cell suspension to pass through the channel.
[0147] Specifically, in S122, the microchannel model constructs a narrow channel or a non-uniform cross-section channel with a height of h. The width at the narrowest point of the narrow channel or the non-uniform cross-section channel is... .
[0148] In S123, the Cartesian coordinate system has its origin at point O, the geometric center of the narrow or non-uniform cross-section channel. The background flow direction is the positive x-axis, the direction of increasing depth of the narrow or non-uniform cross-section channel is the positive y-axis, and the direction of increasing width of the narrow or non-uniform cross-section channel is the positive z-axis. The background fluid flow velocity at the microchannel inlet is denoted as... .
[0149] In S124, pressure difference is used as the driving force to drive the cell suspension through narrow channels or channels with non-uniform cross sections.
[0150] In one embodiment of this application, after S124, the following is also included:
[0151] S125, obtain the simulated mechanical property parameters.
[0152] S126, The above mechanical property parameters are used as input parameters for the cell fluid-structure interaction model for numerical simulation.
[0153] Specifically, the simulation requires four types of parameters: inlet velocity, microchannel structure, cell size, and mechanical properties. The inlet velocity, microchannel structure, and cell size can all be customized and directly set within the fluid-structure interaction model. However, the mechanical properties are more complex.
[0154] Mechanical property parameters include one or more of the following: in-membrane shear modulus, area expansion modulus, compressibility coefficient, membrane flexural modulus, spontaneous membrane curvature, and cell membrane surface viscosity. Mechanical property parameters also include dimensionless parameters describing the cell's deformability and dynamic response in the flow field. These dimensionless parameters specifically include one or more of the following: capillary number, initial cell spacing, scale ratio between cell size and channel size, and relative surface viscosity. These mechanical property parameters collectively determine the cell's deformation behavior, arrangement, and dynamic interactions under fluid excitation in the microchannel, and are the core physical quantities for achieving cell mechanical property inversion and classification in the simulation process in this application. These simulated mechanical property parameters can be obtained in two ways. The first way is by consulting relevant literature. The second way is through experiments, for example, by sequentially placing 1000 single-type cell samples into the microchannel for multiple sets of microfluidic experiments, obtaining experimental flow characteristic data for the 1000 single-type cell samples, and then obtaining the cell's mechanical property parameters based on the experimental flow characteristic data. These mechanical property parameters are then used as the simulated mechanical property parameters.
[0155] In one embodiment of this application, the numerical simulation in S126 adopts a boundary integral-finite element coupled algorithm. The flow field is solved by the boundary integral method, the cell membrane deformation is solved by the finite element method, and the transient coupling equilibrium between the fluid domain and the solid domain is achieved by iterative solution in each simulation time step.
[0156] Specifically, the numerical simulation is a transient simulation, which involves sequentially solving the flow field and cell deformation field at multiple discrete time points, wherein at each time point, the flow field and cell deformation field are coupled into equilibrium through iterative calculation.
[0157] Optionally, the boundary integral-finite element coupling algorithm is used to simulate and reproduce the cellular flow of the sample within the cellular fluid-structure interaction model. This simulation procedure can be used for both isolated flow samples and heterogeneous flow samples. S126 specifically includes:
[0158] S126a, connect to the supercomputer and begin task initialization.
[0159] S126b sets the mechanical parameters for the three-dimensional simulation.
[0160] S126c, based on the system state at the current time step, performs a fluid-structure interaction (FSI) solution and obtains the updated cell geometry features based on the FSI solution.
[0161] S126d determines whether the solution value of fluid-structure interaction meets the convergence condition.
[0162] S126e, if the fluid-structure interaction solution value does not meet the convergence condition, then based on the updated cell geometry, return to the system state based on the current time step, execute a fluid-structure interaction solution, and obtain the updated cell membrane geometric features according to the fluid-structure interaction solution loop until the fluid-structure interaction solution value meets the convergence condition.
[0163] If the fluid-structure interaction solution value satisfies the convergence condition in S126f, then the next fluid-structure interaction solution is executed, that is, the next loop from S226c to S226d is executed.
[0164] S126g obtains the three-dimensional coordinates of each vertex of each cell at each time step based on the simulation results of the evolution of cell geometry over time, and stores the three-dimensional coordinates of each vertex of each cell at each time step as the original simulation data in the database.
[0165] Specifically, in this embodiment, the entire cell is physically abstracted into a geometric spherical structure surrounded by multiple triangular units composed of multiple vertices. By calculating the position coordinates of each vertex, the deformation of the cell at each moment can be calculated. If the fluid-structure interaction solution value in S126f meets the convergence condition, it means that the cell geometry at that moment meets the recording requirements and can be recorded. Then, the study of the next change begins, i.e., the next fluid-structure interaction solution process is executed until the state of each cell at each moment is obtained. The three-dimensional coordinates of each vertex are shown in Table 1. Table 1 shows the data format of the original cell simulation data, where the three columns of Table 1 are the x, y, and z coordinates of the vertex, respectively.
[0166] Table 1 - Original Cell Simulation Data
[0167]
[0168] All units in Table 1 are meters (m). The center point of the microchannel inlet is set to (0, 0, 0), so the x, y, and z axes in Table 1 can be seen to be positive and negative. Each cell has a Table 1 at every moment.
[0169] Following S126, it also includes:
[0170] S127. Post-processing is performed on the original simulation data in the 3D simulation database to obtain simulated cell deformation data. The cell fluid-structure interaction model can obtain cell flow processes consistent with experimental flow. Simulated cell deformation data of the same type of cells as those in the experiment in S100 can be obtained by post-processing the original simulation data in the 3D simulation database (i.e., the data in Table 1). During the post-processing of the original simulation data, this embodiment uses the cell surface mesh vertex data output by the 3D fluid-structure interaction simulation software Fluent as a basis to quantitatively analyze the geometric deformation and dynamic behavior of cells in the microchannel.
[0171] S127 specifically includes:
[0172] (a) Input data and preprocessing
[0173] During the simulation, the three-dimensional vertex coordinates of the cell surface are exported at preset time step intervals (e.g., every 2000 steps), containing N discrete points, each in meters (m). In this embodiment, all time series files are first read, and their three-dimensional coordinates are converted to micrometers (μm) using formula (7). The three-dimensional coordinates here are the three-dimensional coordinates of the vertices obtained by S326g, i.e. .
[0174] (7).
[0175] in, The three-dimensional coordinates of the discrete point before transformation. These are the three-dimensional coordinates of the discrete points after transformation.
[0176] This step standardizes the coordinate scale, which facilitates the calculation of subsequent geometric quantities.
[0177] (II) Construction of two-dimensional cross-sections of cells and extraction of geometric features
[0178] To describe the deformation behavior of cells in the flow direction, this embodiment projects three-dimensional vertices onto a two-dimensional xy plane and calculates cell length, width, and effective cross-sectional area (without a uniform description) based on this projection. Specifically, the calculation of cell length, width, cross-sectional area, deformation parameters, centroid, and velocity is taken as an example. To describe the deformation behavior of cells in the flow direction, this embodiment projects three-dimensional vertices onto the xy plane and calculates cell length, width, and effective cross-sectional area based on this projection. The three-dimensional vertices here are the vertices obtained in S217, where the three-dimensional coordinates of the vertices are projected onto the xy plane.
[0179] (a) Calculate cell length using formula (8).
[0180] (8).
[0181] in, For cell length, This represents the x-coordinate value of the 3D vertex with the largest x-coordinate. The x-coordinate value is the x-coordinate of the three-dimensional vertex with the smallest x-coordinate. Cell length L reflects the degree of cell stretching under the influence of the flow field.
[0182] (b) Calculate the cell width using formula (9).
[0183] (9).
[0184] in, Cell width, This represents the x-coordinate value of the 3D vertex with the largest y-coordinate. The x-coordinate value is the value of the 3D vertex with the smallest y-coordinate. Cell width is used to characterize the amount of lateral compression of the cell after it is squeezed by the channel contraction.
[0185] (3) Calculation of cell cross-sectional area
[0186] Since the order of the cell surface point cloud does not form a closed contour, this embodiment uses the convex hull algorithm to construct the minimum convex polygon of the cell in the xy plane. This method includes: sorting the projected point set by coordinates; then using the monotonic chain algorithm to construct the upper and lower convex hulls; next, merging them to form the complete convex hull boundary; finally, using the Gaussian polygon area formula to calculate the convex hull area, and using the convex hull area as the cell cross-sectional area.
[0187] (III) Calculation of the location of the cell centroid
[0188] The coordinates of the cell centroid are obtained by averaging the coordinates of all vertices at each time step, as shown in formula (10).
[0189] (10).
[0190] in, Using the cell's centroid coordinates, This represents the average x-value of all vertices at each time step of the cell. Let be the average y-value of all vertices in the cell at each time step. This is the average z-value of all vertices at each time step of the cell.
[0191] (iv) Calculation of cell velocity and kinetic parameters
[0192] The cell velocity component is calculated by the centroid position increment and time interval between adjacent time steps, and the calculation formula is shown in formula (11).
[0193] (11).
[0194] in, For cells on the x-axis The velocity component at time t, The coordinates of the cell's centroid are at The value of x at time t, The coordinates of the cell's centroid are at The x value at time t. For cells on the y-axis The velocity component at time t, The coordinates of the cell's centroid are at The y-value at time t. The coordinates of the cell's centroid are at The y-value at time t. For cells on the z-axis The velocity component at time t, The coordinates of the cell's centroid are at The z-value at time t. The coordinates of the cell's centroid are at The z-value at time t.
[0195] Furthermore, the speed is calculated according to formula (12).
[0196] (12).
[0197] in, Let be the magnitude of the cell's velocity at time t. For cells on the x-axis The velocity component at time t, For cells on the y-axis The velocity component at time t, For cells on the z-axis The velocity component at any given moment.
[0198] (v) Extraction of Taylor Deformation Index
[0199] In this embodiment, the Taylor deformability index is used to describe the degree of cell morphological deviation. The Taylor deformability index is a dimensionless parameter. Its calculation method is shown in formula (13).
[0200] (13).
[0201] in, Here, Taylor's deformability index is given, and cell length is given. Cell width, The larger the value, the more significant the stretching of the cell in the flow direction.
[0202] (vi) Results Output and Visualization
[0203] This embodiment outputs time-series data such as cell length, cell width, cell cross-sectional area, deformation parameters (Taylor deformation index), centroid coordinates, velocity components, and resultant velocity to a CSV file, and automatically generates time-varying curves of cell length, cell width, cell cross-sectional area, velocity, and deformation parameters to illustrate the cell deformation process in a hyperbolic microchannel. The horizontal axis of the curve represents time, and the vertical axis represents the different individual cell geometric characteristics and cell population flow characteristics obtained from different numerical simulation processes.
[0204] S130 is used to simulate impedance signals.
[0205] S130 first creates a cell electrical signal simulation model, inputs simulated cell deformation data into the model, and obtains simulated electrical impedance signals corresponding to different single cell types through simulation. The microchannel structure simulation during the simulation is the same as that used in the experiment in S200. After the simulation, the simulated electrical impedance signals are stored in the simulation database.
[0206] After simulation using the fluid-structure interaction model, the flow field is obtained. Because the shape changes of the cells in the flow field, they replace the medium in the electric field, essentially changing the medium filling the space. This changes the dielectric constant of the medium, naturally altering the subsequent electrical signal curve. This is why we use simulated cell deformation data to simulate the electrical impedance signal in S130. The electrical impedance signal simulation primarily uses cell length. Cell width The four quantities are cell volume, cell cross-sectional area, etc.
[0207] Optionally, in S130, the three-dimensional coordinates of each vertex obtained from the cell fluid-structure interaction model can be directly used as input data for the cell electrical signal simulation model. The cell electrical signal simulation model will automatically calculate the cell length, cell width, cell height, and cell cross-sectional area at each time point based on the three-dimensional coordinates of each vertex, and then output the electrical impedance signal simulation curve.
[0208] As mentioned above in S140, the simulated cell deformation data includes time-series data such as cell length, cell width, cell cross-sectional area, Taylor deformation index, centroid coordinates, velocity components, and resultant velocity. These data can be stored in the simulation database in the form of graphs or data matrices. Since the simulated cell deformation data is obtained through simulation of mechanical effect parameters, the correspondence between the simulated mechanical property parameters and the simulated cell deformation data is also stored in the simulation database.
[0209] Since the simulated electrical impedance signal is derived from simulated cell deformation data, when storing the simulated electrical impedance signal in the simulation database, we also store the correspondence between the simulated cell deformation data and the simulated electrical impedance signal. Specifically, we can also generate and store electrical impedance data curves, with time on the horizontal axis and electrical impedance value on the vertical axis.
[0210] In summary, we can establish and store the corresponding data relationships among the simulated mechanical property parameters, simulated cell deformation data, and simulated electrical impedance signals.
[0211] In this embodiment, a joint simulation model of fluid-structure interaction and electrical signal simulation is established, and a database of correspondences between mechanical property parameters, deformation data, and electrical impedance signals is generated in a parameterized manner, thus constructing a digital cell template library covering multiple physics fields. This database provides a complete reference system for matching and comparing experimental data and serves as a key bridge connecting abstract mechanical parameters with specific experimental observations (deformation, electrical signals).
[0212] In one embodiment of this application, S300 includes establishing a theoretical model of cell mechanical properties, obtaining theoretical deformation data of the symmetrical contraction and expansion segments of the cells within the microchannel based on experimental flow characteristic data of the cell samples to be tested, dynamically retrieving at least one set of theoretical cell mechanical property parameters based on the different theoretical deformation data of the cells within the contraction and expansion segments, and performing cell typing on the cell samples to be tested based on the differences in the responses of the contraction and expansion segments to the theoretical cell mechanical property parameters, to obtain theoretical typing results, including:
[0213] S310, based on experimental flow characteristic data of the contraction zone and the first-order constitutive equation, calculates the first theoretical mechanical property parameters of the cell when it is in the contraction zone.
[0214] S320, based on experimental flow characteristic data of the expansion section and the first-order constitutive equation, calculates the second theoretical mechanical property parameters of the cell when it is in the expansion section.
[0215] S330, cell typing is performed based on the first theoretical mechanical property parameter and / or the second theoretical mechanical property parameter to obtain the theoretical typing result.
[0216] Specifically, S310 and S320 can use the same first-order constitutive equation.
[0217] In this embodiment, by calculating the theoretical mechanical parameters of cells during the contraction (loading) and expansion (unloading) phases respectively, and using their differences to characterize the cellular memory effect, a single static parameter measurement is upgraded to an analysis of dynamic loading path response. This deepens the classification basis from a single mechanical parameter value to the capture of dynamic characteristics such as cellular viscoelastic relaxation behavior, significantly enhancing the ability to distinguish cell subtypes with similar steady-state stiffness but different dynamic responses.
[0218] In one embodiment of this application, S310 includes:
[0219] S311, Constructing cell strain based on experimental flow characteristic data of the contraction segment. .
[0220] S312, based on the geometric functions and flow parameters of microchannels, constructs the equivalent shear stress experienced by cells. .
[0221] S313, construct the first-order viscoelastic constitutive equation.
[0222] S314, based on the cell strain recovery curve under the corresponding first-order viscoelastic constitutive equation, the cell relaxation time is obtained by fitting. .
[0223] S315 will strain the cells Cell relaxation time and the equivalent shear stress on the cell Substituting into the first-order viscoelastic constitutive equation, the cell viscosity is calculated by inversion. .
[0224] S316, based on cell viscosity With cell relaxation time Calculate the elastic modulus of a cell To obtain information including cell viscosity and cell elastic modulus The first theoretical mechanical property parameter.
[0225] Specifically, S311 to S316 in this embodiment are the calculation methods for the first theoretical mechanical property parameters when the cell is in the contraction phase. The calculation methods for the second theoretical mechanical property parameters when the cell is in the expansion phase are the same as those for the first theoretical mechanical property parameters, and therefore are omitted.
[0226] In S311, contour segmentation and geometric parameter extraction are performed on the image sequence to obtain parameters of cell changes over time, such as roundness. Select the unstressed region before the cell enters the contraction phase or a designated reference frame as the reference state to determine the reference roundness. .
[0227] Calculate cell strain according to formula (14).
[0228] (14).
[0229] in, Let be the cell strain, which is a dimensionless sequence, and t be the time variable. For cell roundness, Let f be the reference roundness of the cell. If the image frame rate is denoted as f, then the sampling time interval is 1 / f. Substitute t into the i-th sampling time interval.
[0230] In S312, channel structure functions and flow parameters are established, and the equivalent shear stress sequence of the cell is constructed from the channel structure functions and flow parameters. .
[0231] S312 includes:
[0232] S312a, input the flow parameters and level out the units of each flow parameter.
[0233] Specifically, the flow parameters include: channel height: H(x) = H1, in meters; fluid volumetric flow rate: Q = Q1, in cubic meters per second. 3 / s; Fluid dynamic viscosity: μ=μ1, unit is Pa·s; Hyperbolic channel major axis parameter: a, unit is m; Hyperbolic channel minor axis parameter: b, unit is m.
[0234] S312b, Establish the hyperbola channel width function .
[0235] Suppose that the lateral width of the channel varies with the axial coordinate x according to the hyperbolic equation in formula (15):
[0236] (15).
[0237] in, This refers to the major axis parameter of the hyperbola channel. For the minor axis parameters of the hyperbola channel, y represents the axial coordinates of the cell. This represents the horizontal width of the channel.
[0238] Solve Solving for formula (16) yields the solution.
[0239] (16). Among them, It can be defined as the channel half-width, then the width of the entire channel. That's twice as much .
[0240] S312c, the local average velocity is obtained from geometric functions. Under low Reynolds number laminar flow conditions, the local cross-sectional average velocity is established based on continuity relationships. Since the height of the microchannel is constant at H, the cross-sectional area... that is Multiply by H.
[0241] Therefore, the local average flow velocity is as shown in formula (17).
[0242] (17).
[0243] in, Let Q be the average velocity across the local cross section, and Q be the fluid volumetric flow rate. For the cross-sectional area, H represents the width of the entire channel, and H represents the height of the microchannel.
[0244] S312d approximates the local cross-section of the microchannel as a parallel plate laminar flow and establishes the equivalent local shear rate, as shown in formula (18).
[0245] (18).
[0246] in, For local shear rate, Here, H represents the average flow velocity across the local cross-section, and H represents the height of the microchannel.
[0247] Substituting formula (17) into formula (18) yields only the result of The shear rate expression determined by H and Q is shown in formula (19).
[0248] (19).
[0249] S313e, constructing a spatial domain expression of equivalent shear stress experienced by cells. As shown in formula (20).
[0250] (20).
[0251] Substituting formula (19) into formula (20) yields formula (21).
[0252] (twenty one).
[0253] For S312f, any spatial location x corresponds to a specific time t. By tracking the cell's centroid, the curve x(t) representing the change in the cell's axial position over time is obtained, thus expressing the spatial domain of the equivalent shear stress experienced by the cell. Converted to spatial domain expression of equivalent shear stress on cells .
[0254] In S313, a first-order viscoelastic constitutive equation can be established based on the Maxwell model, as shown in equation (22). Of course, other models can also be used as constitutive relations in this embodiment, such as the Kelvin model.
[0255] (twenty two).
[0256] in, It is unknown; another unknown parameter is cell viscosity. .
[0257] In S314, after the cells pass through the contraction phase and enter the expansion phase, their external load gradually decreases, and the cell morphology recovers from the altered state to a steady state. The strain data interval [t1, t2] during this recovery phase is selected for analysis. The exponential recovery model was used for fitting, and the fitting formula is shown in formula (23).
[0258] (twenty three).
[0259] in, To fit the initial strain, For the stable strain after recovery, Let be the cell relaxation time, which describes how quickly a cell recovers its deformation after the external load is removed. The cell relaxation time is obtained by combining equations (14) and (23) using an equivalent fitting method. .
[0260] S315 includes:
[0261] S315a, for and Numerical differentiation yields formula (24).
[0262] (twenty four).
[0263] in, This is the current sampling time point. For the previous time point, This is the next time point to be adopted.
[0264] S315b, substituting the first-order viscoelastic constitutive equation into formula (24), yields the cell viscosity. As shown in formula (25). (25).
[0265] In S316, the cell elastic modulus is calculated according to formula (26). (26). Where E is the cell elastic modulus.
[0266] In this embodiment, a computational path is provided for quantitatively calculating microscopic mechanical intrinsic parameters (viscosity η, elastic modulus E) from macroscopically observable cell deformation data (experimental flow characteristics) by constructing a specific inversion algorithm that combines strain and stress with the first-order viscoelastic constitutive equation. This step is the core algorithmic implementation for transforming image data into theoretical mechanical properties, providing direct parameter basis for quantitative classification.
[0267] In one embodiment of this application, the contraction segment is the portion where the microchannel width gradually changes from 2.5A to 0.85A, and the expansion segment is the portion where the microchannel width gradually changes from 0.85A to 2.5A, where A is the equivalent diameter of the cell.
[0268] Specifically, the microfluidic channel adopted in this application has a hyperbolic symmetrical structure. Therefore, this method can perform mechanical property inversion calculations for a single cell in two symmetrical positions under the same load in a single experiment. Furthermore, to ignore the impact of different cell sizes on the calculation of the mechanical properties of this invention, the calculations are performed in two positional intervals: a channel width of 2.5A to 0.85A and a channel width of 0.85A to 2.5A (where A is the average diameter of the cell). In this case, regardless of the cell size, a corresponding channel segment can be found within the channel, and the load application remains constant for experiments with cells of different sizes.
[0269] Combining the above steps S310 to S330 and this embodiment, it can be seen that, under this embodiment, through the calculations in S310 and S320, two sets of theoretical mechanical property parameters can be obtained. One set is obtained from a pipe width of 2.5A to 0.85A, and the other set is obtained from a pipe width of 0.85A to 2.5A.
[0270] The following describes how S330 is executed in this embodiment, that is, how to perform theoretical classification in order to obtain the theoretical mechanical property parameters of two cell samples to be tested, namely the theoretical mechanical property parameters of cell sample A and cell sample B.
[0271] The first step involves selecting cell data located within the microchannel at a width ranging from 2.5A to 0.85A as input. Assuming there are cells A and B at this point, two sets of mechanical property parameters are obtained. and The first mechanical property calculation and classification were completed by comparing the mechanical parameters of the two.
[0272] Specifically, due to It is possible to base on The values are calculated from E. Therefore, we compare... and , and If at least one set of data is inconsistent, it indicates that cell A and cell B are not the same type of cell, thus completing cell typing. Conversely, if... equal ,and equal This indicates that cell A and cell B cannot be distinguished, and cell typing is not possible.
[0273] The second step uses cell data within the microchannel, specifically the portion of the channel ranging from 0.85 Å to 2.5 Å in width, as input. This yields two sets of mechanical property parameters. and The second mechanical property calculation and classification were completed by comparing the mechanical parameters of the two.
[0274] Specifically, similar to the first step, we compare and , and If at least one set of data is inconsistent, it means that cell A and cell B are not the same type of cell, thus completing cell typing.
[0275] The mechanical properties of cells differ between their anterior and posterior segments because each cell exhibits a unique memory effect. The third step involves cell typing based on the differences between the first and second theoretical mechanical property parameters. These differences characterize the cell's memory effect.
[0276] Step 3, obtain and differences and The differences are then used to classify the types.
[0277] Specifically, I can compare the differences by calculating the average. We compare... and , and If at least one set of data is inconsistent, it means that cell A and cell B are not the same type of cell, thus completing cell typing.
[0278] In this embodiment, we can use any one of the first, second, and third steps to perform cell theoretical typing, or any two steps, or all three steps.
[0279] During the S430 comparison, due to the memory effect—cells recover their shape after being compressed—different cell types recover their shape to varying degrees, resulting in different restored shapes. Generally, different cell types show similar shapes in the first half (2.5A to 0.85A), but the differences are more significant in the second half (0.85A to 2.5A). We compare in two segments to avoid the problem of low randomness, as calculating only the second half would introduce some randomness. While the shape changes in the first half are not significant, unexpected situations can still occur. Using two segments for comparison is more accurate. Secondly, it ensures comprehensive data collection. Even if the differences in the first half are small, the data can still be used for research.
[0280] In this embodiment, by defining the range of the contraction and expansion segments (2.5A to 0.85A) using the cell's equivalent diameter A as a scale benchmark, the detection interval is adaptive to the cell size. This allows cells of different sizes to undergo the same proportion of compression and recovery processes within the channel, ensuring the consistency of mechanical loading conditions and fundamentally eliminating the systematic bias caused by differences in cell size in the mechanical property inversion results.
[0281] This embodiment employs a symmetrical hyperbolic contraction-release structure, providing a dual-load path verification mechanism. The symmetry of the microchannel allows the same cell to undergo two corresponding loading processes at geometrically symmetrical positions: the force increases in the first half and decreases in the second half. This enables independent secondary analysis and cross-validation of the same cell, reducing random errors and single-measurement bias. Furthermore, the introduction of a dynamic inversion algorithm based on cell memory significantly improves the accuracy and stability of parameter inversion.
[0282] Moreover, during the comparison process, by comparing the mechanical response at the same deformation location, the differences in cells under different loading paths (such as memory / viscoelastic effects) are used to perform dynamic inversion, making the inversion parameters of membrane constitutive models such as Skalak closer to the real mechanical process, thus improving the reliability, repeatability and robustness of the inversion.
[0283] In one embodiment of this application, S400 includes matching experimental flow characteristic data of the cell sample to be tested with simulated flow characteristic data in a simulation database, and matching experimental electrical impedance signals of the cell sample to be tested with simulated electrical impedance signals in a simulation database. Based on the matching results, mechanical property inversion and classification of the cell sample to be tested are performed to obtain simulation matching classification results, including:
[0284] S410, the experimental flow characteristic data of the cell sample to be tested is matched with the simulated cell deformation data corresponding to different simulated mechanical property parameters in the simulation database, and the first matching result between the two is obtained.
[0285] S420, the experimental electrical impedance signal of the cell sample to be tested is matched with the simulated electrical impedance signal corresponding to different simulated mechanical property parameters in the simulation database, and the second matching result of the successful matching between the two is obtained;
[0286] S430, when the simulated mechanical property parameters pointed to by the first matching result are consistent with the simulated mechanical property parameters pointed to by the second matching result, the simulated mechanical property parameters of the cell sample to be tested are determined and the typing is completed to obtain the simulated matching typing result.
[0287] Specifically, S510 involves matching and comparing the experimental flow characteristic data of the cell sample to be tested with the deformation data of each group of experimental cells in the simulation database, and classifying and identifying the cell type in the cell sample to be tested based on the mechanical property parameters of the matching results.
[0288] Specifically, the aforementioned processing yields an exported CSV file containing motion characteristics under different mechanical parameters. By comparing the kinematic parameters, the experimental data and simulation data can be matched and compared.
[0289] As shown in Table 2, Table 2 is a Taylor deformation index-membrane bending energy mapping table. The data in Table 2 are from the simulation model provided in this application when the membrane bending energy is changed. Under different membrane bending energies The corresponding deformation parameter data.
[0290] Table 2 - Taylor Deformation Index - Eb Mapping Table
[0291]
[0292] The different rows in Table 2 represent the Taylor deformation index at different times. It should be noted that Table 2 only represents the membrane bending energy for a single cell; that is, there is one Table 2 for each single cell type sample.
[0293] As shown in Table 2, there are several different Cell deformation parameter data under certain conditions; this invention compares the Taylor deformation index of the experimental results with the simulated Taylor deformation index, assuming that the deformation parameters of the experiment are similar to those of the simulated Taylor deformation index. If the deformation parameters are completely or almost identical, it proves that the reorganized simulation parameters are completely consistent with the experimental process, and the mechanical properties of the experimental cells and simulated cells are consistent. At this point, the inversion of the mechanical properties of the experimental cells is complete. Specifically, we can plot the Taylor deformation index obtained from the experiment as a curve according to the order of occurrence, and compare the experimental curve with other curves in the simulation database. For example, during the comparison process, we can iterate through the experimental curve of cell A and compare it with all the simulation curves in the simulation database. We find that the experimental curve of cell A and a simulated curve... The simulation curve showed the highest consistency. Similarly, by performing a comprehensive comparison between the experimental curve of cell B and all simulation curves in the simulation database, we found that the experimental curve of cell B and a simulation curve showed the best consistency. The simulation curves showed the highest degree of agreement. Therefore, we can naturally determine the differences between cell A and cell B, and trace their specific cell types, thus completing the S410 typing and obtaining the first matching result.
[0294] Similarly, in S420, we can also plot the electrical impedance values obtained from the experiment on the cell sample as a curve according to the order of occurrence, such as... Figure 8 and Figure 9 As shown, the experimental electrical impedance curves of the cell samples to be tested will be compared with multiple simulated electrical impedance curves in the simulation database. For example, during the comparison process, the experimental curve of cell A will be compared with multiple simulated electrical impedance curves in the simulation database in a traversal manner. We found that the experimental curve of cell A and a single simulated electrical impedance curve in the database... The simulated electrical impedance curve showed the highest agreement. Similarly, we found that the experimental curve of cell B and a [other cell type] showed the best agreement. The simulated electrical impedance curve showed the highest degree of agreement. Therefore, we can naturally determine the difference between cell A and cell B, and trace their specific cell types, thus completing the S420 typing and obtaining the second matching result.
[0295] In S430, if the curve of the Taylor deformation parameter of the cell sample to be tested and The simulation curve showed the highest agreement, and the experimental impedance curve of the cell sample being tested also matched the simulation curve. The simulation impedance curves showed the highest degree of agreement, indicating that the three were consistent and the model was successfully classified.
[0296] In this embodiment, by jointly matching the experimental deformation parameters and experimental impedance signals with corresponding data in the simulation database, and employing a "successful match for both" judgment logic, collaborative verification of both mechanical and electrical evidence chains is achieved. This method reduces the dependence on matching results of single-mode data and, by utilizing the consistency of multi-physics information, significantly improves the anti-interference capability and confidence level of the classification results.
[0297] Finally, let me explain S500. It compares the theoretical typing results with the simulation matching typing results, and verifies the accuracy of the cell typing of the cell sample to be tested based on the comparison results.
[0298] Specifically, in the theoretical comparison process, the mechanical property parameters we used are: We can obtain a theoretical classification conclusion on whether test cell A and test cell B can be distinguished.
[0299] During the simulation, we can also obtain the classification conclusion of whether test cell A and test cell B can be distinguished.
[0300] If the theoretical typing results and the simulation matching typing results are consistent, it means that the cell typing is accurate and the typing accuracy verification is passed.
[0301] like Figure 10 As shown, Figure 10 This is a comparison result diagram obtained by comparing theoretical classification results with simulation matching classification results in a scale-adaptive dynamic inversion analysis method for cell mechanical-electrical impedance characteristics provided in one embodiment of this application.
[0302] See Figure 10 , Figure 10 (a) shows the comparison results of the contraction segment, and (b) shows the comparison results of the expansion segment. The horizontal axis represents the initial diameter of the cell, characterizing the cell size. The horizontal axis shows the distribution range of the diameters of the four different cell types in the contraction and expansion segments. The gray dots represent IOSE80 cells, the blue dots represent SK0V3 cells, the red dots represent A2780 cells, and the green dots represent OVCAR3 cells. Different colored dots represent different cell types.
[0303] The vertical axis represents the cell relaxation time. .pass Figure 10 It can be seen that the blue and green dots have a high degree of overlap in the distribution area of the contraction segment, indicating low distinguishability and difficulty in classification. However, the overlap between the blue and green distribution areas is not as high in the expansion segment. The red dots do not overlap much with dots of other colors in either the contraction or expansion segment.
[0304] This application also provides a microfluidic experimental system.
[0305] In one embodiment of this application, the microfluidic experimental system is used to implement the scale-adaptive dynamic inversion analysis method of cell mechanical-electrical impedance properties mentioned in any of the foregoing embodiments.
[0306] like Figure 2 As shown, the microfluidic experimental system includes a pressure driving device 10, a flow rate sensor 20, a microfluidic chip 30, a cell impedance meter 40, a high-speed imaging device 50, a data processing device 60, and a recovery device 70.
[0307] The microfluidic chip 30 has microchannels internally constructed. A pressure-driven device 10 drives a single cell sample to flow through the microchannels. A high-speed imaging device 50 continuously captures images of cell flow within the microchannels, forming an image sequence. A flow rate sensor 20 detects flow parameters. A data processing device 60 is communicatively connected to the high-speed imaging device 50. The data processing device 60 is also communicatively connected to the flow rate sensor 20. The data processing device 60 is configured to perform a scale-adaptive dynamic inversion analysis method of cell biomechanical-electrical impedance properties as described in any of the foregoing embodiments.
[0308] Specifically, one embodiment of the structure of the microfluidic chip 30 proposed in this application is, for example... Figure 4 As shown, it mainly consists of an inlet area 310, a filtration area 320, a cell detection area 330, and an outlet area 340.
[0309] Alternatively, one embodiment of the microfluidic experimental system can be as follows: Figure 3 As shown. The collection device 70 of the microfluidic experimental system can be a collection bottle. The pressure driving device 10 can include a syringe pump 210. The syringe pump 210 injects the cell suspension into the inlet region 310 of the microfluidic chip 30, and the cell suspension flows through the cell detection region 330 and into the collection bottle.
[0310] The high-speed imaging device 50 may include a high-speed microscopic imaging module. This module may include a microscope 520 and a CCD high-speed camera 510, enabling real-time monitoring of morphological changes in cells flowing through the detection area. Images recorded by the CCD high-speed camera 510 can be uploaded to a data processing device 60. A flow rate sensor 20 is installed on the pipeline through which the syringe pump injects cell suspension into the microfluidic chip 30. The flow rate sensor 20 can transmit detected flow parameters to the data processing device 60. The data processing device 60 may include an industrial control computer. The industrial control computer can perform real-time monitoring and feedback control of the microfluidic experiment based on the uploaded flow parameters. It can also perform subsequent cell mechanical property inversion based on the images recorded by the CCD high-speed camera 510.
[0311] The technical features of the above embodiments can be combined arbitrarily, and the execution order of the method steps is not restricted. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.
Claims
1. A scale-adaptive dynamic inversion analysis method for cellular mechanical-electrical impedance properties, characterized in that, include: Simulations were used to obtain simulated flow characteristic data and simulated electrical impedance signals for each single cell sample. Based on the simulated flow characteristic data and simulated electrical impedance signals, a simulation database was established that included the correspondence between simulated mechanical property parameters, simulated cell deformation data and simulated electrical impedance signals. Acquire experimental flow characteristics and experimental impedance signals of the cell samples to be tested; A theoretical model of cell mechanical properties is established. The theoretical deformation data of the symmetrical contraction and expansion segments of the cells in the microchannel are obtained based on the experimental flow characteristic data of the cell samples to be tested. Based on the different theoretical deformation data of the cells in the contraction and expansion segments, at least one set of theoretical cell mechanical property parameters is dynamically derived. Based on the difference in the response of the corresponding theoretical cell mechanical property parameters in the contraction and expansion segments, the cell samples to be tested are classified to obtain the theoretical classification results. The contraction section is the part where the microchannel width gradually changes from 2.5A to 0.85A, and the expansion section is the part where the microchannel width gradually changes from 0.85A to 2.5A, where A is the equivalent diameter of the cell; The experimental flow characteristic data of the cell sample to be tested is matched with the simulated flow characteristic data in the simulation database, and the experimental electrical impedance signal of the cell sample to be tested is matched with the simulated electrical impedance signal in the simulation database. Based on the matching results, the mechanical properties of the cell sample to be tested are inverted and classified to obtain the simulation matching classification results. The theoretical typing results are compared with the simulated matching typing results, and the accuracy of the cell typing of the cell sample to be tested is verified based on the comparison results.
2. The scale-adaptive dynamic inversion analysis method for cell biomechanical-electrical impedance properties according to claim 1, characterized in that, The acquisition of experimental flow characteristic data and experimental impedance signals of the cell samples to be tested includes: The cell sample to be tested is injected into the microfluidic chip and the cell flow is driven, while the image sequence of the cell flow and the electrical impedance signal are acquired simultaneously. Image analysis was performed on the image sequence to obtain experimental flow characteristic data of the cells.
3. The scale-adaptive dynamic inversion analysis method for cell mechanical-electrical impedance properties according to claim 1, characterized in that, The microchannel has a hyperbolic symmetrical structure and includes, from left to right, a first straight tube section, a constriction section, a slit section, an expansion section, and a second straight tube section. The diameter gradually decreases from the first straight tube section to the constriction section, the diameter continues to decrease from the constriction section to the slit section, the diameter gradually increases from the slit section to the expansion section, and the diameter continues to increase from the expansion section to the second straight tube section. The center point of the slit section is the point where the diameter of the entire microchannel is the smallest.
4. The scale-adaptive dynamic inversion analysis method for cellular mechanical-electrical impedance properties according to claim 3, characterized in that, A differential electrode pair is installed in each of the first straight tube section, the contraction section, the slit section, the expansion section, and the second straight tube section. A constant current is passed through the differential electrode pairs. By monitoring the voltage changes generated when the cell flows through different regions of each section, the voltage is converted into resistance to obtain the impedance signal.
5. The scale-adaptive dynamic inversion analysis method for cell biomechanical-electrical impedance properties according to claim 1, characterized in that, The process involves acquiring simulated flow characteristic data and simulated electrical impedance signals for each individual cell type through simulation, and establishing a simulation database based on these data, which includes the correspondence between simulated mechanical property parameters, simulated cell deformation data, and simulated electrical impedance signals. Set different combinations of simulated cell mechanical property parameters; Simulated cell deformation data corresponding to different single cell types were obtained through cell fluid-structure interaction model simulation. The simulated cell deformation data is input into the cell electrical signal simulation model, and the simulated electrical impedance signal corresponding to different single cell samples is obtained through the simulation of the cell electrical signal simulation model. Establish and store the corresponding data relationships among the simulated mechanical property parameters, simulated cell deformation data, and simulated electrical impedance signals.
6. The scale-adaptive dynamic inversion analysis method for cellular mechanical-electrical impedance properties according to claim 1, characterized in that, The establishment of a theoretical model for cell mechanical properties involves obtaining theoretical deformation data of the symmetrical contraction and expansion segments of cells within the microchannel based on experimental flow characteristic data of the cell samples under test. Based on the different theoretical deformation data of cells within the contraction and expansion segments, at least one set of theoretical cell mechanical property parameters is dynamically derived. Furthermore, based on the differences in the responses of the corresponding theoretical cell mechanical property parameters between the contraction and expansion segments, cell typing is performed on the cell samples under test to obtain theoretical typing results, including: Based on the experimental flow characteristic data and first-order constitutive equation of the contraction zone, the first theoretical mechanical property parameters of the cell when it is in the contraction zone are calculated. Based on the experimental flow characteristic data and first-order constitutive equation of the expansion section, the second theoretical mechanical property parameters of the cell when it is in the expansion section are calculated. Cell typing is performed based on the first theoretical mechanical property parameter and the second theoretical mechanical property parameter to obtain the theoretical typing result.
7. The scale-adaptive dynamic inversion analysis method for cellular mechanical-electrical impedance properties according to claim 6, characterized in that, The calculation of the first theoretical mechanical property parameters of the cell when it is in the contraction phase, based on the experimental flow characteristic data and the first-order constitutive equation of the contraction phase, includes: Cellular strain was constructed based on experimental flow characteristic data of the contraction segment; Based on the geometric functions and flow parameters of the microchannel, the equivalent shear stress experienced by the cell is constructed; Construct the first-order viscoelastic constitutive equation; Based on the cell strain recovery curve under the corresponding first-order viscoelastic constitutive equation, the cell relaxation time is obtained by fitting; the cell strain, cell relaxation time and the equivalent shear stress on the cell are substituted into the first-order viscoelastic constitutive equation to calculate the cell viscosity by inversion. The cell elastic modulus is calculated based on the cell viscosity and the cell relaxation time to obtain the first theoretical mechanical property parameter that includes cell viscosity and cell elastic modulus.
8. The scale-adaptive dynamic inversion analysis method for cell biomechanical-electrical impedance properties according to claim 1, characterized in that, The process involves matching the experimental flow characteristic data of the cell sample to be tested with the simulated flow characteristic data in the simulation database, and matching the experimental electrical impedance signal of the cell sample to be tested with the simulated electrical impedance signal in the simulation database. Based on the matching results, the mechanical properties of the cell sample to be tested are inverted and classified to obtain the simulation matching and classification results, including: The experimental flow characteristic data of the cell sample to be tested are matched with the simulated cell deformation data corresponding to different simulated mechanical property parameters in the simulation database, and the first matching result between the two is obtained. The experimental electrical impedance signal of the cell sample to be tested is matched with the simulated electrical impedance signal corresponding to different simulated mechanical property parameters in the simulation database to obtain the second matching result of the successful match between the two. When the simulated mechanical property parameters pointed to by the first matching result are consistent with the simulated mechanical property parameters pointed to by the second matching result, the simulated mechanical property parameters of the cell sample to be tested are determined and the typing is completed, thus obtaining the simulated matching typing result.
9. The scale-adaptive dynamic inversion analysis method for cell mechanical-electrical impedance properties according to claim 1, characterized in that, The experimental flow characteristic data includes the geometric characteristics of individual cells; The individual geometric features of the cell include at least one of the following: cell minor axis length, cell major axis length, cell equivalent area, cell equivalent perimeter, cell centroid coordinates, cell roundness, and cell Taylor deformation index.
Citation Information
Patent Citations
Biological living cell surface mechanical property evaluation method based on GSA optimization neural network
CN115547405A
Device and system for measuring properties of cells and method of measuring properties of cells using the same
KR1020100136181A
