Tumor cell culture environment regulation and control system based on programmable mechanical characteristics

Abstract

The invention relates to the technical field of cell culture, in particular to a tumor cell culture environment regulation and control system and method based on programmable mechanical characteristics. The system comprises a mechanical characteristic adjustment module, a mechanical sensor signal collection module, a mechanical sensor calibration module, a cell-mechanical signal coupling relation calculation module and a cell biological behavior regulation and control module, data related to mechanical characteristics are collected, transmitted, stored and processed by simulating the mechanical characteristics of an extracellular matrix, a mechanical sensor is calibrated, and the cell biological behavior is regulated and controlled. According to the system, the topological size and density of the microcarrier are adjusted, the mechanical characteristics of an extracellular matrix are accurately simulated, and culture conditions closer to the in-vivo environment are provided for tumor cells.

Description

Technical Field

[0001] The present invention relates to the field of cell culture technology, and specifically to a tumor cell culture environment control system and method based on programmable mechanical characteristics. The system is particularly suitable for accurately simulating the in vivo mechanical microenvironment and achieving precise control of the biological behavior of tumor cells. Background Art

[0002] The biological behavior of tumor cells is significantly influenced by their surrounding microenvironment. In particular, the mechanical characteristics of the microenvironment play a key role in regulating tumor cell proliferation, differentiation, migration, and invasion. Traditional in vitro tumor cell culture systems typically use two-dimensional rigid culture media that cannot accurately simulate the complex three-dimensional mechanical environment in the body, resulting in significant discrepancies between experimental results and in vivo conditions.

[0003] Some existing studies have focused on the impact of the mechanical microenvironment on tumor cell behavior, for example by modifying the stiffness and topology of the culture substrate to study cellular responses. However, these approaches typically focus on a single mechanical parameter, making it difficult to comprehensively simulate the complex in vivo mechanical environment. Furthermore, they lack quantitative analysis of the relationship between mechanical characteristics and cellular responses, hindering precise regulation. Furthermore, existing systems generally lack real-time monitoring and feedback mechanisms, making it difficult to dynamically adjust the mechanical environment based on cellular responses.

[0004] Therefore, there is an urgent need for a system that can accurately simulate and regulate the mechanical characteristics of the tumor cell culture environment, achieve precise regulation of the biological behavior of tumor cells, and provide an experimental platform for tumor research and drug screening that is closer to the in vivo environment. Summary of the Invention

[0005] The purpose of the present invention is to provide a tumor cell culture environment control system and method based on programmable mechanical characteristics, which can achieve precise control of the biological behavior of tumor cells by accurately simulating and controlling the mechanical characteristics of the cell culture environment.

[0006] The present invention proposes a tumor cell culture environment control system based on programmable mechanical characteristics, comprising:

[0007] Mechanical characteristics regulation module, used to simulate the mechanical characteristics of the extracellular matrix, including the extracellular matrix mechanical characteristics simulation unit and the biomaterial screening unit;

[0008] a mechanical sensor signal collection module, connected to the mechanical characteristic adjustment module, for collecting, transmitting, storing and processing data related to the mechanical characteristics, comprising a mechanical sensor signal input unit, a mechanical sensor signal processing unit and a mechanical sensor signal input and output unit;

[0009] a mechanical sensor calibration module, connected to the mechanical sensor signal collection module, for calibrating the mechanical sensor, comprising a mechanical sensor calibration signal input unit, a mechanical sensor calibration unit, and a mechanical sensor calibration signal output unit;

[0010] a cell-mechanical signal coupling relationship calculation module, connected to the mechanical sensor calibration module and the mechanical sensor signal collection module, for calculating the cell mechanical response threshold, comprising a signal extraction unit, a signal analysis unit, and an output unit, wherein the signal analysis unit establishes a dynamic coupling model based on the multidimensional mechanical signal and the cell response data, and calculates the cell mechanical response threshold through an adaptive algorithm;

[0011] The cell biological behavior regulation module is connected to the cell-mechanical signal coupling relationship calculation module and the mechanical characteristic adjustment module, and is used to regulate the cell biological behavior. It includes a mechanical sensor signal input unit, an adjustment threshold setting unit and an adjustment signal generation unit. The adjustment signal generation unit sends the adjusted signal to the mechanical characteristic adjustment module, and the mechanical characteristic adjustment module receives the adjustment signal and outputs the adjusted mechanical characteristics.

[0012] Preferably, the mechanical sensor includes at least one of a piezoresistive sensor, a resistive sensor, a capacitive sensor, a fiber optic sensor, an electromagnetic sensor, a friction sensor, and a piezoelectric sensor.

[0013] Preferably, the extracellular matrix mechanical characteristic simulation unit is used to simulate the topological structure of the extracellular matrix, and the topological structure is divided into porous or continuous.

[0014] Preferably, the extracellular matrix mechanical characteristic simulation unit adjusts the mechanical characteristics of the microcarriers by selecting collagen microcarriers and / or elastin microcarriers and adjusting the topological size and / or topological density of the microcarriers.

[0015] Preferably, the signal analysis unit comprises:

[0016] The subunit for constructing a dynamic coupling model of mechanical and biological signals is used to establish a multidimensional mapping relationship between mechanical signals and cellular responses based on tensor analysis, and introduces timing analysis to capture the time delay effect between changes in the mechanical environment and cellular responses;

[0017] An adaptive mechanical response threshold calculation subunit, which is used to automatically identify the critical threshold of cell mechanical response based on a Bayesian optimization framework and improve the calculation accuracy in the critical threshold range;

[0018] The cell behavior prediction subunit is used to construct a cell proliferation / differentiation behavior prediction model and predict cell fate under different mechanical environments.

[0019] Preferably, the adaptive mechanical response threshold calculation subunit calculates the cell mechanical response threshold by the following steps:

[0020] Initialize the prior threshold parameter distribution;

[0021] Collect cell-mechanical response data;

[0022] Compute the posterior distribution;

[0023] identifying a most likely threshold interval based on the posterior distribution;

[0024] Applying mesh refinement in the critical interval to improve accuracy;

[0025] Output the optimized threshold value and its uncertainty estimate.

[0026] Preferably, the signal analysis unit also includes an intelligent mechanical-biosignal deep mapping network sub-unit, which is used to construct a deep learning architecture, capture complex mechanical-biosignal mapping relationships, and introduce an attention mechanism to automatically identify the mechanical feature combinations that have the most significant impact on cell behavior.

[0027] Preferably, the network structure of the intelligent mechanical-biological signal deep mapping network subunit includes:

[0028] Input layer, used to receive multi-dimensional mechanical feature vectors;

[0029] Feature extraction layer, used to capture the spatiotemporal characteristics of mechanical features;

[0030] Attention layer, used to automatically identify key mechanical features and their weights;

[0031] Mapping layer, used to establish the mapping relationship between mechanical characteristics and cell behavior;

[0032] The output layer is used to generate cell mechanical response thresholds and uncertainty estimates.

[0033] Preferably, the mechanical characteristic regulation-related data includes at least one of microcarrier density, topological density, and topological size.

[0034] Preferably, the tumor cell culture environment control system comprises the following steps:

[0035] Step 1: The mechanical characteristics adjustment module selects collagen microcarriers and / or elastin microcarriers, and adjusts the topological size and / or topological density of the microcarriers to obtain mechanical characteristics that simulate the extracellular matrix;

[0036] Step 2: Adjust the stress-strain relationship and displacement change of the mechanical sensor to adjust the sensing sensitivity;

[0037] Step 3: the mechanical sensor signal input unit inputs the mechanical characteristic control related data into the mechanical sensor signal processing unit and the mechanical sensor calibration signal input unit;

[0038] Step 4: the mechanical sensor signal processing unit processes the mechanical characteristic control related data and sends the processed data to the mechanical sensor signal input and output unit;

[0039] Step 5: the mechanical sensor calibration signal input unit sends the mechanical characteristic control related data to the mechanical sensor calibration unit, and the mechanical sensor calibration unit adjusts and outputs the calibration signal to the mechanical sensor calibration signal output unit;

[0040] Step 6: the mechanical sensor calibration signal output unit outputs the calibration signal to the mechanical sensor signal input unit;

[0041] Step 7: the mechanical sensor signal input unit sends the received mechanical characteristic control related data to the mechanical sensor signal processing unit;

[0042] Step 8: The mechanical sensor signal processing unit processes the mechanical characteristic control related data and sends the processed data to the mechanical sensor signal input and output unit;

[0043] Step nine, the mechanical sensor signal input and output unit sends the mechanical characteristic control related data and calibration data to the mechanical sensor to control the mechanical characteristics of the microcarrier;

[0044] Step 10, the mechanical sensor collects and transmits the microcarrier mechanical characteristic regulation related data to the signal extraction unit;

[0045] Step 11, the signal analysis unit calculates the cell mechanical response threshold according to the mechanical characteristic regulation-related signal and the cell-mechanical sensor signal coupling relationship, and outputs the calculated cell mechanical response threshold to the adjustment threshold setting unit;

[0046] Step 12: the mechanical sensor signal input unit sends the received mechanical characteristic control related data to the adjustment threshold setting unit;

[0047] Step 13: the adjustment threshold setting unit compares the received mechanical characteristic regulation related data with the cell mechanical response threshold, and sends the comparison result to the adjustment signal generating unit;

[0048] Step 14: the adjustment signal generating unit sends the adjusted mechanical characteristic control related data to the mechanical characteristic adjustment module;

[0049] In step 15, the mechanical characteristic regulation module regulates the microcarrier according to the received mechanical characteristic regulation related data to complete the tumor cell culture environment regulation process.

[0050] The beneficial effects of the present invention include:

[0051] 1. High-precision mechanical environment simulation: By adjusting the topological size and density of microcarriers, the mechanical characteristics of the extracellular matrix are accurately simulated, providing tumor cells with culture conditions that are closer to the in vivo environment;

[0052] 2. Real-time dynamic regulation: Based on a closed-loop feedback mechanism, the mechanical environment is adjusted in real time according to the cell response, achieving precise dynamic regulation of the biological behavior of tumor cells;

[0053] 3. Personalized and precise regulation: Through the adaptive threshold calculation algorithm, the optimal mechanical response threshold is automatically identified for different cell types, achieving personalized and precise regulation;

[0054] 4. Improve experimental repeatability: The system integrates calibration and standardization mechanisms to significantly reduce experimental variability, improve result repeatability, and thus reduce experimental costs;

[0055] 5. Expanded application scope: Supports multiple tumor cell types and mechanical parameter regulation, providing a unified technical platform for tumor research and drug screening. BRIEF DESCRIPTION OF THE DRAWINGS

[0056] Figure 1 Schematic diagram of the overall architecture of the tumor cell culture environment control system based on programmable mechanical characteristics of the present invention;

[0057] Figure 2 Schematic diagram of the structure of the mechanical characteristics adjustment module of the present invention;

[0058] Figure 3 This is a structural diagram of the mechanical sensor signal collection module of the present invention;

[0059] Figure 4 This is a schematic structural diagram of a mechanical sensor calibration module according to the present invention;

[0060] Figure 5 Schematic diagram of the structure of the cell-mechanical signal coupling relationship calculation module of the present invention;

[0061] Figure 6 This is a schematic diagram of the structure of the cell biological behavior regulation module of the present invention;

[0062] Figure 7 This is a flow chart of the adaptive mechanical response threshold calculation algorithm of the present invention;

[0063] Figure 8This is a schematic diagram of the intelligent mechanical-biological signal deep mapping network structure of the present invention;

[0064] Figure 9 This is a flow chart of the method for regulating the tumor cell culture environment based on programmable mechanical characteristics of the present invention. DETAILED DESCRIPTION

[0065] Please refer to the attached Figure 1-9 , the specific implementation of the present invention is described in detail below with reference to the accompanying drawings. It should be noted that, in the absence of conflict, the embodiments and features in the embodiments of the present invention can be combined with each other.

[0066] Reference Figure 1 The programmable mechanical characteristics-based tumor cell culture environment control system 100 provided by the present invention includes a mechanical characteristics adjustment module 10, a mechanical sensor signal collection module 20, a mechanical sensor calibration module 30, a cell-mechanical signal coupling relationship calculation module 40, and a cell biological behavior control module 50. These five modules form a complete closed-loop control system, achieving precise control of the entire process from mechanical environment simulation to cell behavior regulation.

[0067] Reference Figure 2 The mechanical characteristic adjustment module 10 is used to simulate the mechanical characteristics of the extracellular matrix, and includes an extracellular matrix mechanical characteristic simulation unit 11 and a biomaterial screening unit 12.

[0068] The extracellular matrix mechanical characteristics simulation unit 11 is used to construct a microcarrier that simulates the mechanical structure of the extracellular matrix. In one embodiment of the present invention, the unit can simulate the mechanical characteristics of the extracellular matrix by adjusting the topological structure of the microcarrier. Figure 2 As shown, the topology can be either porous or continuous to accommodate the culture requirements of different tumor cell types. Preferably, for pancreatic cancer cells, a porous structure with a pore size of 5-10 μm and a porosity of 60-70% can be used to simulate the mechanical characteristics of the pancreatic cancer microenvironment; while for breast cancer cells, a porous structure with a pore size of 15-20 μm and a porosity of 40-50% can be used.

[0069] The biomaterial screening unit 12 is used to screen collagen and elastin microcarriers, the primary components of the extracellular matrix. In a preferred embodiment of the present invention, collagen microcarriers are primarily used to simulate the rigid structure of the extracellular matrix, with a mechanical modulus range of 5-20 kPa; while elastin microcarriers are primarily used to simulate the elastic structure of the extracellular matrix, with a mechanical modulus range of 1-5 kPa. By adjusting the ratio of these two microcarriers, the mechanical characteristics of different tumor microenvironments can be accurately simulated.

[0070] Preferably, the extracellular matrix mechanical characteristic simulation unit 11 can adjust the mechanical characteristics of the microcarriers by selecting collagen microcarriers and / or elastin microcarriers and adjusting the microcarrier topological size and / or topological density. For example, for a tumor cell culture environment requiring higher rigidity, the proportion of collagen microcarriers can be increased and the topological size can be reduced; for an environment requiring higher elasticity, the proportion of elastin microcarriers can be increased and the topological density can be increased.

[0071] In a specific embodiment of the present invention, for the culture of pancreatic cancer cells, the ratio of collagen microcarriers to elastin microcarriers can be set to 7:3, the topological size can be set to 8 μm, and the topological density can be set to 65% to simulate the mechanical characteristics of the pancreatic cancer microenvironment; while for breast cancer cells, the ratio can be set to 4:6, the topological size can be set to 18 μm, and the topological density can be set to 45%.

[0072] Reference Figure 3 The mechanical sensor signal collection module 20 is connected to the mechanical characteristic adjustment module 10 and is used to collect, transmit, store and process data related to the mechanical characteristics. This module includes a mechanical sensor signal input unit 21, a mechanical sensor signal processing unit 22 and a mechanical sensor signal input and output unit 23.

[0073] The mechanosensor signal input unit 21 inputs data related to mechanical characteristic regulation from the mechanosensor to the mechanosensor signal processing unit 22. In one embodiment of the present invention, this unit can utilize a multi-channel signal acquisition system to simultaneously collect mechanical characteristic data such as microcarrier density, topological density, and topological size. Preferably, the acquisition frequency can be set to 5-10 Hz to ensure that dynamic changes in cellular responses can be captured.

[0074] The mechanical sensor signal processing unit 22 receives and processes data related to mechanical characteristic control and transmits the processed data to the mechanical sensor signal input and output unit 23. During processing, this unit employs a signal preprocessing algorithm to reduce noise, normalize, and extract features from the raw mechanical signals. Specifically, wavelet transforms are used for signal noise reduction, improving the signal-to-noise ratio by up to 15-20 dB; Z-score normalization is used to make different mechanical signal types comparable; and principal component analysis is used for feature extraction, reducing data dimensionality while retaining key information.

[0075] The mechanical sensor signal input / output unit 23 receives data processed by the mechanical sensor signal processing unit 22 and outputs the processed data to the mechanical sensor. This unit is also responsible for data storage and management, ensuring data integrity and traceability. Preferably, a time-series database can be used for storage, supporting high-frequency writing and fast querying, providing a data foundation for subsequent analysis.

[0076] Reference Figure 4 The mechanical sensor calibration module 30 is connected to the mechanical sensor signal collection module 20 and is used to calibrate the mechanical sensor to ensure the accuracy of mechanical characteristic measurement. This module includes a mechanical sensor calibration signal input unit 31, a mechanical sensor calibration unit 32, and a mechanical sensor calibration signal output unit 33.

[0077] The mechanical sensor calibration signal input unit 31 inputs the signal required for calibrating the mechanical sensor to the mechanical sensor calibration unit 32. In the present invention, the mechanical sensor may include at least one of a piezoresistive sensor, a resistive sensor, a capacitive sensor, an optical fiber sensor, an electromagnetic sensor, a friction sensor, and a piezoelectric sensor. Preferably, different types of sensors can be selected according to different mechanical characteristic measurement requirements. For example, for the measurement of microcarrier density, a piezoresistive sensor can be preferably selected, and its sensitivity can reach 0.1 g / cm 3 For the measurement of topological density, capacitive sensors are preferred, with a sensitivity of up to 1% porosity change.

[0078] The mechanical sensor calibration unit 32 calibrates the mechanical sensor and inputs the calibrated mechanical sensor signal into the mechanical sensor calibration signal output unit 33. The calibration process involves adjusting parameters such as the stress-strain relationship and displacement change of the mechanical sensor to adjust the sensor sensitivity. Specifically, for piezoresistive sensors, the calibration unit adjusts sensitivity by adjusting the gauge factor between 0.8 and 1.2; for capacitive sensors, sensitivity adjustment is achieved by adjusting the capacitor spacing between 5 and 15 μm.

[0079] The mechanical sensor calibration signal output unit 33 outputs the mechanical sensor calibration signal and outputs the calibration signal to the mechanical sensor signal input unit 21. In this way, when the mechanical sensor performs subsequent mechanical characteristic measurements, the accuracy and reliability of the measurement results will be ensured based on the calibrated parameters.

[0080] In a specific embodiment of the present invention, for the measurement of microcarrier density, a standard sample (e.g., a known density of 1.05 g / cm 3 The piezoresistive sensor was calibrated by adjusting the gauge factor to 1.05 so that it could be adjusted within the range of 1.0-1.1 g / cm 3 The linearity within the range reaches more than 99%; the calibration signal is then transmitted to the signal collection module to ensure the accuracy of subsequent actual sample measurements.

[0081] Reference Figure 5The cell-mechanical signal coupling relationship calculation module 40 is connected to the mechanical sensor calibration module 30 and the mechanical sensor signal collection module 20. It is the core innovative module of the present invention and is used to calculate the cell mechanical response threshold. This module includes a signal extraction unit 41, a signal analysis unit 42, and an output unit 43.

[0082] The signal extraction unit 41 is used to receive calibration signals from the mechanical sensors. This unit simultaneously receives signals from different types of sensors through a multi-channel data interface and performs preliminary data integration. Preferably, a JSON formatted data structure is used for transmission and storage to ensure data structuring and scalability. A typical data structure is as follows:

[0083]

[0084]

[0085] The signal analysis unit 42 calculates the cell mechanical response threshold based on the mechanical characteristic-related signals sent by the mechanical sensor signal collection module 20 and the mechanical sensor calibration module 30 and the cell-mechanical sensor signal coupling relationship. This is the core innovation of the present invention and includes multiple subunits:

[0086] The mechanical-biological signal dynamic coupling model construction subunit 421 is used to establish a multidimensional mapping relationship between mechanical signals and cellular responses based on tensor analysis, and introduces time series analysis to capture the time delay effect between mechanical environment changes and cellular responses. The mathematical expression of this model is:

[0087] R(t)=F(M(t-τ),H,C),

[0088] Where: R(t) is the cell response function, which represents the cell biological response at time t, and the unit is the number of cells / hour (for proliferation rate) or percentage (for differentiation rate); M(t-τ) is the mechanical stimulation function with time delay τ, which represents the mechanical environment parameter vector at time t-τ, including the microcarrier density (g / cm 3 ), topological density (%), topological size (μm) and other parameters; τ is the time delay parameter, the unit is minutes, which reflects the time lag between mechanical stimulation and cell response; H is the historical response matrix with a dimension of m×n, where m is the number of historical time points and n is the number of response parameters, which is used to capture the historical information of cell response; C is the cell type-related parameter matrix with a dimension of k×p, where k is the number of cell types and p is the number of related parameters, which reflects the difference in sensitivity of different cell types to mechanical stimulation; F is a nonlinear mapping function, which is implemented using a multi-layer perceptron structure and is used to process the nonlinear relationship between mechanical stimulation and cell response.

[0089] In practical applications, the time delay parameter τ is typically set in the range of 5-120 minutes and automatically adjusted according to different cell types. For example, for breast cancer cells that respond quickly to mechanical stimuli, τ can be set to 15-30 minutes; while for pancreatic cancer cells that respond more slowly, τ can be set to 60-90 minutes. The nonlinear coefficient matrix uses a 4×4 dimension to capture the interactions between different mechanical characteristics. The threshold sensitivity is set in the range of 0.01-0.1 kPa and can be dynamically adjusted according to the cell type.

[0090] The adaptive mechanical response threshold calculation subunit 422 is used to automatically identify the critical threshold of cell mechanical response based on the Bayesian optimization framework and improve the calculation accuracy in the critical threshold range. The process of the algorithm is as follows Figure 7 As shown, the following steps are included:

[0091] 1) Initialize the prior threshold parameter distribution P(θ), where θ represents the threshold parameter vector with dimension d, corresponding to d mechanical characteristic parameters;

[0092] 2) Collect cell-mechanical response data D, which includes mechanical environment parameters and corresponding cell response results. The data set size is N sets of experimental data;

[0093] 3) Calculate the posterior distribution P(θ|D) using the Bayesian formula:

[0094]

[0095] Where: P(D|θ) is the likelihood function, which represents the probability of observing the data set D under a given threshold parameter θ; P(θ) is the prior distribution, which is usually assumed to be a multivariate normal distribution N(μ0,∑0), where μ0 is the prior mean vector and ∑0 is the prior covariance matrix; P(D) is the evidence factor, which is used to normalize the posterior distribution. The calculation formula is P(D) = ∫P(D|θ)P(θ)dθ.

[0096] 4) Identify the most likely threshold interval based on the posterior distribution and calculate the maximum posterior probability estimate:

[0097]

[0098] Where: θMAP is the maximum posterior probability estimate, that is, the mode of the posterior distribution P(θ|D); argmax θ It represents the value of θ that maximizes the function P(θ|D).

[0099] 5) Apply mesh refinement in key intervals to improve accuracy. Specifically, in θ MAP Within the range of ±20%, the grid density is increased by 10 times to achieve a refined search. After the grid is refined, the parameter search step size Δθ is reduced from the original step size Δθ0 to Δθ0 / 10.

[0100] 6) Output the optimized threshold value θ * and its uncertainty estimate σ * , where θ * is the optimal threshold parameter vector after refined search, σ * is the corresponding standard deviation vector, reflecting the uncertainty of parameter estimation.

[0101] In one embodiment of the present invention, for breast cancer cells, the initial prior distribution can be set to a normal distribution N(5,2 2 )(mean is 5kPa, standard deviation is 2kPa); after collecting 50 sets of experimental data, the optimal threshold θ is calculated * is 3.8 kPa, with uncertainty σ * It is 0.5kPa, indicating that breast cancer cells have the best biological response under the mechanical environment of 3.8±0.5kPa.

[0102] Cell behavior prediction subunit 423 is used to construct a cell proliferation / differentiation behavior prediction model to predict cell fate under different mechanical environments. This subunit uses a multi-output regression model to simultaneously predict cell proliferation rate, differentiation rate, and migration rate. The mathematical expression of the model is:

[0103] Y=g(X,W),

[0104] Where: Y is the cell behavior vector with a dimension of q, including q indicators such as proliferation rate (number of cells / hour) and differentiation rate (%); X is the mechanical environment parameter vector with a dimension of d, including microcarrier density (g / cm 3 ), topological density (%), topological size (μm), etc.; W is the model weight matrix with dimension d×q, which is obtained by optimizing the training data; g is a nonlinear prediction function, which can be implemented by a deep neural network, including multiple hidden layers and nonlinear activation functions.

[0105] In practical applications, the weight matrix W is obtained by training with historical data and is regularly updated to adapt to new experimental conditions. Preferably, a 5-fold cross-validation is used to evaluate the model performance to ensure that the prediction accuracy is above 85%. The mean square error (MSE) is used as the loss function for model training, and the formula is:

[0106]

[0107] Where: N is the number of training samples; Y i is the true cell behavior vector of the i-th sample; is the cell behavior vector predicted by the model.

[0108] The intelligent mechanical-biological signal deep mapping network subunit 424 is used to build a deep learning architecture to capture the complex mechanical-biological signal mapping relationship and introduce an attention mechanism to automatically identify the mechanical feature combination that has the most significant impact on cell behavior. The structure of the network is as follows Figure 8 As shown, including:

[0109] Input layer: receives a multidimensional mechanical feature vector with a dimension of d, including d features such as microcarrier density, topological density, and topological size;

[0110] Feature extraction layer: A multi-layer convolutional network is used, which includes L convolutional layers, each with f l A convolution kernel with a kernel size of k×k is used to capture the spatiotemporal characteristics of mechanical features;

[0111] Attention layer: uses the self-attention mechanism, and the calculation formula is:

[0112]

[0113] in: is the query matrix, which is obtained by linear transformation of the input features, n is the sequence length, d k is the dimension of the key vector; is the key matrix, which is obtained by linear transformation of the input features; is a value matrix obtained by linear transformation of input features, d v is the dimension of the value vector; QK T It represents the transpose multiplication of the query matrix and the key matrix, and the result dimension is n×n; is a scaling factor used to prevent the softmax gradient from disappearing due to excessive dot product values; softmax is a normalization function. Make sure the attention weights sum to 1.

[0114] Mapping layer: A deep fully connected network is used, which includes M hidden layers, each with hm neurons, to establish the mapping relationship between mechanical characteristics and cell behavior;

[0115] Output layer: Generates cell mechanical response thresholds and uncertainty estimates. The output dimension is 2q, where the first q elements are the thresholds of each indicator and the last q elements are the corresponding uncertainty estimates.

[0116] During the network training process, the learning rate is initially set to 0.001, and the cosine annealing adjustment strategy is adopted. The learning rate change formula is:

[0117]

[0118] Where: η t is the learning rate of the tth round of training; η minis the minimum learning rate, set to 0.0001; η max is the maximum learning rate, set to 0.001; t is the current training round number; T is the total number of training rounds.

[0119] The batch size is 32-128, dynamically adjusted according to the amount of data; the regularization coefficient is 0.0001-0.001 to prevent overfitting; an early stopping strategy is used, and training is stopped when the performance of the validation set does not improve for 5 consecutive rounds. The training loss function uses the mean square error with regularization:

[0120]

[0121] Where: MSE is the mean square error; λ is the regularization coefficient; P is the total number of model parameters; w i is the i-th model parameter.

[0122] Preferably, different network parameters can be preset for different types of tumor cells. For example, for breast cancer cells, the number of convolutional layers L can be set to 4, the number of attention heads to 4, and the hidden layer dimension to 256; while for pancreatic cancer cells, the number of convolutional layers L can be set to 6, the number of attention heads to 8, and the hidden layer dimension to 512.

[0123] The output unit 43 outputs the calculated cell mechanical response threshold to the cell biological behavior control module 50. This unit is responsible for converting complex calculation results into a standard format to ensure compatibility with downstream modules. In one embodiment of the present invention, the output data structure is as follows:

[0124]

[0125]

[0126] Reference Figure 6 The cell biological behavior regulation module 50 is used to regulate cell biological behavior, and includes a mechanical sensor signal input unit 51, an adjustment threshold setting unit 52 and an adjustment signal generating unit 53.

[0127] The mechanical sensor signal input unit 51 receives the cell mechanical response threshold sent by the cell-mechanical sensor signal coupling relationship calculation module 40 and the mechanical characteristic regulation-related data sent by the mechanical characteristic module 10. This unit adopts a dual-channel design to ensure the simultaneous reception and processing of the two types of data.

[0128] The adjustment threshold setting unit 52 compares the mechanical characteristic regulation related data with the cell biological behavior threshold and sends the comparison result to the adjustment signal generating unit 53. The comparison process adopts the weighted Euclidean distance calculation formula:

[0129]

[0130] Where: d is the distance between the current mechanical environment and the target threshold, the unit depends on the unit of the parameter, usually dimensionless processing; x i is the current i-th mechanical characteristic parameter, such as microcarrier density (g / cm 3 ), topological density (%), etc.; θ i is the corresponding target threshold parameter; w i is the feature weight, dimensionless, ranging from [0,1], and satisfies Reflects the importance of different features; n is the number of features; Indicates the summation of n features.

[0131] Preferably, different feature weights can be set for different tumor cell types. For example, for breast cancer cells, the weights w1, w2, and w3 of microcarrier density, topological density, and topological size can be set to 0.4, 0.4, and 0.2, respectively; while for pancreatic cancer cells, they can be set to 0.3, 0.5, and 0.2, respectively.

[0132] The adjustment signal generating unit 53 sends the adjusted signal to the mechanical characteristic adjustment module 10. This unit generates the optimized mechanical characteristic control parameters according to the comparison results. The specific algorithm is:

[0133]

[0134] in: is the adjusted i-th mechanical characteristic parameter, and its unit is the same as the original parameter; is the current parameter value; α is the adjustment rate, dimensionless, usually set between 0.1-0.5, controlling the amplitude of each adjustment; w i is the feature weight, dimensionless; θ i is the target threshold parameter; Indicates the gap between the target value and the current value, and determines the direction and magnitude of the adjustment.

[0135] The mechanical characteristic adjustment module 10 receives the adjustment signal and outputs the adjusted mechanical characteristics to complete closed-loop control.

[0136] In a specific embodiment of the present invention, for breast cancer cells, if the current microcarrier density is 0.95 g / cm 3 , while the target threshold is 0.90g / cm 3 , the weight is 0.4, and the adjustment rate is 0.3, then the adjusted microcarrier density is: 0.95 + 0.3 × 0.4 × (0.90-0.95) = 0.942 g / cm 3 Through this gradual adjustment, the system can smoothly adjust the mechanical environment parameters to the optimal range, avoiding instability caused by violent fluctuations.

[0137] In the present invention, the data related to mechanical characteristic regulation include at least one of microcarrier density, topological density, and topological size. These data are the basis for the system to regulate the mechanical environment.

[0138] Microcarrier density is usually expressed in g / cm 3 The unit reflects the mass distribution of microcarriers, which usually ranges from 0.8 to 1.2 g / cm 3 Different types of tumor cells have different preferences for microcarrier density. For example, breast cancer cells prefer a density of 0.85-0.95 g / cm 3 The optimal microcarrier density for pancreatic cancer cells is 0.95-1.05 g / cm 3 cultured at a density of .

[0139] Topological density, expressed as a percentage of porosity, reflects the porousness of the microcarrier structure and typically ranges from 40-80%. A high topological density provides more space for cell migration and proliferation. For example, for breast cancer cells, which require high migration ability, a topological density of 60-70% is ideal, while for relatively quiescent pancreatic cancer cells, a topological density of 50-60% is ideal.

[0140] The topological dimension, measured in μm, reflects the characteristic dimensions of the microcarrier structure, such as pore size, and typically ranges from 5 to 20 μm. To provide appropriate mechanical stimulation, the topological dimension must be aligned with the cell size. For example, for breast cancer cells with a diameter of approximately 10 to 15 μm, the optimal topological dimension is 12 to 18 μm; for pancreatic cancer cells with a diameter of approximately 8 to 12 μm, the optimal topological dimension is 7 to 14 μm.

[0141] Reference Figure 9 The method for controlling the tumor cell culture environment based on programmable mechanical characteristics of the present invention is implemented based on the above system and includes the following steps:

[0142] In step 1, the mechanical properties adjustment module 10 selects collagen microcarriers and / or elastin microcarriers and adjusts the microcarrier topological size and / or topological density to obtain mechanical properties that simulate the extracellular matrix. In one embodiment of the present invention, a mixture of collagen microcarriers and elastin microcarriers in a ratio of 6:4 is selected, with the topological size set to 15 μm and the topological density set to 65%, to preliminarily simulate a suitable microenvironment for breast cancer cells.

[0143] Step 2: Adjust the stress-strain relationship and displacement of the mechanical sensor to adjust the sensor sensitivity. Specifically, for piezoresistive sensors, the gauge factor can be adjusted to 1.0; for capacitive sensors, the capacitor spacing can be adjusted to 10μm to ensure that the measurement accuracy meets the experimental requirements.

[0144] Step 3: The mechanical sensor signal input unit 21 inputs the mechanical characteristic control related data into the mechanical sensor signal processing unit 22 and the mechanical sensor calibration signal input unit 31. The data transmission uses a standardized interface to ensure compatibility between different modules.

[0145] In step 4, the mechanical sensor signal processing unit 22 processes the mechanical characteristic control related data and sends the processed data to the mechanical sensor signal input and output unit 23. The processing process includes noise reduction, normalization and feature extraction to improve data quality.

[0146] In step 5, the mechanical sensor calibration signal input unit 31 sends the mechanical characteristic control related data to the mechanical sensor calibration unit 32. The mechanical sensor calibration unit 32 adjusts and outputs the calibration signal to the mechanical sensor calibration signal output unit 33. The calibration process ensures the accuracy and consistency of the measurement.

[0147] In step 6, the mechanical sensor calibration signal output unit 33 outputs the calibration signal to the mechanical sensor signal input unit 21. In this way, subsequent measurements will be performed based on the calibrated parameters, improving data reliability.

[0148] In step seven, the mechanical sensor signal input unit 21 sends the received mechanical characteristic control related data to the mechanical sensor signal processing unit 22. This step ensures that the processing unit obtains the latest data for analysis.

[0149] In step eight, the mechanical sensor signal processing unit 22 processes the mechanical characteristic control related data and sends the processed data to the mechanical sensor signal input and output unit 23. The processed data is more suitable for subsequent analysis and utilization.

[0150] In step nine, the mechanical sensor signal input and output unit 23 sends the mechanical characteristic control related data and calibration data to the mechanical sensor to control the mechanical characteristics of the microcarrier. This step achieves a preliminary adjustment of the mechanical environment of the microcarrier.

[0151] Step 10: The mechanical sensor collects and transmits the data related to the mechanical characteristics of the microcarrier to the signal extraction unit 41. This step ensures that the cell-mechanical signal coupling relationship calculation module 40 obtains the latest mechanical environment data.

[0152] In step 11, the signal analysis unit 42 calculates the cell mechanical response threshold based on the mechanical characteristic regulation-related signals and the cell-mechanical sensor signal coupling relationship, and outputs the calculated cell mechanical response threshold to the adjustment threshold setting unit 52. This is the core step of the entire system, determining the optimal mechanical response threshold of the cell through complex calculations and analysis.

[0153] In step 12, the mechanical sensor signal input unit 51 sends the received mechanical characteristic control related data to the adjustment threshold setting unit 52. This step ensures that the adjustment unit simultaneously obtains the current mechanical environment data and the target threshold data.

[0154] In step 13, the adjustment threshold setting unit 52 compares the received mechanical characteristic regulation related data with the cell mechanical response threshold and sends the comparison result to the adjustment signal generating unit 53. The comparison process uses weighted Euclidean distance calculation to quantify the difference between the current environment and the target environment.

[0155] In step 14, the adjustment signal generating unit 53 sends the adjusted mechanical characteristic adjustment related data to the mechanical characteristic adjustment module 10. This step generates specific adjustment parameters to guide subsequent environmental adjustments.

[0156] In step 15, the mechanical characteristic adjustment module 10 adjusts the microcarriers according to the received mechanical characteristic adjustment data, completing the process of regulating the tumor cell culture environment. By adjusting the physical properties of the microcarriers, precise regulation of the cell culture environment is achieved.

[0157] In a specific embodiment of the present invention, for the culture of breast cancer cells, the initial microcarrier density is set to 0.90 g / cm 3 , topological density is 65%, and topological size is 15μm. After 10 iterations of control, the system automatically adjusts the microcarrier density to 0.88g / cm 3 , the topological density was adjusted to 68%, and the topological size was adjusted to 16μm. At this time, the proliferation rate of breast cancer cells increased by 35% and the differentiation index increased by 28%, indicating that the system can effectively identify and achieve optimal culture conditions.

[0158] The above description is merely a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any modifications or substitutions that can be easily conceived by a person skilled in the art within the technical scope disclosed in the present invention should be included within the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be based on the appended claims.

Claims

1. A tumor cell culture environment control system based on programmable mechanical characteristics, characterized in that: include: Mechanical characteristics regulation module, used to simulate the mechanical characteristics of the extracellular matrix, including the extracellular matrix mechanical characteristics simulation unit and the biomaterial screening unit; a mechanical sensor signal collection module, connected to the mechanical characteristic adjustment module, for collecting, transmitting, storing and processing data related to the mechanical characteristics, comprising a mechanical sensor signal input unit, a mechanical sensor signal processing unit and a mechanical sensor signal input and output unit; a mechanical sensor calibration module, connected to the mechanical sensor signal collection module, for calibrating the mechanical sensor, comprising a mechanical sensor calibration signal input unit, a mechanical sensor calibration unit, and a mechanical sensor calibration signal output unit; a cell-mechanical signal coupling relationship calculation module, connected to the mechanical sensor calibration module and the mechanical sensor signal collection module, for calculating the cell mechanical response threshold, comprising a signal extraction unit, a signal analysis unit, and an output unit, wherein the signal analysis unit establishes a dynamic coupling model based on the multidimensional mechanical signal and the cell response data, and calculates the cell mechanical response threshold through an adaptive algorithm; The cell biological behavior regulation module is connected to the cell-mechanical signal coupling relationship calculation module and the mechanical characteristic adjustment module, and is used to regulate the cell biological behavior. It includes a mechanical sensor signal input unit, an adjustment threshold setting unit and an adjustment signal generation unit. The adjustment signal generation unit sends the adjusted signal to the mechanical characteristic adjustment module, and the mechanical characteristic adjustment module receives the adjustment signal and outputs the adjusted mechanical characteristics.

2. The tumor cell culture environment control system based on programmable mechanical characteristics according to claim 1, characterized in that: The mechanical sensor includes at least one of a piezoresistive sensor, a resistive sensor, a capacitive sensor, an optical fiber sensor, an electromagnetic sensor, a friction sensor, and a piezoelectric sensor.

3. The tumor cell culture environment control system based on programmable mechanical characteristics according to claim 1, characterized in that: The extracellular matrix mechanical characteristic simulation unit is used to simulate the topological structure of the extracellular matrix, and the topological structure is divided into porous or continuous.

4. The tumor cell culture environment control system based on programmable mechanical characteristics according to claim 1, characterized in that: The extracellular matrix mechanical characteristic simulation unit adjusts the mechanical characteristics of the microcarriers by selecting collagen microcarriers and / or elastin microcarriers and adjusting the topological size and / or topological density of the microcarriers.

5. The tumor cell culture environment control system based on programmable mechanical characteristics according to claim 1, characterized in that: The signal analysis unit includes: The subunit for constructing a dynamic coupling model of mechanical and biological signals is used to establish a multidimensional mapping relationship between mechanical signals and cellular responses based on tensor analysis, and introduces timing analysis to capture the time delay effect between changes in the mechanical environment and cellular responses; An adaptive mechanical response threshold calculation subunit, which is used to automatically identify the critical threshold of cell mechanical response based on a Bayesian optimization framework and improve the calculation accuracy in the critical threshold range; The cell behavior prediction subunit is used to construct a cell proliferation / differentiation behavior prediction model and predict cell fate under different mechanical environments.

6. The tumor cell culture environment control system based on programmable mechanical characteristics according to claim 5, characterized in that: The adaptive mechanical response threshold calculation subunit calculates the cell mechanical response threshold by the following steps: Initialize the prior threshold parameter distribution; Collect cell-mechanical response data; Compute the posterior distribution; identifying a most likely threshold interval based on the posterior distribution; Applying mesh refinement in the critical interval to improve accuracy; Output the optimized threshold value and its uncertainty estimate.

7. The tumor cell culture environment control system based on programmable mechanical characteristics according to claim 5, characterized in that: The signal analysis unit also includes an intelligent mechanical-biosignal deep mapping network sub-unit, which is used to build a deep learning architecture, capture complex mechanical-biosignal mapping relationships, and introduce an attention mechanism to automatically identify the mechanical feature combinations that have the most significant impact on cell behavior.

8. The tumor cell culture environment control system based on programmable mechanical characteristics according to claim 7, characterized in that: The network structure of the intelligent mechanical-biological signal deep mapping network subunit includes: Input layer, used to receive multi-dimensional mechanical feature vectors; Feature extraction layer, used to capture the spatiotemporal characteristics of mechanical features; Attention layer, used to automatically identify key mechanical features and their weights; Mapping layer, used to establish the mapping relationship between mechanical characteristics and cell behavior; The output layer is used to generate cell mechanical response thresholds and uncertainty estimates.

9. The tumor cell culture environment control system based on programmable mechanical characteristics according to claim 1, characterized in that: The mechanical characteristic regulation-related data includes at least one of microcarrier density, topological density, and topological size.

10. A method for regulating the tumor cell culture environment based on programmable mechanical characteristics, characterized in that: The tumor cell culture environment control system according to any one of claims 1 to 9 comprises the following steps: Step 1: The mechanical characteristics adjustment module selects collagen microcarriers and / or elastin microcarriers, and adjusts the topological size and / or topological density of the microcarriers to obtain mechanical characteristics that simulate the extracellular matrix; Step 2: Adjust the stress-strain relationship and displacement change of the mechanical sensor to adjust the sensing sensitivity; Step 3: the mechanical sensor signal input unit inputs the mechanical characteristic control related data into the mechanical sensor signal processing unit and the mechanical sensor calibration signal input unit; Step 4: the mechanical sensor signal processing unit processes the mechanical characteristic control related data and sends the processed data to the mechanical sensor signal input and output unit; Step 5: the mechanical sensor calibration signal input unit sends the mechanical characteristic control related data to the mechanical sensor calibration unit, and the mechanical sensor calibration unit adjusts and outputs the calibration signal to the mechanical sensor calibration signal output unit; Step 6: the mechanical sensor calibration signal output unit outputs the calibration signal to the mechanical sensor signal input unit; Step 7: the mechanical sensor signal input unit sends the received mechanical characteristic control related data to the mechanical sensor signal processing unit; Step 8: The mechanical sensor signal processing unit processes the mechanical characteristic control related data and sends the processed data to the mechanical sensor signal input and output unit; Step nine, the mechanical sensor signal input and output unit sends the mechanical characteristic control related data and calibration data to the mechanical sensor to control the mechanical characteristics of the microcarrier; Step 10, the mechanical sensor collects and transmits the microcarrier mechanical characteristic regulation related data to the signal extraction unit; Step 11, the signal analysis unit calculates the cell mechanical response threshold according to the mechanical characteristic regulation-related signal and the cell-mechanical sensor signal coupling relationship, and outputs the calculated cell mechanical response threshold to the adjustment threshold setting unit; Step 12: the mechanical sensor signal input unit sends the received mechanical characteristic control related data to the adjustment threshold setting unit; Step 13: the adjustment threshold setting unit compares the received mechanical characteristic regulation related data with the cell mechanical response threshold, and sends the comparison result to the adjustment signal generating unit; Step 14: the adjustment signal generating unit sends the adjusted mechanical characteristic control related data to the mechanical characteristic adjustment module; In step 15, the mechanical characteristic regulation module regulates the microcarrier according to the received mechanical characteristic regulation related data to complete the tumor cell culture environment regulation process.

Images