A cell subtype classification method based on the combined parameters of cell migration persistence

Abstract

The present invention relates to the technical fields of biology and the like, and in particular to a cell subtype classification method based on a cell migration persistence joint parameter, comprising: calculating the migration speed of a cell, and calculating its migration angular displacement based on the migration speed; determining the number of sampling intervals and the number of groups, drawing a statistical histogram based on the angular displacement, and forming a probability density distribution diagram; calculating the angular displacement distribution entropy that varies with the number of intervals based on the probability density distribution diagram; determining the initial angular displacement distribution entropy of the cell and calculating the stabilization time based on the cell angular displacement distribution entropy, and using the initial angular displacement distribution entropy and the stabilization time to form a cell joint parameter; obtaining the joint parameter of all cells in the same cell population, and drawing a two-dimensional kernel density map based on the joint parameter, and then obtaining the number of cell subtypes existing in the cell population based on the numerical distribution in the two-dimensional kernel density map; the method of the present invention has the characteristics of being simple, accurate and rapid, and can provide technical support for quantitative analysis of population cell dynamics.

Description

Technical Field

[0001] The present invention relates to the technical fields of biology, physics, statistics, information theory and the like, and in particular to a cell subtype classification method based on a cell migration persistence combined parameter. Background Art

[0002] Cell migration is essential for the normal development of tissues and organs, as well as many physiological and pathological processes, such as immune response, wound healing, angiogenesis, and cancer metastasis. The migration ability of cells can usually be quantified using both duration and migration speed, and there are currently a variety of analysis methods available for researchers to use.

[0003] A cell population includes one or more subtypes of cells. In this field, it is usually necessary to establish a dynamic model of the cells, analyze the migration ability of the cells through the dynamic model, and then determine the subtype of the cells based on the migration ability of the cells. For example, the above two parameters can be derived by fitting the mean square displacement. The method based on Bayesian inference can reveal the time-varying characteristics of persistence, and the method based on wavelet transform can also explore the characteristics of persistence. However, the existing analysis methods still have three main problems: first, they rely on the assumed dynamic model; second, they involve complex calculations and inferences; and third, they are easily affected by the correlation between cell migration speed and time. Therefore, it is very important to develop a simple, efficient, and model-independent analysis method for cell dynamics research. Summary of the invention

[0004] In order to determine the number of cell subtypes present in a cell population, the present invention proposes a cell subtype classification method based on cell migration persistence combined parameters, which specifically includes the following steps:

[0005] Get the two-dimensional trajectory coordinates r of a cell migration in a cell population i , calculate the cell migration velocity v i , and calculate its migration angular displacement θ based on the migration velocity i ;

[0006] According to the migration angle displacement θ i , determine the sampling interval number τ and the group number n, draw a statistical histogram based on angular displacement, and form a probability density distribution diagram;

[0007] Substitute the probability value in the probability density distribution diagram into the Shannon entropy expression to calculate the angular displacement distribution entropy that changes with the interval number τ;

[0008] Determine the initial angular displacement distribution entropy of the cell and calculate the stabilization time according to the angular displacement distribution entropy of the cell, and use the initial angular displacement distribution entropy and the stabilization time to jointly constitute the cell joint parameter;

[0009] The joint parameters of all cells in the same cell population are obtained, and a two-dimensional kernel density map is drawn according to the joint parameters. Then, the number of cell subtypes existing in the cell population is obtained according to the numerical distribution in the two-dimensional kernel density map.

[0010] Furthermore, the cell migration velocity v was calculated. i =(v x,i ,v y,i ) process includes:

[0011] v x,i =(x i -x i-1 ) / Δt

[0012] v y,i =(y i -y i-1 ) / Δt

[0013] Where Δt is the sampling time, x i and i represents the horizontal and vertical coordinates of the i-th sampling point on the two-dimensional trajectory of the cell, respectively. i-1 and i-1 Respectively represent the horizontal and vertical coordinates of the i-1th sampling point on the two-dimensional trajectory of the cell, v x,i and v y,i They represent the velocity components of the cell in the horizontal and vertical directions calculated based on the i-1th and ith sampling points, respectively.

[0014] Furthermore, the migration angular displacement θ is calculated based on the migration velocity. i When , first calculate the angular displacement θ according to the migration speed i The absolute value of is calculated as follows:

[0015]

[0016] Among them, |θ i | is the migration angular displacement θ calculated based on the migration velocity i The absolute value of cos -1 (·) represents the arccosine function, v i represents the cell velocity vector at the i-th sampling point, |v i | represents the modulus of the cell velocity vector at the i-th sampling point.

[0017] Furthermore, the angular displacement θ i The calculation includes:

[0018] Arbitrarily select a cell velocity vector v that is consistent with the cell at the i-1th sampling point i-1 The vertical direction vector;

[0019] Calculate the cell velocity vector v at the i-th sampling point i dot product with the direction vector;

[0020] If the dot product result is greater than 0, then θ i =|θ i |, otherwise θ i =-|θ i |.

[0021] Furthermore, the process of obtaining the probability density distribution map of the cells includes: determining a statistical histogram of the migration angular displacement when a given number of sampling intervals and a given number of groups are given, and obtaining a two-dimensional probability density distribution of the angular displacement by normalization.

[0022] Furthermore, the calculation process of the cell's angular displacement distribution entropy (EAD) includes:

[0023]

[0024] Among them, EAD[p(Δθ)] represents the angular displacement distribution entropy corresponding to the probability distribution p(Δθ) when the group interval is Δθ, and the group interval Δθ is expressed as p j (Δθ) represents the probability of the jth group.

[0025] Furthermore, the stabilization time is calculated according to the cell angular displacement distribution entropy, that is, the sampling point number corresponding to when the value of the angular displacement distribution entropy reaches the set threshold is obtained, and the product of the number and the sampling interval is used as the stabilization time.

[0026] Furthermore, according to the cell angular displacement distribution entropy, the entropy value when the sampling point sequence number is 1 is obtained, and the value is used as the initial angular displacement distribution entropy.

[0027] Furthermore, the initial angular displacement distribution entropy and the stabilization time are used together to form the cell joint parameter. The distribution of the values ​​in the figure is determined by drawing a two-dimensional kernel density map, and the number of cell subtypes existing in the cell population is obtained.

[0028] Furthermore, the angular displacement distribution entropy of the cell is used as a dynamic index, that is, if the angular displacement distribution entropy is closer to 0, the cell movement is closer to pure ballistic motion, and if the angular displacement distribution entropy is closer to 1, the cell movement is closer to Brownian motion.

[0029] Compared with the prior art, the present invention has the following beneficial effects:

[0030] (1) The present invention combines cell steering dynamics and Shannon entropy ideas to develop an angular displacement distribution entropy EAD method. Compared with other existing technologies (such as Bayesian inference, wavelet transform, least squares fitting, etc.), firstly, the present invention does not involve complex calculations and derivations, and can quickly output calculation results; secondly, since the migration angular displacement does not depend on the pre-assumed kinetic model, the problem that the kinetic model used cannot describe the actual cell movement is avoided; finally, the results obtained based on the present invention do not depend on the cell speed and time correlation, and therefore have strong unbiased characteristics and robustness.

[0031] (2) The angular displacement distribution entropy EAD proposed in the present invention corresponds to the Gibbs entropy in thermodynamics. The angular displacement distribution entropy can quantify the directional continuity of cell migration, and further characterize the degree of order or disorder of the migration movement. For example, an entropy value of 1 corresponds to pure Brownian motion, and 0 corresponds to pure ballistic motion. Therefore, this method can be used to intuitively evaluate the similarity between any cell movement and Brownian motion or ballistic motion. Through this invention, not only can the characteristics of cell migration patterns (such as the continuity of changes over time) be evaluated, but also the distribution and evolution of migration patterns in cell populations can be further determined. Therefore, it can provide important guidance for the study of collective cell migration (such as cancer cell invasion and metastasis). BRIEF DESCRIPTION OF THE DRAWINGS

[0032] Figure 1 It is a flow chart of a method for classifying cell subtypes based on cell migration persistence combined parameters of the present invention;

[0033] Figure 2 is a cell migration trajectory diagram of an example of the present invention;

[0034] Figure 3 is a graph of cell migration speed of an example of the present invention;

[0035] Figure 4 is a diagram of the angular displacement of cell migration according to an example of the present invention;

[0036] Figure 5 is a probability density distribution diagram of individual and overall averages of examples of the present invention;

[0037] Figure 6 is a graph showing the relationship between the individual and overall average angular displacement distribution entropy of the examples of the present invention and the change over time;

[0038] Figure 7 It is a two-dimensional kernel density map drawn based on joint parameters of an example of the present invention. DETAILED DESCRIPTION

[0039] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0040] The present invention proposes a cell subtype classification method based on cell migration persistence combined parameters, which specifically includes the following steps:

[0041] Get the two-dimensional trajectory coordinates r of a cell migration in a cell population i , calculate the cell migration velocity v i , and calculate its migration angular displacement θ based on the migration velocity i ;

[0042] According to the migration angle displacement θ i , determine the sampling interval number τ and the group number n, draw a statistical histogram based on angular displacement, and form a probability density distribution diagram;

[0043] Substitute the probability value in the probability density distribution diagram into the Shannon entropy expression to calculate the angular displacement distribution entropy that changes with the interval number τ;

[0044] Determine the initial angular displacement distribution entropy of the cell and calculate the stabilization time according to the angular displacement distribution entropy of the cell, and use the initial angular displacement distribution entropy and the stabilization time to jointly constitute the cell joint parameter;

[0045] The joint parameters of all cells in the same cell population are obtained, and a two-dimensional kernel density map is drawn according to the joint parameters. Then, the number of cell subtypes existing in the cell population is obtained according to the numerical distribution in the two-dimensional kernel density map.

[0046] Further, according to the migration coordinate r i Calculate the migration angle θ i The steps include:

[0047] ① Calculate the speed based on any two consecutive coordinates and get i-1 speeds;

[0048] ② Calculate the angular displacement based on any two consecutive velocities and obtain i-2 angular displacements;

[0049] ③ According to the deviation of the second speed relative to the first speed, the angular displacement θ i Mapped to the interval -π~π.

[0050] Furthermore, the migration velocity (v x,i ,v y,i ) is expressed as:

[0051] v x,i =(x i -x i-1 ) / Δt

[0052] v y,i =(y i -y i-1 ) / Δt

[0053] Among them, v x,i and v y,i represents the x and y components of the ith velocity, x i and i represents the coordinate component of the i-th position, x i-1 and i-1 represents the coordinate component of the i-1th position, and Δt is the sampling time.

[0054] Further, the migration angular displacement |θ is calculated i The formula for | is:

[0055]

[0056] where |θ i | is in the range of 0 to π, cos -1 Represents the inverse cosine function.

[0057] Furthermore, the above angular displacement |θ i |The method of mapping to the interval -π~π is:

[0058] ① Arbitrarily select a cell velocity vector v that is consistent with the cell at the i-1th sampling point i-1 The vertical direction vector is generally expressed as (v y,i ,-v x,i ) or (-v y,i ,v x,i );

[0059] ② Calculate the cell velocity vector v at the i-th sampling point i dot product with the direction vector;

[0060] ③If the result of the dot product is greater than zero, then let θ i =|θ i |, otherwise θ i =-|θ i |.

[0061] Further, according to the angular displacement θ i The steps for drawing a three-dimensional probability density distribution function (PDF) include:

[0062] ① Determine the value range of the number of groups n and the number of sampling intervals τ (1, 2, ..., N);

[0063] ② According to the number of groups n, draw the angular displacement statistical histogram with interval number τ = 1, and normalize it to obtain a two-dimensional PDF, where the angular displacement values ​​are θ1, θ2, θ3, …, θ i ;

[0064] ③ Cyclic change of the interval number τ, respectively calculate the two-dimensional PDF under each τ value. In this embodiment, the interval number τ refers to the sampling interval number. For example, τ=1 means sampling one point in every two points, τ=2 means sampling one point in every three points, and so on.

[0065] ④ Stack all two-dimensional PDFs in the direction of increasing interval number τ to form a three-dimensional PDF.

[0066] Furthermore, the formula for calculating EAD based on Shannon entropy is:

[0067]

[0068] Among them, EAD[p(Δθ)] represents the angular displacement distribution entropy corresponding to the probability distribution p(Δθ) when the class interval is Δθ, and the class interval Δθ of the statistical histogram is expressed as p j (Δθ) represents the probability of the jth group of the statistical histogram, log2(·) represents the logarithmic function with base 2, and log2(n) is the normalization factor. The function of the normalization factor is to eliminate the influence of the number of groups n.

[0069] Further, the steps of constructing joint parameters according to EAD include:

[0070] ① Draw the relationship curve between EAD and the interval number τ;

[0071] ② Extract the value of EAD when τ = 1, making it equal to the initial entropy value EAD i ;

[0072] ③ Extract the value of τ when the EAD curve reaches stability, so that (τ·Δt) is equal to the stable time TIME s In this embodiment, the stabilization time refers to the time when the EAD curve tends to be stable. Those skilled in the art can determine whether the EAD curve tends to be stable by setting a threshold value, for example, when the EAD value is greater than or equal to 0.9 to 0.99, those skilled in the art generally believe that it tends to be stable;

[0073] ④ Combine the initial entropy value and the stabilization time to construct the joint parameter (EAD i ,TIME s ).

[0074] Further, the stable time TIME s As the horizontal axis, the initial entropy value EAD i The ordinate is used to draw the corresponding two-dimensional nuclear density map, and then the cell subtype and dynamic characteristics are determined according to the nuclear density map.

[0075] This example uses Dictyostelium discoideum cell migration data published in a journal article as a data analysis example. In this example, the characteristics of the cell migration data are as follows:

[0076] (1.1) This data mainly studies the effect of the Arpin gene on the movement of Dictyostelium discoideum. This gene can transcribe Arpin protein, thereby weakening the directional persistence of migration; when this gene is knocked out, the directional persistence of Dictyostelium discoideum migration will be significantly improved;

[0077] (1.2) This data contains 38 cell trajectories, each trajectory consists of 163 points, the time interval between adjacent points is 1 minute, and each point represents the position of the cell nucleus on the two-dimensional plane.

[0078] The above-mentioned cell migration data can be adjusted by those skilled in the art according to actual needs. The present invention does not limit the source and type of the data, and the data can also be the movement data of microorganisms, active particles, etc.

[0079] Based on the above cell migration data, the velocity and angular displacement are calculated, and then the Shannon entropy is introduced to calculate the angular displacement distribution entropy and joint parameters to characterize the directional continuity and distribution of cell migration, and then determine the cell subtype and kinetic characteristics ( Figure 1 ). Specifically include the following steps:

[0080] (2.1) First, based on the above cell migration data, draw the cell migration trajectory (x i ,y i ), qualitative analysis of cell migration ( Figure 2 ); Secondly, the migration speed (v x,i ,v y,i ), the calculation formula is as follows:

[0081] v x,i =(x i -x i-1 ) / Δt

[0082] v y,i =(y i -y i-1 ) / Δt

[0083] where v x,i and v y,i represents the x and y components of the ith velocity, xi and i represents the coordinate component of the i-th position, x i-1 and i-1 represents the coordinate component of the i-1th position, and Δt is the sampling time.

[0084] After this, further use Calculate the velocity modulus and plot the velocity modulus versus time curve ( Figure 3 ).

[0085] (2.2) Based on the above migration speed (v x,i ,v y,i ), calculate the corresponding migration angular displacement θ i , the formula used is as follows:

[0086]

[0087] where |θ i | is in the range of 0 to π, cos -1 represents the arccosine function. Further transforming the above angular displacement |θ i |The method of mapping to the interval -π~π is:

[0088] ① Determine and arbitrarily select a cell velocity vector v at the i-1th sampling point i-1 The vertical direction vector (v y,i ,-v x,i ) or (-v y,i ,v x,i );

[0089] ② Calculate the cell velocity vector v at the i-th sampling point i dot product with the direction vector;

[0090] ③If the result of the dot product is greater than zero, then let θ i =|θ i |, otherwise θ i =-|θ i |.

[0091] The angular displacement θ is calculated i Afterwards, plot the angular displacement θ i The relationship curve with time ( Figure 4 ).

[0092] (2.3) Based on the angular displacement θ calculated above i , calculate the corresponding three-dimensional PDF, the main steps include:

[0093] ① Determine the number of groups n=10, and the value range of the sampling interval number τ in this embodiment is 1, 2, ..., 30;

[0094] ② Plot the histogram of the angular displacement statistics when the interval number τ = 1, and normalize it by dividing it by the total number of angular displacements in this case to obtain a two-dimensional PDF, where the angular displacements are θ1, θ2, θ3, ...;

[0095] ③Change the interval number τ and repeat step ② to obtain the two-dimensional PDF at each τ value, where the angular displacement is θ1, θ 1+τ ,θ 1+2τ , …;

[0096] ④Finally, all PDFs are stacked in the direction of increasing interval number τ to form a three-dimensional PDF ( Figure 5 a);

[0097] ⑤ Repeat steps ② to ④ above to calculate the 3D PDF of each cell respectively, and then obtain the overall average 3D PDF ( Figure 5 b).

[0098] (2.4) For the three-dimensional PDF of a single cell, the two-dimensional PDFs of τ = 1, 2, ... are substituted into the Shannon entropy formula in turn to obtain the EAD ( Figure 6 a), the calculation formula is as follows:

[0099]

[0100] Among them, EAD[p(Δθ)] represents the angular displacement distribution entropy corresponding to the probability distribution p(Δθ) when the group interval is Δθ, and the group interval Δθ of the statistical histogram is expressed as log2 represents a logarithmic function with base 2, and log2(n) represents a normalization factor, thereby eliminating the influence of the number of groups n;

[0101] Then repeat the above process to calculate the EAD of all cells and obtain the overall average EAD curve ( Figure 6 b).

[0102] (2.5) Based on the calculated EAD curve, extract the initial entropy value EAD i and stabilization time TIME s , and then construct the joint parameters. The specific method is as follows:

[0103] ① For the EAD curve of a single cell, extract the EAD value when τ = 1 to make it equal to the initial entropy value EAD i ;

[0104] ② Extract the value of τ when the EAD curve reaches stability for the first time (for example, the value of the angular displacement distribution entropy is ≥ 0.99 for the first time to determine that the EAD curve is stable) so that (τ·Δt) is equal to the stable time TIME s ;

[0105] ③ Combine the initial entropy value and the stabilization time to construct the joint parameter (EAD i ,TIME s ), and repeat the above process to obtain the joint parameter set of all cells.

[0106] (2.6) Based on the joint parameter set, the stable time TIME s As the horizontal axis, the initial entropy value EAD i As the vertical axis, draw a two-dimensional kernel density map ( Figure 7 ), and then analyze the directional continuity of cell migration and determine the types of cell subtypes and dynamic characteristics.

[0107] according to Figure 5 From the three-dimensional PDF shown in b, we can see that when the interval number τ is small, the two-dimensional PDF is mainly concentrated near 0rad, showing the characteristics of uneven distribution, that is, "high in the middle and low at both ends"; as the interval number τ gradually increases, the two-dimensional PDF will gradually change to a uniform distribution, resulting in similar heights in the middle and at both ends. The above uneven PDF indicates that cells are more inclined to migrate along the previous direction, resulting in angular displacement mainly concentrated near 0rad, while the uniform PDF indicates that cells are more likely to turn, resulting in angular displacement uniformly distributed in the range of -π to π. Therefore, based on the changes in the above PDF, it can be qualitatively judged that as the interval number τ increases, the direction of cell migration will gradually weaken.

[0108] The EAD proposed in the present invention quantifies the changes in the above PDF from the perspective of Shannon entropy. The main principle is: when the probability of a random event producing any result is equal, it represents the most disordered state (i.e., the result cannot be determined), and its entropy value is 1; and when the probabilities are all unequal, it represents the most ordered state (i.e., the result can be determined), and its entropy value is 0. Therefore, the smaller the EAD, the stronger the directional persistence, which can correspond to the most ordered dynamics, i.e., pure ballistic motion; conversely, the weaker the directional persistence, which can correspond to the most disordered dynamics, i.e., pure Brownian motion. Figure 6 The results shown in b can quantitatively determine that as the interval number τ increases, the directional persistence of cell migration will gradually weaken.

[0109] In order to further explore the changes and distribution of EAD, this embodiment further extracts joint parameters and plots Figure 7 The two-dimensional kernel density diagram is shown. The diagram shows that the embodiment of the present invention is mainly composed of three clusters, whose centers are located at (0.910, 0.907), (2.247, 0.879) and (4.218, 0.826), and their stable time TIME s Gradually increased from 0.910 to 4.218, while the initial entropy value EAD iGradually decreased from 0.907 to 0.826. Therefore, it can be seen that the third cluster in the figure corresponds to the largest directional persistence, the second cluster is second, and the first cluster is the smallest. Since the data analyzed in this embodiment are derived from the same cell population, the above three clusters can represent different motor subtypes, that is, the third cluster has the strongest motility, while the first cluster has the weakest motility. Through the analysis of the two-dimensional nuclear density map, we can reveal the existence of different cell subtypes in the same cell population, which to a certain extent explains the coexistence and role exchange of leading cells and follower cells in collective cell invasion.

[0110] As a specific implementation, the angular displacement distribution entropy and the two-dimensional kernel density map can be calculated based on the cell migration trajectory, and then the distribution characteristics of the directional persistence in the cell population can be analyzed to further determine the subtype and kinetic characteristics of the cell. In the specific implementation, the present invention does not limit the cell data, the number of groups, the number of intervals, etc., and those skilled in the art can adaptively adjust the various parameters involved in the present invention according to the actual cell types to be studied to obtain the best results.

[0111] The experimental results of this embodiment show the analysis of cell migration subtypes of Dictyostelium discoideum, that is, those skilled in the art can classify cell subtypes in a cell population based on the technical solution of the present application and perform corresponding kinetic analysis on the cells. The process of obtaining the analysis results based on the solution of the present invention is highly innovative, and in practical applications, the cell subtype and kinetic analysis method proposed in the present invention can be used to further study the kinetic characteristics of different types of cancer cells during migration, such as invasive breast cancer MDA-MB-231 cells and MCF-10A cells. The method of the present application can improve the accuracy and robustness of cell kinetic analysis.

[0112] Although embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions and variations may be made to the embodiments without departing from the principles and spirit of the present invention, and that the scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. A cell subtype classification method based on cell migration persistence combined parameters, characterized in that: The specific steps include: Get the two-dimensional trajectory coordinates r of a cell migration in a cell population i , calculate the cell migration velocity v i , and calculate its migration angular displacement θ based on the migration velocity i ; According to the migration angle displacement θ i , determine the sampling interval number τ and the group number n, draw a statistical histogram based on angular displacement, and form a probability density distribution diagram; Substitute the probability value in the probability density distribution diagram into the Shannon entropy expression to calculate the angular displacement distribution entropy that changes with the interval number τ, including: Determine the value range of the number of groups n and the number of sampling intervals τ; According to the number of groups n, a statistical histogram of the angular displacement with an interval number τ = 1 is drawn and normalized to obtain a two-dimensional probability density distribution diagram, where the angular displacement values ​​are θ1, θ2, θ3, …, θ i ; The interval number τ is changed cyclically, and the two-dimensional probability density distribution graph under each τ value is calculated respectively, and all the two-dimensional probability density distribution graphs are stacked along the direction of increasing interval number τ, thereby forming a three-dimensional probability density distribution graph; Determine the initial angular displacement distribution entropy of the cell and calculate the stabilization time according to the angular displacement distribution entropy of the cell, and use the initial angular displacement distribution entropy and the stabilization time to jointly constitute the cell joint parameter. The initial angular displacement distribution entropy of the cell is the entropy value when the sampling point sequence number is 1, that is, the entropy value when the sampling point sequence number is 1 is used as the initial angular displacement distribution entropy, and the stabilization time is the product of the number of intervals when the angular displacement distribution entropy of the cell is greater than or equal to the set threshold for the first time and the sampling interval; The calculation process of the cell's angular displacement distribution entropy includes: Among them, EAD[p(Δθ)] represents the angular displacement distribution entropy corresponding to the probability distribution p(Δθ) when the group interval is Δθ, and the group interval Δθ of the statistical histogram is expressed as pj(Δθ) represents the probability of the jth group; The joint parameters of all cells in the same cell population are obtained, and a two-dimensional kernel density map is drawn according to the joint parameters. Then, the number of cell subtypes existing in the cell population is obtained according to the numerical distribution in the two-dimensional kernel density map.

2. A cell subtype classification method based on cell migration persistence combined parameters according to claim 1, characterized in that: Calculate the cell migration velocity v i =(v x，i , v y，i ) include: v x，i =(x i -x i-1 ) / Δt v y,i =(y i -y i-1 ) / Δt Where Δt is the sampling time, x i and i represents the horizontal and vertical coordinates of the i-th sampling point on the two-dimensional trajectory of the cell, respectively. i-1 and i-1 Respectively represent the horizontal and vertical coordinates of the i-1th sampling point on the two-dimensional trajectory of the cell, v x，i and v y，i They represent the velocity components of the cell in the horizontal and vertical directions calculated based on the i-1th and ith sampling points, respectively.

3. A cell subtype classification method based on cell migration persistence combined parameters according to claim 2, characterized in that: Calculate the migration angle θ based on the migration velocity i When , first calculate the angular displacement θ according to the migration speed i The absolute value of include: Among them, |θ i | is the migration angular displacement θ calculated based on the migration velocity i The absolute value of cos -1 (·) represents the arccosine function, v i represents the cell velocity vector at the i-th sampling point, |v i | represents the modulus of the cell velocity vector at the i-th sampling point.

4. A cell subtype classification method based on cell migration persistence combined parameters according to claim 3, characterized in that: Angular displacement θ i The calculation includes: Arbitrarily select a cell velocity vector v that is consistent with the cell at the i-1th sampling point i-1 The vertical direction vector; Calculate the cell velocity vector v at the i-th sampling point i dot product with the direction vector; If the dot product result is greater than 0, then θ i =|θ i |, otherwise θ i =-|θ i |.

5. The method for classifying cell subtypes based on cell migration persistence combined parameters according to claim 1, characterized in that: The angular displacement distribution entropy of the cell is used as a dynamic index, that is, the closer the angular displacement distribution entropy is to 0, the closer the cell movement is to pure ballistic motion, and the closer the angular displacement distribution entropy is to 1, the closer the cell movement is to Brownian motion.

Images