Activity analysis system of stem cell exosome uric acid reducing preparation

Abstract

The invention relates to the technical field of biomedicine optimization, and discloses an activity analysis system of a stem cell exosome uric acid reducing preparation, comprising: a data acquisition module for acquiring characteristic data of a uric acid reducing preparation in K experiments and representing the characteristic data through a first characteristic sequence; the data preprocessing module is used for preprocessing the first feature sequence to obtain a second feature sequence; the uric acid reduction effect prediction module is used for constructing a uric acid reduction effect prediction model according to the second feature sequence and outputting a decline value of uric acid; and the parameter combination generation module is used for generating an optimal parameter combination according to the improved hedera helix optimization algorithm in combination with the uric acid reduction effect prediction model. According to the invention, the uric acid reduction value is accurately predicted through the uric acid reduction effect prediction model, and the balance of curative effect maximization and side effect minimization is realized in combination with the improved hedera helix optimization algorithm.

Description

Technical Field

[0001] The present invention belongs to the technical field of biomedicine optimization, and specifically relates to an activity analysis system for a stem cell exosome uric acid-lowering preparation. Background Art

[0002] Although traditional uric acid-lowering drugs have certain therapeutic effects, they have problems such as large side effects and poor individualized effects. More accurate and safe new treatment methods are urgently needed. With the rapid development of stem cell exosome technology, studies have found that exosomes are rich in bioactive substances (such as RNA, proteins, etc.), which can regulate inflammatory responses and metabolic pathways, providing a new idea for uric acid-lowering treatment. However, the current development of exosome preparations faces multiple challenges, including complex optimization of parameters such as exosome sources, separation methods, component ratios and concentrations. Traditional experimental design methods are time-consuming and inefficient, and it is difficult to achieve global optimization in high-dimensional parameter space. In addition, the lack of intelligent efficacy prediction and parameter control methods makes it difficult to improve efficacy while reducing costs. Summary of the invention

[0003] The present invention provides an activity analysis system for a stem cell exosome uric acid-lowering preparation, which solves the technical problems of low efficiency and complex parameter optimization in related technologies.

[0004] The present invention provides an activity analysis system for a stem cell exosome uric acid-lowering preparation, comprising:

[0005] A data acquisition module is used to obtain characteristic data of uric acid-lowering preparations in K experiments and represent them through a first characteristic sequence. The characteristic data of uric acid-lowering preparations include: exosome concentration, RNA component ratio, protein component ratio, stem cell source and separation method, where K is a custom parameter;

[0006] The first feature sequence includes K units, and the i-th unit represents the characteristic data of the uric acid-lowering preparation obtained in the i-th experiment, where 1≤i≤K;

[0007] A data preprocessing module, used for preprocessing the first feature sequence to obtain a second feature sequence;

[0008] A uric acid-lowering effect prediction module is used to construct a uric acid-lowering effect prediction model based on the second feature sequence. The uric acid-lowering effect prediction model is constructed based on an improved long short-term memory network. The input of the model is the second feature sequence, and the output is the uric acid reduction value in the g-th experiment, where g is a custom parameter.

[0009] The parameter combination generation module is used to generate the optimal parameter combination based on the improved ivy optimization algorithm combined with the uric acid-lowering effect prediction model. The optimal parameter combination includes: exosome concentration, RNA component ratio, protein component ratio, stem cell source and separation method.

[0010] Furthermore, stem cell sources include: adipose-derived stem cells and bone marrow-derived stem cells, and the stem cell sources are represented by real number codes; separation methods include: ultracentrifugation, ultrafiltration and immunoaffinity separation, and the separation methods are represented by one-hot codes.

[0011] Further, the first feature sequence is preprocessed to obtain a second feature sequence, and the specific steps include:

[0012] Step S201, for missing values ​​in the first feature sequence, use the average value of data in the same dimension of the first feature sequence except for the missing values ​​to perform filling processing;

[0013] Step S202, extracting the RNA and protein ratio interaction feature according to the first feature sequence, the calculation formula of the RNA and protein ratio interaction feature is: ,in, It represents the interactive characteristics of RNA and protein ratios. R represents the RNA component ratio, ranging from 0 to 1. P represents the protein component ratio, ranging from 0 to 1. and Respectively represent the RNA component weight coefficient and the protein component weight coefficient, which are used to control the basic influence of RNA and protein components, respectively. and Represent the RNA square term weight coefficient and protein square term weight coefficient, respectively. represents the interaction inhibition coefficient, It represents the difference penalty coefficient, which is used to measure the impact of RNA and protein imbalance on the overall effect. exp represents the exponential function, and log represents the logarithmic function. It represents the absolute value of the difference between RNA component and protein component. The interactive feature of RNA and protein ratio takes into account the synergistic effect of RNA component ratio and protein component ratio.

[0014] Step S203, extracting the fluctuation characteristics of the ratio of RNA to protein according to the first characteristic sequence, the calculation formula of the fluctuation characteristics of the ratio of RNA to protein is: ,in, Indicates the fluctuation characteristics of the ratio of RNA to protein, and They represent the proportion of RNA components and protein components in the i-th experiment, respectively, and i represents the index of the number of experiments;

[0015] Step S204: extracting concentration synergistic correction features according to the first feature sequence. The calculation formula of the concentration synergistic correction features is: ,in, represents the concentration synergistic correction feature, represents the hyperbolic tangent function, represents the concentration correction factor, represents the exosome concentration;

[0016] Step S205, using a z-score normalization method to normalize the processed first feature sequence to obtain a second feature sequence.

[0017] Furthermore, the structure of the uric acid-lowering effect prediction model includes: an input layer, an improved long short-term memory network layer, an attention mechanism layer, and an output layer;

[0018] The input layer is used to input the second feature sequence;

[0019] The improved long short-term memory network layer is used to capture the dependency between the second feature sequence and the uric acid drop value;

[0020] The attention mechanism layer is used to dynamically assign weights of the features of the second feature sequence to the uric acid decrease value;

[0021] The output layer is used to output the uric acid decrease value.

[0022] Furthermore, the structure of the improved long short-term memory network layer includes: an input gate, a forget gate, an output gate, a candidate memory unit, a memory unit update and a hidden state update, and the calculation formula of the improved long short-term memory network layer includes:

[0023] The calculation formula of the input gate is: ;

[0024] The calculation formula of the forget gate is: ;

[0025] The calculation formula of the output gate is: ;

[0026] The calculation formula for candidate memory cells is: ;

[0027] The calculation formula for memory unit update is: ;

[0028] The calculation formula for hidden state update is: ;

[0029] in, , , , , and Represent the input gate, forget gate, output gate, candidate memory unit, memory unit update and hidden state update of the i-th unit respectively. represents the i-th unit of the second feature sequence, and Represent the hidden states of the outputs of the i-th unit and the i-1-th unit, respectively. , and Respectively represent the first weight parameter, the first bias parameter and the second weight parameter of the input gate, Indicates that the i-th unit and the hidden state of the output of the i-1th unit To splice, represents the memory cell update of the i-1th unit, , and Respectively represent the first weight parameter, the first bias parameter and the second weight parameter of the forget gate, , and Respectively represent the first weight parameter, the first bias parameter and the second weight parameter of the output gate, and They represent the first weight parameter and the first bias parameter of the candidate memory unit respectively, sigmoid represents the sigmoid activation function, and tanh represents the hyperbolic tangent activation function.

[0030] Furthermore, the calculation formula of the attention mechanism layer is: ;

[0031] in, represents the weighted hidden state of each unit and the unit's feature weight, , Respectively represent the first weight parameter and the first bias parameter of each unit, , and They respectively represent the first weight parameter, the second weight parameter and the first bias parameter of the attention mechanism within the unit, and i represents the unit index of the second feature sequence.

[0032] Furthermore, the optimal parameter combination is generated according to the improved ivy optimization algorithm combined with the uric acid-lowering effect prediction model. The specific steps include:

[0033] Step S301, generating an initialization population with Q individuals meeting the constraint conditions through Latin hypercube sampling, where Q is a custom parameter;

[0034] The individual positions of the initialized population are represented by the uric acid-lowering feature vector, and the encoding format of the uric acid-lowering feature vector is: ,in, represents the uric acid lowering feature vector, indicates the source of stem cells, and A indicates the analysis method;

[0035] Step S302, constructing an objective function according to the uric acid-lowering effect prediction model;

[0036] Step S303, calculating the growth rate change of each individual according to the growth rate formula, wherein the growth rate formula is: ,in, and represents the change in growth rate of individual j at the k+1th and kth iterations, k is the index of the iteration number, j and a are the indexes of the individuals, represents the dynamic enhancement coefficient, and represents the growth rate of individuals a and j at the kth iteration, exp represents the exponential function, represents the growth rate regulation coefficient, represents the maximum growth rate of individual j, represents the time attenuation coefficient, represents the square of a uniformly distributed random number, represents a random vector that obeys a standard normal distribution, and the dimension of the solution space is D. Represents element-wise multiplication;

[0037] Step S304, update the position of each individual according to the climbing formula and the fitness of the individual obtained by the objective function. The climbing formula is: ,in, represents the updated position of individual j at the kth iteration, represents the position of individual j at the kth iteration, represents the position of the individual with the highest fitness among the neighbors of individual j, represents the climb enhancement factor, represents the position of the individual with the highest fitness in the population at the kth iteration, Represents the cubic value of a uniformly distributed random number, represents the nonlinear amplification factor;

[0038] Step S305, calculate the fitness of each individual according to the objective function, and select and update the individuals through the update mechanism. The formula of the update mechanism is: ,in, represents the individual position after the individual is processed by the update mechanism, represents a D-dimensional uniformly distributed random number vector, represents the dynamic threshold parameter, and represents the fitness of individual j at the kth iteration and the fitness of the individual with the highest fitness in the population at the kth iteration, Represents the direction control coefficient, which controls the step size of the individual when updating the direction;

[0039] Step S306, the current number of iterations is increased by 1. When the current number of iterations reaches the maximum number of iterations, the individual with the highest current fitness is output as the optimal parameter combination. Otherwise, return to step S303, where the maximum number of iterations is a custom parameter.

[0040] Furthermore, the specific steps of step S301 include:

[0041] Step S3011, determining the upper and lower bounds of the individual solution space, that is, the maximum and minimum values ​​of each dimension in the individual's uric acid-lowering feature vector;

[0042] Step S3012, divide the solution space into isometric subspaces, randomly select a value in each subspace to form a D-dimensional coordinate;

[0043] Step S3013, randomly arrange between different dimensions to form a Latin hypercube sampling matrix. The Latin hypercube sampling matrix is ​​a A matrix with each row corresponding to a sample point and each column corresponding to a sampling value of a dimension, ranging from 0 to 1;

[0044] Step S3014, mapping each row of the Latin hypercube sampling matrix to the solution space according to a mapping formula as an individual of the initial population, wherein the mapping formula is: ,in, represents the individual position a of the initialized population, and They represent the upper and lower bounds of the solution space, represents the i-th row of the Latin hypercube sampling matrix.

[0045] Furthermore, the calculation formula of the objective function is: ;

[0046] in, represents the fitness of an individual, represents the initial efficacy weight coefficient, represents the therapeutic effect weight step length, It indicates the uric acid reduction value obtained by the individual according to the uric acid reduction effect prediction model. represents the side effect weight coefficient, represents the side effect weight step size, , and Respectively represent the weight coefficients of exosome concentration, RNA component ratio and protein component ratio for the total cost.

[0047] The beneficial effects of the present invention are as follows: the present invention uses an improved long short-term memory network layer to enable the model to efficiently capture the nonlinear relationship between complex features such as exosome concentration, RNA and protein ratio and uric acid reduction value, especially the dependency between features, and provide reliable prediction support for optimizing exosome preparations;

[0048] The attention mechanism layer of the present invention dynamically allocates input feature weights, highlights the influence of key parameters on the decrease in uric acid value, provides quantitative results of feature contribution, and provides a scientific basis for adjusting the combination of formulation parameters;

[0049] The present invention achieves a balance between maximizing efficacy and minimizing side effects by accurately predicting the decrease in uric acid through a prediction model, and combining it with an improved ivy optimization algorithm. The prediction model provides accurate efficacy evaluation, and the optimization algorithm explores the optimal parameter combination based on the prediction results. BRIEF DESCRIPTION OF THE DRAWINGS

[0050] Figure 1 It is a schematic diagram of a module of the activity analysis system of the stem cell exosome uric acid-lowering preparation of the present invention;

[0051] Figure 2 is a flow chart of preprocessing the first feature sequence of the present invention;

[0052] Figure 3 It is a flow chart of the improved ivy optimization algorithm of the present invention.

[0053] In the figure: a data acquisition module 101, a data preprocessing module 102, a uric acid lowering effect prediction module 103 and a parameter combination generation module 104. DETAILED DESCRIPTION

[0054] The subject matter described herein will now be discussed with reference to example embodiments. It should be understood that the discussion of these embodiments is only to enable those skilled in the art to better understand and implement the subject matter described herein, and the functions and arrangements of the elements discussed may be changed without departing from the scope of protection of the contents of this specification. Each example may omit, replace or add various processes or components as needed. In addition, the features described relative to some examples may also be combined in other examples.

[0055] like Figure 1-Figure 3 As shown, the activity analysis system of stem cell exosome uric acid lowering preparation includes:

[0056] The data acquisition module 101 is used to obtain characteristic data of the uric acid-lowering preparation in K experiments and represent it through a first characteristic sequence, wherein the characteristic data of the uric acid-lowering preparation includes: exosome concentration, RNA component ratio, protein component ratio, stem cell source and separation method, wherein K is a custom parameter;

[0057] The first feature sequence includes K units, and the i-th unit represents the characteristic data of the uric acid-lowering preparation obtained in the i-th experiment, where 1≤i≤K;

[0058] A data preprocessing module 102 is used to preprocess the first feature sequence to obtain a second feature sequence;

[0059] The uric acid-lowering effect prediction module 103 is used to construct a uric acid-lowering effect prediction model based on the second feature sequence. The uric acid-lowering effect prediction model is constructed based on an improved long short-term memory network. The input of the model is the second feature sequence, and the output is the uric acid reduction value in the g-th experiment, where g is a custom parameter.

[0060] The parameter combination generation module 104 is used to generate an optimal parameter combination according to the improved ivy optimization algorithm combined with the uric acid-lowering effect prediction model, and the optimal parameter combination includes: exosome concentration, RNA component ratio, protein component ratio, stem cell source and separation method.

[0061] In one embodiment of the present invention, the exosome concentration is determined using BCA protein concentration, the RNA-seq high-throughput sequencing method is used to analyze the ratio of RNA components in the exosomes, the mass spectrometry analysis method is used to measure the protein content, and the protein component ratio is calculated, the stem cell type and isolation method in the experiment are recorded, and the uric acid concentration is determined using the uricase method in an in vitro experiment to obtain the uric acid concentration decrease value.

[0062] In one embodiment of the present invention, the sources of stem cells include: adipose-derived stem cells and bone marrow-derived stem cells, and the sources of stem cells are represented by real number codes; the separation methods include: ultracentrifugation, ultrafiltration and immunoaffinity separation, and the separation methods are represented by one-hot codes.

[0063] In one embodiment of the present invention, the first feature sequence is preprocessed to obtain the second feature sequence, and the specific steps include:

[0064] Step S201, for missing values ​​in the first feature sequence, use the average value of data in the same dimension of the first feature sequence except for the missing values ​​to perform filling processing;

[0065] Step S202, extracting the RNA and protein ratio interaction feature according to the first feature sequence, the calculation formula of the RNA and protein ratio interaction feature is: ,in, It represents the interactive characteristics of RNA and protein ratios. R represents the RNA component ratio, ranging from 0 to 1. P represents the protein component ratio, ranging from 0 to 1. and Respectively represent the RNA component weight coefficient and the protein component weight coefficient, which are used to control the basic influence of RNA and protein components, respectively. and Represent the RNA square term weight coefficient and protein square term weight coefficient, respectively. represents the interaction inhibition coefficient, It represents the difference penalty coefficient, which is used to measure the impact of RNA and protein imbalance on the overall effect. exp represents the exponential function, and log represents the logarithmic function. It represents the absolute value of the difference between RNA component and protein component. The interactive feature of RNA and protein ratio takes into account the synergistic effect of RNA component ratio and protein component ratio.

[0066] Step S203, extracting the fluctuation characteristics of the ratio of RNA to protein according to the first characteristic sequence, the calculation formula of the fluctuation characteristics of the ratio of RNA to protein is: ,in, Indicates the fluctuation characteristics of the ratio of RNA to protein, and They represent the proportion of RNA components and protein components in the i-th experiment, respectively, and i represents the index of the number of experiments;

[0067] Step S204: extracting concentration synergistic correction features according to the first feature sequence. The calculation formula of the concentration synergistic correction features is: ,in, represents the concentration synergistic correction feature, represents the hyperbolic tangent function, represents the concentration correction factor, represents the exosome concentration;

[0068] Step S205, using a z-score normalization method to normalize the processed first feature sequence to obtain a second feature sequence.

[0069] In one embodiment of the present invention, the structure of the uric acid-lowering effect prediction model includes: an input layer, an improved long short-term memory network layer, an attention mechanism layer, and an output layer;

[0070] The input layer is used to input the second feature sequence;

[0071] The improved long short-term memory network layer is used to capture the dependency between the second feature sequence and the uric acid drop value;

[0072] The attention mechanism layer is used to dynamically assign weights of the features of the second feature sequence to the uric acid decrease value;

[0073] The output layer is used to output the uric acid decrease value.

[0074] In one embodiment of the present invention, the structure of the improved long short-term memory network layer includes: an input gate, a forget gate, an output gate, a candidate memory unit, a memory unit update, and a hidden state update. The calculation formula of the improved long short-term memory network layer includes:

[0075] The calculation formula of the input gate is: ;

[0076] The calculation formula of the forget gate is: ;

[0077] The calculation formula of the output gate is: ;

[0078] The calculation formula for candidate memory cells is: ;

[0079] The calculation formula for memory unit update is: ;

[0080] The calculation formula for hidden state update is: ;

[0081] in, , , , , and Represent the input gate, forget gate, output gate, candidate memory unit, memory unit update and hidden state update of the i-th unit respectively. represents the i-th unit of the second feature sequence, and Represent the hidden states of the outputs of the i-th unit and the i-1-th unit, respectively. , and Respectively represent the first weight parameter, the first bias parameter and the second weight parameter of the input gate, The enhancement weight representing the memory cell state of the previous unit is used to enhance the sensitivity of the current unit to long-term memory. Indicates that the i-th unit and the hidden state of the output of the i-1th unit To splice, represents the memory cell update of the i-1th unit, , and Respectively represent the first weight parameter, the first bias parameter and the second weight parameter of the forget gate, Used to emphasize the nonlinear interaction between the current input and the hidden state. , and Respectively represent the first weight parameter, the first bias parameter and the second weight parameter of the output gate, The input gate and memory cell state are combined to dynamically weight the output. and They represent the first weight parameter and the first bias parameter of the candidate memory unit respectively, sigmoid represents the sigmoid activation function, and tanh represents the hyperbolic tangent activation function.

[0082] In one embodiment of the present invention, the improved long short-term memory network layer can more accurately capture the effects of different exosome concentrations, RNA and protein ratios on uric acid reduction by optimizing the gating mechanism, thereby enhancing the adaptability of the uric acid-lowering effect prediction model to the nonlinear relationship between complex multidimensional features.

[0083] In one embodiment of the present invention, the calculation formula of the attention mechanism layer is: ;

[0084] in, represents the weighted hidden state of each unit and the unit's feature weight, , Respectively represent the first weight parameter and the first bias parameter of each unit, , and They represent the first weight parameter, the second weight parameter and the first bias parameter of the attention mechanism within the unit, i represents the unit index of the second feature sequence, Used to indicate the importance of each unit to the final output. Indicates the importance of each feature within a specific unit.

[0085] In one embodiment of the present invention, the attention mechanism layer dynamically adjusts the weights according to the importance of the input features, highlights the features that have a great impact on the decrease in uric acid values, and directly quantifies the contribution of the features, providing a scientific basis for optimizing the formulation parameters.

[0086] In one embodiment of the present invention, the ivy optimization algorithm is derived from the growth pattern of the ivy plant. The algorithm simulates the orderly coordinated growth of the ivy population and the diffusion evolution process of the ivy plant. The algorithm uses the knowledge of nearby ivy plants to determine the growth direction and imitates the behavior of the ivy plant by selecting the closest and most important neighbors for self-improvement.

[0087] In one embodiment of the present invention, the optimal parameter combination is generated according to the improved ivy optimization algorithm combined with the uric acid-lowering effect prediction model, and the specific steps include:

[0088] Step S301, generating an initialization population with Q individuals meeting the constraint conditions through Latin hypercube sampling, where Q is a custom parameter;

[0089] Constraints include: each variable in the characteristic data of uric acid-lowering preparations must meet physical or experimental condition constraints, for example, the sum of the RNA component ratio and the protein component ratio must not exceed 1;

[0090] The individual positions of the initialized population are represented by the uric acid-lowering feature vector, and the encoding format of the uric acid-lowering feature vector is: ,in, represents the uric acid lowering feature vector, indicates the source of stem cells, and A indicates the analysis method;

[0091] Step S302, constructing an objective function according to the uric acid-lowering effect prediction model;

[0092] Step S303, calculating the growth rate change of each individual according to the growth rate formula, wherein the growth rate formula is: ,in, and represents the change in growth rate of individual j at the k+1th and kth iterations, k is the index of the iteration number, j and a are the indexes of the individuals, represents the dynamic enhancement coefficient, and represents the growth rate of individuals a and j at the kth iteration, exp represents the exponential function, represents the growth rate regulation coefficient, represents the maximum growth rate of individual j, represents the time attenuation coefficient, represents the square of a uniformly distributed random number, represents a random vector that obeys a standard normal distribution, and the dimension of the solution space is D. Represents element-wise multiplication;

[0093] Step S304, update the position of each individual according to the climbing formula and the fitness of the individual obtained by the objective function. The climbing formula is: ,in, represents the updated position of individual j at the kth iteration, represents the position of individual j at the kth iteration, represents the position of the individual with the highest fitness among the neighbors of individual j, represents the climb enhancement factor, represents the position of the individual with the highest fitness in the population at the kth iteration, Represents the cubic value of a uniformly distributed random number, represents the nonlinear amplification coefficient. This step updates the individual position through the climbing formula to make the individual approach the neighborhood optimal solution and the global optimal solution, and introduces random disturbance to enhance the search diversity.

[0094] Step S305, calculate the fitness of each individual according to the objective function, and select and update the individuals through the update mechanism. The formula of the update mechanism is: ,in, represents the individual position after the individual is processed by the update mechanism, represents a D-dimensional uniformly distributed random number vector, represents the dynamic threshold parameter, and represents the fitness of individual j at the kth iteration and the fitness of the individual with the highest fitness in the population at the kth iteration, Represents the direction control coefficient, which controls the step size of the individual when updating the direction. This step is used to screen out better individuals through dynamic threshold parameters and update them. For other individuals with poor fitness, direction control is used to move them closer to the optimal solution to avoid population degradation;

[0095] Step S306, the current number of iterations is increased by 1. When the current number of iterations reaches the maximum number of iterations, the individual with the highest current fitness is output as the optimal parameter combination. Otherwise, return to step S303, where the maximum number of iterations is a custom parameter.

[0096] In one embodiment of the present invention, the specific steps of step S301 include:

[0097] Step S3011, determining the upper and lower bounds of the solution space, that is, the maximum and minimum values ​​of each dimension in the individual's uric acid-lowering feature vector;

[0098] Step S3012, divide the solution space into isometric subspaces, randomly select a value in each subspace to form a D-dimensional coordinate;

[0099] Step S3013, randomly arrange between different dimensions to form a Latin hypercube sampling matrix. The Latin hypercube sampling matrix is ​​a A matrix with each row corresponding to a sample point and each column corresponding to a sampling value of a dimension, ranging from 0 to 1;

[0100] Step S3014, mapping each row of the Latin hypercube sampling matrix to the solution space according to a mapping formula as an individual of the initial population, wherein the mapping formula is: ,in, represents the individual position a of the initialized population, and They represent the upper and lower bounds of the solution space, represents the i-th row of the Latin hypercube sampling matrix.

[0101] In one embodiment of the present invention, the calculation formula of the objective function is: ;

[0102] in, represents the fitness of an individual, represents the initial efficacy weight coefficient, represents the therapeutic effect weight step length, It indicates the uric acid reduction value obtained by the individual according to the uric acid reduction effect prediction model. represents the side effect weight coefficient, represents the side effect weight step size, , and Respectively represent the weight coefficients of exosome concentration, RNA component ratio and protein component ratio for the total cost. Indicates the efficacy part, It represents the side effect part, and in the early stage of iteration, priority is given to reducing side effects and improving population quality. As the number of iterations increases, the efficacy part increases and the side effects decrease, gradually shifting to a balance between efficacy and side effects. In the later stage of iteration, the efficacy is mainly used to enhance the effect of lowering uric acid values.

[0103] The above describes an embodiment of the present embodiment, but the present embodiment is not limited to the above-mentioned specific implementation methods. The above-mentioned specific implementation methods are merely illustrative and not restrictive. Under the guidance of the present embodiment, ordinary technicians in this field can also make many forms, all of which are within the protection of the present embodiment.

Claims

1. An activity analysis system for stem cell exosome uric acid-lowering preparations, characterized in that: include: A data acquisition module is used to obtain characteristic data of uric acid-lowering preparations in K experiments and represent them through a first characteristic sequence. The characteristic data of uric acid-lowering preparations include: exosome concentration, RNA component ratio, protein component ratio, stem cell source and separation method, where K is a custom parameter; The first feature sequence includes K units, and the i-th unit represents the characteristic data of the uric acid-lowering preparation obtained in the i-th experiment, where 1≤i≤K; A data preprocessing module, used for preprocessing the first feature sequence to obtain a second feature sequence; A uric acid-lowering effect prediction module is used to construct a uric acid-lowering effect prediction model based on the second feature sequence. The uric acid-lowering effect prediction model is constructed based on an improved long short-term memory network. The input of the model is the second feature sequence, and the output is the uric acid reduction value in the g-th experiment, where g is a custom parameter. The parameter combination generation module is used to generate the optimal parameter combination based on the improved ivy optimization algorithm combined with the uric acid-lowering effect prediction model. The optimal parameter combination includes: exosome concentration, RNA component ratio, protein component ratio, stem cell source and separation method.

2. The activity analysis system of the stem cell exosome uric acid-lowering preparation according to claim 1, characterized in that: The sources of stem cells include: adipose-derived stem cells and bone marrow-derived stem cells, and the sources of stem cells are represented by real number codes; the separation methods include: ultracentrifugation, ultrafiltration and immunoaffinity separation, and the separation methods are represented by one-hot codes.

3. The activity analysis system of the stem cell exosome uric acid-lowering preparation according to claim 1, characterized in that: Preprocessing the first feature sequence to obtain a second feature sequence includes: Step S201, for missing values ​​in the first feature sequence, use the average value of data in the same dimension of the first feature sequence except for the missing values ​​to perform filling processing; Step S202, extracting the RNA and protein ratio interaction feature according to the first feature sequence, the calculation formula of the RNA and protein ratio interaction feature is: ,in, It represents the interactive characteristics of RNA and protein ratios. R represents the RNA component ratio, ranging from 0 to 1. P represents the protein component ratio, ranging from 0 to 1. and Respectively represent the RNA component weight coefficient and the protein component weight coefficient, which are used to control the basic influence of RNA and protein components, respectively. and Respectively represent the RNA square term weight coefficient and protein square term weight coefficient, represents the interaction inhibition coefficient, It represents the difference penalty coefficient, which is used to measure the impact of RNA and protein imbalance on the overall effect. exp represents the exponential function, and log represents the logarithmic function. It represents the absolute value of the difference between RNA component and protein component. The interactive feature of RNA and protein ratio takes into account the synergistic effect of RNA component ratio and protein component ratio. Step S203, extracting the fluctuation characteristics of the ratio of RNA to protein according to the first characteristic sequence, the calculation formula of the fluctuation characteristics of the ratio of RNA to protein is: ,in, Indicates the fluctuation characteristics of the ratio of RNA to protein, and They represent the proportion of RNA components and protein components in the i-th experiment, respectively, and i represents the index of the number of experiments; Step S204: extracting concentration synergistic correction features according to the first feature sequence. The calculation formula of the concentration synergistic correction features is: ,in, represents the concentration synergistic correction feature, represents the hyperbolic tangent function, represents the concentration correction factor, represents the exosome concentration; Step S205, using a z-score normalization method to normalize the processed first feature sequence to obtain a second feature sequence.

4. The activity analysis system of the stem cell exosome uric acid lowering preparation according to claim 1, characterized in that: The structure of the uric acid-lowering effect prediction model includes: input layer, improved long short-term memory network layer, attention mechanism layer, and output layer; The input layer is used to input the second feature sequence; The improved long short-term memory network layer is used to capture the dependency between the second feature sequence and the uric acid drop value; The attention mechanism layer is used to dynamically assign weights of the features of the second feature sequence to the uric acid decrease value; The output layer is used to output the uric acid decrease value.

5. The activity analysis system of the stem cell exosome uric acid lowering preparation according to claim 4, characterized in that: The structure of the improved long short-term memory network layer includes: input gate, forget gate, output gate, candidate memory unit, memory unit update and hidden state update. The calculation formula of the improved long short-term memory network layer includes: The calculation formula of the input gate is: ; The calculation formula of the forget gate is: ; The calculation formula of the output gate is: ; The calculation formula for candidate memory cells is: ; The calculation formula for memory unit update is: ; The calculation formula for hidden state update is: ; in, , , , , and They represent the input gate, forget gate, output gate, candidate memory unit, memory unit update, and hidden state update of the i-th unit, respectively. represents the i-th unit of the second feature sequence, and Represent the hidden states of the outputs of the i-th unit and the i-1-th unit, respectively. , and Respectively represent the first weight parameter, the first bias parameter and the second weight parameter of the input gate, Indicates that the i-th unit and the hidden state of the output of the i-1th unit To splice, represents the memory cell update of the i-1th unit, , and Respectively represent the first weight parameter, the first bias parameter and the second weight parameter of the forget gate, , and Respectively represent the first weight parameter, the first bias parameter and the second weight parameter of the output gate, and They represent the first weight parameter and the first bias parameter of the candidate memory unit respectively, sigmoid represents the sigmoid activation function, and tanh represents the hyperbolic tangent activation function.

6. The activity analysis system of the stem cell exosome uric acid-lowering preparation according to claim 4, characterized in that: The calculation formula of the attention mechanism layer is: ； in, represents the weighted hidden state of each unit and the unit's feature weight, , Respectively represent the first weight parameter and the first bias parameter of each unit, , and They respectively represent the first weight parameter, the second weight parameter and the first bias parameter of the attention mechanism within the unit, and i represents the unit index of the second feature sequence.

7. The activity analysis system of the stem cell exosome uric acid lowering preparation according to claim 1, characterized in that: The optimal parameter combination is generated based on the improved ivy optimization algorithm combined with the uric acid-lowering effect prediction model. The specific steps include: Step S301, generating an initialization population with Q individuals meeting the constraint conditions through Latin hypercube sampling, where Q is a custom parameter; The individual positions of the initialized population are represented by the uric acid-lowering feature vector, and the encoding format of the uric acid-lowering feature vector is: ,in, represents the uric acid lowering feature vector, indicates the source of stem cells, and A indicates the analysis method; Step S302, constructing an objective function according to the uric acid-lowering effect prediction model; Step S303, calculating the growth rate change of each individual according to the growth rate formula, wherein the growth rate formula is: ,in, and represents the change in growth rate of individual j at the k+1th and kth iterations, k is the index of the iteration number, j and a are the indexes of the individuals, represents the dynamic enhancement coefficient, and represents the growth rate of individuals a and j at the kth iteration, exp represents the exponential function, represents the growth rate regulation coefficient, represents the maximum growth rate of individual j, represents the time attenuation coefficient, represents the square of a uniformly distributed random number, represents a random vector that obeys a standard normal distribution, and the dimension of the solution space is D. Represents element-wise multiplication; Step S304, updating the position of each individual according to the climbing formula and the fitness of the individual obtained by the objective function, the climbing formula is: ,in, represents the updated position of individual j at the kth iteration, represents the position of individual j at the kth iteration, represents the position of the individual with the highest fitness among the neighbors of individual j, represents the climb enhancement factor, represents the position of the individual with the highest fitness in the population at the kth iteration, Represents the cubic value of a uniformly distributed random number, represents the nonlinear amplification factor; Step S305, calculate the fitness of each individual according to the objective function, and select and update the individuals through the update mechanism. The formula of the update mechanism is: ,in, represents the individual position after the individual is processed by the update mechanism, represents a D-dimensional uniformly distributed random number vector, represents the dynamic threshold parameter, and represents the fitness of individual j at the kth iteration and the fitness of the individual with the highest fitness in the population at the kth iteration, Represents the direction control coefficient, which controls the step size of the individual when updating the direction; Step S306, the current number of iterations is increased by 1. When the current number of iterations reaches the maximum number of iterations, the individual with the highest current fitness is output as the optimal parameter combination. Otherwise, return to step S303, where the maximum number of iterations is a custom parameter.

8. The activity analysis system of the stem cell exosome uric acid lowering preparation according to claim 7, characterized in that: The specific steps of step S301 include: Step S3011, determining the upper and lower bounds of the individual solution space, that is, the maximum and minimum values ​​of each dimension in the individual's uric acid-lowering feature vector; Step S3012, divide the solution space into isometric subspaces, randomly select a value in each subspace to form a D-dimensional coordinate; Step S3013, randomly arrange between different dimensions to form a Latin hypercube sampling matrix. The Latin hypercube sampling matrix is ​​a A matrix with each row corresponding to a sample point and each column corresponding to a sampling value of a dimension, ranging from 0 to 1; Step S3014, mapping each row of the Latin hypercube sampling matrix to the solution space according to a mapping formula as an individual of the initial population, wherein the mapping formula is: ,in, represents the individual position a of the initialized population, and They represent the upper and lower bounds of the solution space, represents the i-th row of the Latin hypercube sampling matrix.

9. The activity analysis system of the stem cell exosome uric acid-lowering preparation according to claim 7, characterized in that: The calculation formula of the objective function is: ； in, represents the fitness of an individual, represents the initial efficacy weight coefficient, represents the therapeutic effect weight step length, It indicates the uric acid reduction value obtained by the individual according to the uric acid reduction effect prediction model. represents the side effect weight coefficient, represents the side effect weight step size, , and Respectively represent the weight coefficients of exosome concentration, RNA component ratio and protein component ratio for the total cost.

Images