A gene selection method based on multi-target feature selection for biological trait expression

Abstract

The application discloses a gene selection method based on biological character expression of multi-target feature selection, which comprises the following steps: step 1, inputting biological gene samples; step 2, grouping features of the input gene samples and generating a sub-population for each group; step 3, independently evolving each sub-population; step 4, gradually merging all the sub-populations; and step 5, finally merging all the sub-populations into one population, searching on the original genes, finally obtaining the best partial genes on the original gene set in expression effect, and taking the partial genes as the finally selected genes. The application can screen a small number of important genes from a large number of gene features, thereby effectively improving the accuracy of gene selection of high-dimensional biological character expression, and is also beneficial to shortening a research period and reducing research cost.

Description

Technical Field

[0001] This invention belongs to the field of biotechnology, specifically a gene selection method for biological trait expression based on multi-target feature selection. Background Technology

[0002] Biological genomic data often has several characteristics compared to data from other fields: First, it has high gene dimensionality and a small sample size. Second, subtle changes in gene expression can lead to completely different traits. Third, high-dimensional genes often contain noisy, redundant genes, while the number of genes containing truly useful information is very small. However, finding the genome corresponding to a specific trait expression within such a large number of biological genes is crucial, for example, identifying which gene expression triggers a disease for early prevention and treatment. Existing methods for selecting genes expressing a specific trait, such as traditional optical sequencing, require manual operation by researchers and expensive equipment. Their biggest drawback is that, given the large number of genes and their diverse traits, existing methods struggle to efficiently identify the expression genomes corresponding to different traits. In short, these existing methods often face the problem of enormous research costs and high inefficiency. Summary of the Invention

[0003] This invention aims to address the existing problem of gene selection targeting a specific trait expression in high-dimensional biological genes. It proposes a gene selection method for biological trait expression based on multi-target feature selection, with the goal of screening out a smaller number of genes containing effective information from a large number of genes, thereby effectively improving the accuracy of gene selection for high-dimensional biological trait expression.

[0004] To achieve the above-mentioned objectives, the present invention adopts the following technical solution:

[0005] The gene selection method for biological trait expression based on multi-target feature selection of the present invention is characterized by the following steps:

[0006] Step 1: Obtain an X×D dimensional gene feature set, where X represents the number of samples in the gene feature set and D represents the number of genes in the gene feature set; divide the D-dimensional genes of any sample in the gene feature set into k groups of genes, where the i-th group of genes is denoted as G. i ={g (i-1)×d+1 ,g (i-1)×d+2 ,…,g (i-1)×d+j ,…,g i×d}, g (i-1)×d+j Let represent the j-th dimension gene of the i-th gene group; d represents the number of genes in each group, and d = D / k; i ∈ [1, k];

[0007] Step 2: Set the number of sub-populations as k; the size of each sub-population as N;

[0008] Define the i-th sub-population P i = {p i,1 ,p i,2 ,…,p i,j ,…,p i,N}, wherein p i,j represents the j-th individual in the i-th sub-population P i , j ∈ [1, N], and p i,j = {f i,j,1 ,f i,j,2 ,…,f i,j,s ,…,f i,j,d}, f i,k,s represents whether the s-th gene feature of the j-th individual p i in P i,，j is selected; if f i,，j,s = 1, it represents that the s-th gene is selected, if f i,j,s = 0, it represents that the s-th gene is not selected; s ∈ [1, d] ;

[0009] Let the genome searched by the i-th sub-population P i be G i ;

[0010] Step 3: Define the maximum number of iterations as Tmax, and the maximum number of iterations in the first stage as T;

[0011] Step 3.1: Define and initialize the current generation number t = 1, and randomly initialize the k sub-populations of the t-th generation;

[0012] Step 3.2: Initialize i = 1;

[0013] Step 3.3: Calculate the error rate of the selected gene features of the j-th individual p in the i-th sub-population P of the t-th generation according to formula (1)

[0014]

[0015] In formula (1), represents the number of samples classified incorrectly using the selected gene features of the j-th individual p in the i-th sub-population P of the t-th generation;

[0016] Step 3.3: Calculate the ratio of the selected gene features to the original number of genes of the j-th individual p in the i-th sub-population P of the t-th generation according to formula (2)

[0017]

[0018] In equation (2), This indicates the use of the i-th subpopulation in generation t. The j-th individual The number of selected gene features;

[0019] Step 4: Generate the i-th offspring population in generation t;

[0020] Step 4.1: Starting from the i-th subpopulation of generation t. Select two individuals and At the same time from Randomly select a gene feature f random , where (i-1)×d <random≤i×d;α∈[1,N];β∈[1,N];

[0021] Step 4.2: Exchange individuals and The [random, i×d]-dimensional genetic characteristics are used to generate two new individuals. and

[0022] Step 4.3: Generate N new individuals following the process in Steps 4.1-4.2, and form the i-th offspring population of the t-th generation.

[0023] Step 4.4: Assign i+1 to i. If i≤k, return to step 4.1 and execute sequentially; otherwise, execute step 5.

[0024] Step 5: Initialize i = 1;

[0025] Step 5.0: Define and initialize variable q = 1; set the population of the i-th offspring in generation t. and the i-th subpopulation Merged into the i-th merged population of generation t. Its population size is

[0026] Step 5.1: Define and initialize the i-th merged population in generation t. The set of individuals expressing the qth gene It is an empty set;

[0027] Step 5.2: Initialize u = 1;

[0028] Step 5.3: If Then calculate the i-th merged population in generation t. the u-th individual in two objective functions of and if the individual is a non-dominated individual, add it to the set otherwise, directly execute step 5.4;

[0029] step 5.4: assign u+1 to u, return to step 5.3 until ;

[0030] step 5.5: delete all individuals in the set from the population ;

[0031] step 5.6: if the size of the population is not 0, reassign the size of the population after deletion to and assign q+1 to q, then return to step 5.1; otherwise, indicate that the set of q gene expression individuals is generated and execute step 5.7;

[0032] step 5.7: add the individuals in to to the i-th offspring population of the (t+1)-th generation in order until the number of individuals in the offspring population reaches N;

[0033] step 5.8: assign i+1 to i, if i>k, execute step 6, otherwise, return to step 5.0 in order;

[0034] step 6: assign t+1 to t, if t>T, execute step 7; otherwise return to step 4 in order;

[0035] step 7: set the number of generations of independent evolution as M; and initialize i=1;

[0036] step 7.1: select one individual from the i-th offspring population of the T-th generation and one individual from the (i+1)-th offspring population of the T-th generation and generate a gene feature set denote the number of genes in the gene feature set

[0037] step 7.2: define and initialize a variable o=1;

[0038] step 7.3: generate two random numbers​​​ and if then select the othegene from add the gene to the gene set

[0039] Step 7.4: assign o+1 to o, if then return to step 7.3; otherwise, select all genes in the gene set and generate a new individual

[0040] Step 7.5: repeat steps 7.1-7.4 until 2xN new individuals are generated and form the (T+1)th sub-population

[0041] Step 7.6: assign i+2 to i, if i<k, then return to step 7.1 for sequential execution; otherwise, indicate that all sub-populations of the (T+1)th generation are obtained and step 8 is executed;

[0042] Step 8: assign 2xN to N and 2xd to d; assign to k and T+1 to T, so that all sub-populations of the Tth generation are obtained and denoted as wherein denotes the ith sub-population of the Tth generation; mod denotes the remainder calculation;

[0043] if k=1, then the Tth generation population P is composed of only one sub-population T , and jump to step 10; otherwise, step 9 is executed;

[0044] Step 9: initialize m=1;

[0045] Step 9.1: re-initialize i=1;

[0046] Step 9.2: select two individuals and of the Tth generation from the ith sub-population and let the variables of the Tth generation be then the gene features with value "1" in and represent different selection modes of the same gene for the two individuals and

[0047] Step 9.3: add all gene features with value "1" in to the gene feature set of the Tth generation, so that is obtained, wherein Represents the set of gene characteristics in generation T. In the c-th dimension, the genes are represented, and gd is the set of gene features. The number of genes; c∈[1,gd];

[0048] Step 9.4: Generate a set of 1×gd-dimensional Boolean vectors for the Tth generation. in, Represents the Boolean vector of generation T. The c-th dimension Boolean value in, and

[0049] Step 9.5: If the Boolean vector of generation T... In Then exchange the two individuals in generation T. and The c-th gene characteristic produces two new individuals in the T-th generation. and

[0050] Step 9.6: Generate a population of size N for the Tth generation following the steps 9.2-9.5. and the population of generation T. After merging, the population of generation T+1 is selected using non-dominated sorting.

[0051] Step 9.7: Assign i+1 to i. If i≤k, return to step 9.2; otherwise, proceed to step 9.8.

[0052] Step 9.8: Assign T+1 to T. If T>Tmax, it means that the population P of generation Tmax has been obtained. Tmax Then select the best individual from them as the final gene selection result output; otherwise, proceed to step 9.9;

[0053] Step 9.9: Assign m+1 to m; if m≤M, then return to step 9.1; otherwise, proceed to step 7.

[0054] The gene selection method for biological trait expression based on multi-target feature selection described in this invention is also characterized in that step 5.3 includes:

[0055] Step 5.3.0: Initialize v = 1;

[0056] Step 5.3.1: If v≠u and If so, proceed to step 5.3.2; otherwise, proceed to step 5.3.3.

[0057] Step 5.3.2: Calculate the merged population The vth individual in the population Two objective functions and if and then the individual is a non-dominated individual in the population and is added to the set Otherwise, step 5.3.3 is executed directly.

[0058] Step 5.3.3: v+1 is assigned to v, and step 5.3.1 is returned until .

[0059] The step 7.1 comprises:

[0060] Step 7.1.1: the gene features with value 1 in the individual are composed into a set The gene features with value 1 in the individual are composed into a set

[0061] Step 7.1.2: the gene feature set and the gene feature set are merged into a new gene feature set

[0062] The electronic device comprises a memory and a processor, and is characterized in that the memory is used for storing a program supporting the processor to execute the gene selection method, and the processor is configured to execute the program stored in the memory.

[0063] The computer readable storage medium stores a computer program, and the computer program is characterized in that when the computer program is run by a processor, the steps of the gene selection method are executed.

[0064] Compared with the prior art, the present application has the beneficial effects that:

[0065] 1、The method can find some expression genes containing effective information and extremely small in number in high-dimensional original genes by using the feature selection method on the original genes.

[0066] ​2、The present application can solve the difficulty of finding effective genes in the huge search space generated by high-dimensional genes when facing original genes, and can effectively provide a solution when facing various high-dimensional and large-scale gene problems in practical problems.

[0067] 3、The present application can overcome the contradiction between classification accuracy and feature number in the traditional method, and can effectively reduce the research cost and shorten the research period. BRIEF DESCRIPTION OF DRAWINGS

[0068] Figure 1 The flowchart of the gene selection method for biological trait expression based on multi-objective feature selection. DETAILED DESCRIPTION

[0069] In this embodiment, a gene selection method for biological trait expression based on multi-objective feature selection is used to select a small number of genes with high importance from a gene expression dataset. These genes are much smaller in number than the original genes, but contain a large amount of effective information. Specifically, as shown in the figure, the gene selection method is performed according to the following steps: Figure 1

[0070] Step 1: Obtain a set of XxD-dimensional gene feature sets, where X represents the number of samples in the gene feature set, and D represents the number of genes in the gene feature set; divide the D-dimensional genes of any one sample in the gene feature set into k groups of genes, where the i-th group of genes is denoted as G i ={g (i-1)×d+1 ,g (i-1)×d+2 ,…,g (i-1)×d+j ,…,g i×d}, g (i-1)d+j represents the j-th gene in the i-th group of genes; d represents the number of genes in each group, and d=D / k; i∈[1,k]; obtain the minimum value g y,min and the maximum value g y,max of the y-th gene expressed in the dataset.

[0071] Step 1.1: Obtain the value g y,origin of the y-th gene of the x-th sample expressed in the dataset.​

[0072] Step 1.2: Calculate the new gene expression value g y,origin =(g y,origin -g y,min ) / g y,max -g y,min ).

[0073] Step 2: Set the number of subpopulations to k; the size of each subpopulation to N;

[0074] Define the i-th subpopulation P i ={p i,1 ,p i,2 ,…,p i,j ,…,p i,N}, where p i,j Let P represent the i-th subpopulation. i The j-th individual, j∈[1,N], and p i,j ={f i,j,1 ,f i,j,2 ,…,f i,j,s ,…,f i,j,d}, f i,j,s P represents i The j-th individual p i,j Whether to select the s-th gene trait; if f i,j,s =1, which means selecting the s-th gene. If f i,j,s =0, which means that the s-th gene was not selected; s∈[1,d];

[0075] Let the i-th subpopulation P i The searched genome is G. i ;

[0076] Step 3: Define the maximum number of iterations as Tmax, and the maximum number of iterations in the first stage as T;

[0077] Step 3.1: Define and initialize the current generation t=1, and randomly initialize the k subpopulations of the t-th generation;

[0078] Step 3.2: Initialize i = 1;

[0079] Step 3.3: Calculate the i-th subpopulation in generation t according to equation (1). The j-th individual Error rate of selected gene features

[0080]

[0081] In equation (1), This indicates the use of the i-th subpopulation in generation t. The j-th individual the number of samples misclassified by the selected gene features;

[0082] Step 3.3: Calculate the ith sub-population of the tth generation according to formula (2) the jth individual in the ith sub-population of the tth generation the ratio of the number of selected gene features to the number of original genes

[0083]

[0084] In formula (2), denotes using the ith sub-population of the tth generation the jth individual in the ith sub-population of the tth generation the number of selected gene features;

[0085] Step 4: Generate the ith offspring population of the tth generation;

[0086] Step 4.1: Select two individuals and from the ith sub-population of the tth generation and randomly select a gene feature f from random where (i-1)xd< random≤i×d; α∈[1, N]; β∈[1, N];

[0087] Step 4.2: Exchange the individuals and and the gene features in the dimension of [random, i×d] between them, thereby generating two new individuals and

[0088] Step 4.3: Generate N new individuals according to the process of Step 4.1-Step 4.2 and compose the ith offspring population of the tth generation

[0089] Step 4.4: Assign i+1 to i, and if i≤k, return to Step 4.1 for sequential execution; otherwise, execute Step 5;

[0090] Step 5: Initialize i=1;

[0091] Step 5.0: Define and initialize the variable q=1; merge the ith offspring population of the tth generation and the ith sub-population into the ith merged population of the tth generation with a population size of

[0092] Step 5.1: Define and initialize the ith merged population of the tth generation ​the qth gene expression individual set is empty;

[0093] Step 5.2: initialize u = 1;

[0094] Step 5.3: if calculate the two objective functions and of the u-th individual in the i-th merged population of the t-th generation if the individual is a non-dominated individual, add it to the set otherwise, directly execute Step 5.4; Step 5.3.0: initialize v = 1;

[0095] Step 5.3.1: if v ≠ u and

[0096] execute Step 5.3.2; otherwise, execute Step 5.3.3; Step 5.3.2: calculate the two objective functions

[0097] and of the v-th individual in the merged population if and indicate that the individual is a non-dominated individual in the population and add to the set otherwise, directly execute Step 5.3.3; Step 5.3.3: assign v + 1 to v and return to Step 5.3.1 until .

[0098] Step 5.4: assign u + 1 to u and return to Step 5.3 until .

[0099] Step 5.5: remove all individuals in the set from the population

[0100] Step 5.6: if the size of the population is not 0, reassign the size of the population after removal to and assign q + 1 to q, then return to Step 5.1; otherwise, indicate that q gene expression individual sets

[0101] containing individuals are generated, the number of individuals in each set is ​​​​and execute step 5.7;

[0102] Step 5.7: Add the individuals in to the i-th sub-population of the t+1-th generation in order until the number of individuals in the sub-population is N; for the first η Pareto sets, we have: and and then select individuals in according to the following rules:

[0103] Step 5.7.1: Calculate the crowding distance of the μ-th individual in

[0104] Step 5.7.2: Sort all individuals in according to the crowding distance , and select the top individuals into the next generation sub-population

[0105] Step 5.8: Assign i+1 to i, if i>k, execute step 6, otherwise, return to step 5.0 and execute in order;

[0106] Step 6: Assign t+1 to t, if t>T, execute step 7; otherwise return to step 4 and execute in order;

[0107] Step 7: Set the number of independent evolution generations as M; note that M≤(Tmax-T) / log2k should be satisfied to avoid that all sub-populations have not completed merging when the maximum number of iterations Tmax is reached. Set i=1;

[0108] Step 7.1: Select one individual from the i-th sub-population of the T-th generation and one individual from the i+1-th sub-population of the T-th generation, and generate the gene feature set Let the number of genes in the gene feature set be Step 7.1.1: Form the set from the genes with value 1 in the individual Form the set

[0109] from the genes with value 1 in the individual ​​​​​

[0110] Step 7.1.2: Combine the gene feature set and the gene feature set to form a new gene feature set

[0111] Step 7.2: Define and initialize variable o = 1;

[0112] Step 7.2.1: Generate two random numbers for the Tth generation and If then select the oth gene from and add it to the gene set

[0113] Step 7.2.2: Assign o + 1 to o, if then return to Step 7.2.1; otherwise, select all genes from the gene set and generate a new individual For genes belonging to G i and G i+1 but not to , all the genes in the individual are set to 0, indicating that they are not selected.

[0114] Step 7.3: Repeat Steps 7.1-7.2.2 until 2 x N new individuals are generated and form the (T + 1)th generation of sub-populations population The search range of genes expands to G i ∪ G i+1 , and the length of the genes represented by the individual becomes 2 x d.

[0115] Step 7.4: Assign i + 2 to i, if i < k, then return to Step 7.1 for sequential execution; otherwise, indicate that all sub-populations of the (T + 1)th generation are obtained, and execute Step 8;

[0116] Step 8: Assign 2 x N to N, and 2 x d to d; assign to k, and T + 1 to T, thus obtaining all sub-populations of the Tth generation, denoted as wherein represents the ith sub-population of the Tth generation; mod represents the remainder calculation; at this time, the number of sub-populations becomes half of the original, but the population size of the sub-populations and the number of genes represented by the individuals in the population become twice the original.

[0117] If k = 1, it indicates that the Tth generation population PT is composed of only one sub-population, and jump to Step 10; otherwise, execute Step 9;

[0118] ​Step 9: initialize m = 1;

[0119] Step 9.1: reinitialize i = 1;

[0120] Step 9.2: select the i-th sub-population of the T-th generation select two individuals of the T-th generation and Let the variables of the T-th generation be then the gene features with value "1" in and different selection methods for the same gene; ^ represents XOR calculation;

[0121] Step 9.3: add all gene features with value "1" in to the gene feature set of the T-th generation to obtain wherein, represents the gene feature set of the T-th generation the gene in the c-th dimension of the gene feature set gd is the number of genes in the gene feature set c ∈ [1, gd]; when different sub-populations are gradually merged, the number of genes represented by each individual increases exponentially, at this time, using the previous results, important local genes can be quickly selected, that is the genes corresponding to the bits with value 1 in

[0122] selecting this small number of important genes can effectively solve the difficulty of high-dimensional gene selection. Step 9.4: generate a set of 1 × gd-dimensional Boolean vectors of the T-th generation wherein, represents the c-th Boolean value in the Boolean vector of the T-th generation

[0123] Step 9.5: if in the Boolean vector of the T-th generation then exchange the c-th gene features of the two individuals of the T-th generation and and generate two new individuals of the T-th generation and

[0124] Step 9.6: generate the offspring population of the T-th generation with size N according to the process of Step 9.2-Step 9.5 and merge with the population of the T-th generation to select the population of the T+1-th generation

[0125] Step 9.7: i+1 is assigned to i, if i≤k, then return to execute step 9.2; otherwise, execute step 9.8;

[0126] Step 9.8: T+1 is assigned to T, if T>Tmax, then it means that the Tmax population P Tmax is obtained, and the optimal individual is selected from the population as the final gene selection result, otherwise, execute step 9.9;

[0127] Step 9.9: m+1 is assigned to m; if m≤M, then return to execute step 9.1; otherwise, return to execute step 7.

[0128] In the embodiment, an electronic device includes a memory for storing a program supporting a processor to execute the above method, and the processor is configured to execute the program stored in the memory.

[0129] In the embodiment, a computer readable storage medium has a computer program stored thereon, and the computer program is executed by a processor to execute the steps of the above method.

[0130] The specific implementation method of the present application is illustrated by an example of real gene data set, and the effect of the method of the present application is verified.

[0131] 1) Preparation of standard data set

[0132] The gse19804 data set is used as a standard data set to verify the effectiveness of the gene selection method based on multi-target feature selection of biological trait expression, and the gse19804 data set is a standard data set widely used in gene feature selection problems. In the gse19804 data set, 120 data are included, i.e. corresponding to collected gene samples, wherein each sample includes 21655-dimensional features, indicating that each sample includes 21655 genes. All samples are used to verify the performance of the selected expression genes of the present application. For the 120 training samples, they are divided into training sample and test sample data.

[0133] 2) Evaluation index

[0134] The classification accuracy and the number of selected features are used as the evaluation index of the present example, which is used to evaluate the performance of the feature selection of the present application on different biological gene expression samples. The higher the classification accuracy value and the fewer the number of features, the better the feature selection effect.

[0135] 3) Experiment on standard data set

[0136] In order to verify the effectiveness of the method, the algorithm (SOEA) of the present application is compared with the classical MOFSBDE algorithm, the VS-CCPSO algorithm and the PSO-EMT algorithm in modeling and prediction on the gse19804 dataset, and the classification accuracy and the number of features are compared. The experimental results are shown in Table 1.

[0137] Table 1 Comparison of experimental results

[0138]

[0139] As can be seen from Table 1, in the standard dataset gse19804, the SOEA gene feature selection algorithm of the present application can obtain higher classification accuracy while obtaining lower number of features compared with the other three algorithms, thereby reflecting the effectiveness of the present application.

Claims

1. A gene selection method for biological trait expression based on multi-objective feature selection, characterized by, is performed as follows: Step 1: Obtain a set of gene features of XxD dimensions, wherein X represents the number of samples in the gene feature set, and D represents the number of genes in the gene feature set; divide the D-dimensional genes of any one sample in the gene feature set into k groups of genes, wherein the i-th group of genes is denoted as G i ={g (i-1)×d+1 , g (i-1)×d+2 ,..., g (i-1)×d+f ,..., g i×d}, g (i-1)×d+j represents the j-th gene in the i-th group of genes; d represents the number of genes in each group, and d=D / k; i∈[1, k] Step 2: Set the number of sub-populations as k; the size of each sub-population as N; Definition of the ith sub-population P i = {p i，1 , p i，2 ,..., p i，j ,..., p i，N}, wherein p i，j represents the jth individual in the ith sub-population P i , j ∈ [1, N], and p i，j = {f i，j，1 , f i，j，2 ,..., f i，j，s ,..., f i，j，d}, f i，j，s represents whether the s th gene feature of the jth individual p i in P i，j is selected; if f i，j，s = 1, it represents that the s th gene is selected, and if f i，j，s = 0, it represents that the s th gene is not selected; s ∈ [1, d] Let the ith subpopulation P i The genome searched is G i ; Step 3: Define the maximum iteration number as Tmax, and the maximum iteration number of the first stage as T; Step 3.1: Define and initialize the current generation number t = 1, and randomly initialize the k sub-populations of the tth generation; Step 3.2: Initialize i = 1; Step 3.3: Calculate the ith sub-population of the tth generation according to formula (1) the jth individual in the middle error rate of the selected genetic features In formula (1), denotes the use of the i-th subpopulation of the t-th generation the j-th individual in the i-th subpopulation of the t-th generation the number of samples classified incorrectly by the selected gene signature Step 3.3: Calculate the ith sub-population of the tth generation according to formula (2) the jth individual in the population the ratio of the selected gene features to the number of original genes In formula (2), denotes the use of the i-th subpopulation of the t-th generation the j-th individual in the i-th subpopulation of the t-th generation the selected number of genetic features; Step 4: Generate the ith sub-population of the tth generation; Step 4.1: Select two individuals from the ith subpopulation of the tth generation and randomly select one gene feature f from random where (i-1)xd< random < ixd; a e [1, N]; β e [1, N].​​ Step 4.2: Interchange individuals and the gene features of dimension [random, i x d], resulting in two new individuals and Step 4.3: N new individuals are generated following the procedure of Step 4.1-Step 4.2 and form the i-th subpopulation of the t-th generation Step 4.4: Assign i + 1 to i, and if i ≤ k, return to Step 4.1 for sequential execution; otherwise, execute Step 5; Step 5: Initialize i = 1; Step 5.0: Define and initialize variable q = 1; merge the ith subpopulation of the tth generation into the ith merged population of the tth generation and the ith subpopulation of the tth generation into the ith merged population of the tth generation with population size Step 5.1 : Define and initialize the ith merged population of the tth generation of the qth set of gene expression individuals is an empty set; Step 5.2: Initialize u = 1; Step 5.3: If then calculate two objective functions and of the u-th individual in the i-th merged population of the t-th generation and If the individual is non-dominated, add it to the set Otherwise, go directly to Step 5.4; Step 5.4: Assign u+1 to u, go back to step 5.3 until ; Step 5.5: Remove all individuals from the population in the set in the set Step 5.6: If the size of the population is not 0, the size of the population after the deletion is re-assigned to and q+1 is assigned to q, and then the step 5.1 is returned; otherwise, it indicates that q sets of gene expression individuals are generated and the step 5.7 is executed. Step 5.7: Add the individuals in to the i-th subpopulation of the t+1 -th generation in order until the number of individuals in the subpopulation reaches N. Step 5.8: Repeat steps 5.1 to 5.7 until the number of generations reaches G. Step 5.8: Assign i + 1 to i, and if i > k, execute Step 6, otherwise, return to Step 5.0 for sequential execution; Step 6: Assign t + 1 to t, and if t > T, execute Step 7; otherwise, return to Step 4 for sequential execution; Step 7: Set the number of independent evolution generations as M; and initialize i = 1; Step 7.1 : Selecting one individual from the i-th sub-population of the T-th generation and the i+1-th sub-population of the T-th generation respectively and and generating a gene feature set denoted as with the number of genes Step 7.2: Define and initialize the variable o = 1; Step 7.3: Generate two random numbers for the Tth generation and if then select the oth gene from add the gene to the set of genes ​ Step 7.4: Assign o + 1 to o, and if then return to step 7.3; otherwise, select all genes in the gene set and create a new individual Step 7.5: Steps 7.1 - 7.4 are repeated until 2 x N new individuals are produced and a subpopulation of Generation T+1 is formed Step 7.6: Assign i + 2 to i, and if i < k, return to Step 7.1 for sequential execution; otherwise, indicate that all sub-populations of the T+1th generation are obtained, and execute Step 8; Step 8: Assign 2xN to N and 2xd to d; assign k to k and T+1 to T, and thus obtain all the sub-populations of the Tth generation, denoted as wherein, wherein, represents the ith sub-population of the Tth generation; mod represents the remainder calculation. If k = 1, it means that the Tth generation population P is composed of only one sub-population T , jump to step 10; otherwise, execute step 9; Step 9: Initialize m = 1; Step 9.1: Reinitialize i = 1; Step 9.2: Selecting the ith sub-population of the Tth generation Selecting two individuals of the Tth generation and Simultaneously setting the variable of the Tth generation Then The gene feature with value "1" in the Tth generation indicates that the two individuals of the Tth generation and Different selection methods for the same gene; represents XOR calculation; Step 9.3: [The text appears to be incomplete and contains several grammatical errors. A more accurate translation would require the full context.] All gene traits with a value of "1" are added to the gene trait set of generation T. Thus obtain in, Represents the set of gene characteristics in generation T. In the c-th dimension, the genes are represented, and gd is the set of gene features. The number of genes; c∈[1, gd]; Step 9.4: Generate a set of 1 x gd dimensional Boolean vectors for the Tth generation where, represents the cth Boolean value in the Tth generation Boolean vector ​ Step 9.5: If the Boolean vector in the Tth generation then exchange the cth gene feature of the two individuals in the Tth generation and and produce two new individuals in the Tth generation and ​ Step 9.6: The offspring population of size N for generation T is generated following the procedure of steps 9.2 - 9.5 and combined with the population of generation T After the combination, the population of generation T+1 is selected by non-dominated sorting Step 9.7: Assign i + 1 to i, and if i ≤ k, return to Step 9.2 for execution; otherwise, execute Step 9.8; Step 9.8: T+1 is assigned to T, and if T>Tmax, it indicates that the Tth population P is obtained Tmax and the optimal individual is selected as the final genetic selection result output; otherwise, step 9.9 is executed. Step 9.9: Assign m + 1 to m; if m ≤ M, return to Step 9.1 for execution; otherwise, execute Step 7.

2. The method according to claim 1, wherein, The Step 5.3 comprises: Step 5.3.0: Initialize v = 1; Step 5.3.1 : if v≠u and Step 5.3.2 is performed; otherwise, Step 5.3.3 is performed. Step 5.3.2: Calculate the merged population The vth individual in the population Two objective functions and If and then individual is a non-dominated individual in the population and add to the set Otherwise, proceed directly to Step 5.3.3; Step 5.3.3: Assign v+1 to v, go back to step 5.3.1 until .

3. The method of claim 1, wherein the method is based on multi-objective feature selection of biological trait expression. The Step 7.1 comprises: Step 7.1.1 : The individual The set of gene features with value 1 in the individual The individual The set of gene features with value 1 in the individual Step 7.1.2: Combining gene feature sets and gene feature sets for new gene feature sets 4. An electronic device comprising a memory and a processor, characterized in that The memory is configured to store a program supporting the processor to execute the genetic selection method of any one of claims 1-3, and the processor is configured to execute the program stored in the memory.

5. A computer-readable storage medium having stored thereon a computer program, characterized in that The computer program is executed by the processor to perform the steps of the genetic selection method of any one of claims 1-3. The computer program is executed by the processor to perform the steps of the genetic selection method of any one of claims 1-3.

Images