Character selection method and device

Abstract

The embodiment of the invention provides a centrum selection method and device. The method comprises the steps that a background channel, a medicine set and patient data are determined; determining a drug disorder pathway based on the background pathway and the drug set, and determining a disease disorder pathway based on the background pathway and the patient data; determining a molecular herbal database based on the disease disorder pathway and the drug disorder pathway; determining an initial group based on the molecular herbal database; wherein the initial group comprises a plurality of individuals, and the individuals comprise charms; a fitness function is determined, a genetic algorithm is carried out based on the initial population and the fitness function, a target charm is determined, and the fitness function comprises a number parameter and a multi-index parameter of the cell line. The traditional Chinese medicine prescriptions are screened from multiple perspectives of reversal of disorder pathways and drug effect scores. Compared with the limitation that a traditional method depends on a single index to carry out optimization screening, the multi-dimensional screening strategy can better reflect the overall drug effect and the treatment effect of the traditional Chinese medicine prescription, and therefore the application value is improved.

Description

Technical Field

[0001] The embodiments of this specification relate to the technical field of traditional Chinese medicine development, and in particular to a method for selecting a good prescription. Background Art

[0002] Current methods for formulating traditional Chinese medicine include exhaustive method and TCMFP.

[0003] The exhaustive method is based on a pre-set TCM drug pool and enumerates all TCM compatibility combinations to find the best prescription. Although this method can ensure the comprehensive exploration and capture of all potential beneficial formulas and ensure the comprehensiveness of the screening conclusion, it is extremely time-consuming when the number of good prescriptions is large. In addition, the exhaustive method itself does not support the direct optimization of drug dosage configuration, which is another major limitation in its practicality.

[0004] TCMFP is a prediction method for good Chinese medicine prescriptions (TCMFP) that integrates the experience of good Chinese medicine prescriptions, modern pharmacology, network pharmacology and genetic algorithms. This method uses disease-related genes, protein-protein interactions (PPIs) and Chinese medicine components to construct a "target-molecule-TCM" network, establish a TCM database, and collect drug pairs in classic prescriptions from databases, ancient books or literature, and organize the data to form a TCM pair dataset. Hscore is calculated based on the TCM dataset and the "target-molecule-TCM" network, and the random walk algorithm is used to screen out important TCMs. Combined with the TCM drug pair dataset, Pscore is calculated to determine the high-frequency TCM combination, and finally Hscore and Pscore are integrated to calculate FmapScore as the evaluation score of the herbal formula. Genetic algorithms are used to predict the best prescriptions, including prescription initialization, calculation of FmapScore, retention of the top 50% of high-scoring prescriptions, and crossover mutation. Then, the predicted TCM prescriptions are analyzed for effectiveness. Finally, based on the protein interaction network, KEGG pathway and GO enrichment analysis, the predicted best herbal formula is evaluated for its impact on the disease target.

[0005] The human body is a complex system with multiple cells, tissues and organs working in a highly coordinated manner. Studying the effect of a drug on a single cell line cannot fully reflect the complex reaction of the drug in a multi-cellular physiological environment. When screening traditional Chinese medicine prescriptions, the effects of the prescription on multiple cell lines should be comprehensively considered.

[0006] At present, the exhaustive method can only optimize a certain cell line and then evaluate it on other cell lines. This method cannot take into account the global optimal solution of all cell lines at the same time, resulting in the optimization results being biased towards the local area and it is difficult to find the global optimal solution in multiple cell lines. When searching for the optimal solution, general heuristic genetic algorithms often use a single indicator for selection evolution, and the screening results perform better in one aspect. Therefore, a better solution is urgently needed. Summary of the invention

[0007] In view of this, the embodiments of this specification provide a good party selection method. One or more embodiments of this specification also relate to a good party selection device, a computing device, a computer-readable storage medium and a computer program to solve the technical defects existing in the prior art.

[0008] According to a first aspect of an embodiment of this specification, a method for selecting a good party is provided, comprising:

[0009] Identify background pathways, drug collections, and patient data;

[0010] Identify drug-dysregulated pathways based on background pathways and drug collections, and identify disease-dysregulated pathways based on background pathways and patient data;

[0011] Identify molecular herbal databases based on disease- and drug-disordered pathways;

[0012] Determine an initial group based on a molecular herbal database; wherein the initial group includes multiple individuals, and the individuals include Shan Fang;

[0013] Determine the fitness function, perform a genetic algorithm based on the initial population and the fitness function, and determine the target solution; wherein the fitness function includes the quantity parameter and multi-index parameter of the cell line.

[0014] In one possible implementation, determining a drug disorder pathway based on a background pathway and a drug set includes:

[0015] Determine drug expression profile data by intervening cells with a drug collection;

[0016] Gene enrichment analysis based on drug expression profile data and background pathways was performed to identify drug dysregulation pathways.

[0017] In one possible implementation, determining a drug disorder pathway based on a background pathway and a drug set includes:

[0018] Perform differential gene analysis based on patient data to determine disease expression profile data;

[0019] Identify disease dysregulated pathways based on disease expression profiling data and background pathways.

[0020] In a possible implementation, determining an initial population based on a molecular herbal database includes:

[0021] Determine the number of good parties, and determine the group size parameter based on the number of good parties;

[0022] Determine the number of compatible Chinese medicines, and determine the individual length parameters based on the number of compatible Chinese medicines;

[0023] The initial population is determined based on the population size parameter and the individual length parameter.

[0024] In a possible implementation, a genetic algorithm is performed based on an initial population and a fitness function to determine a target method, including:

[0025] Perform non-dominated sorting on the initial population and determine the individual ranks;

[0026] Perform selection operations based on individual levels and fitness functions, and perform crossover and mutation operations to determine the first generation population;

[0027] Based on the initial population and the first generation population, the parent population is determined by merging;

[0028] Based on the parent population, termination conditions and fitness function, a genetic algorithm is used to determine the target solution.

[0029] In a possible implementation, the method further includes:

[0030] Based on the merger of the initial population and the first generation population, when no parent population is generated, the initial population and the first generation population are non-dominated and sorted to determine the individual level;

[0031] Determine the reference point, and select target individuals from the initial population and the first generation population based on the individual level, reference point and fitness function;

[0032] The parent population is determined based on the target individual.

[0033] In a possible implementation, a genetic algorithm is performed based on an initial population and a fitness function to determine a target method, including:

[0034] Determine a target generation number, and after performing a genetic algorithm to the target generation number based on an initial population and a fitness function, determine at least two sub-evolution processes;

[0035] A genetic algorithm is performed in parallel based on at least two sub-evolution processes to determine a target solution.

[0036] According to a second aspect of the embodiments of this specification, a good method selection device is provided, including:

[0037] a data determination module configured to determine background pathway, drug set, and patient data;

[0038] a pathway determination module configured to determine a drug disorder pathway based on a background pathway and a drug set, and to determine a disease disorder pathway based on the background pathway and patient data;

[0039] a database determination module configured to determine a molecular herbal database based on a disease disorder pathway and a drug disorder pathway;

[0040] A group determination module is configured to determine an initial group based on a molecular herbal database; wherein the initial group includes a plurality of individuals, and the individuals include a good prescription;

[0041] The good recipe selection module is configured to determine a fitness function, and to perform a genetic algorithm based on the initial population and the fitness function to determine a target good recipe; wherein the fitness function includes a quantity parameter and a multi-index parameter of the cell line.

[0042] According to a third aspect of an embodiment of this specification, a computing device is provided, including:

[0043] Memory and processor;

[0044] The memory is used to store computer executable instructions, and the processor is used to execute the computer executable instructions. When the computer executable instructions are executed by the processor, the steps of the above-mentioned good method selection method are implemented.

[0045] According to a fourth aspect of the embodiments of this specification, a computer-readable storage medium is provided, which stores computer-executable instructions, and when the instructions are executed by a processor, the steps of the above-mentioned good party selection method are implemented.

[0046] According to a fifth aspect of the embodiments of this specification, a computer program is provided, wherein when the computer program is executed in a computer, the computer is caused to execute the steps of the above-mentioned good party selection method.

[0047] The embodiments of this specification provide a method and device for selecting a good prescription, wherein the method includes: determining background pathways, drug sets and patient data; determining drug disorder pathways based on background pathways and drug sets, and determining disease disorder pathways based on background pathways and patient data; determining a molecular herbal database based on disease disorder pathways and drug disorder pathways; determining an initial population based on the molecular herbal database; wherein the initial population includes multiple individuals, and the individuals include a good prescription; determining a fitness function, performing a genetic algorithm based on the initial population and the fitness function, and determining a target good prescription, wherein the fitness function includes a number parameter and a multi-index parameter of a cell line. Screening of Chinese medicine prescriptions from multiple perspectives of reversing disordered pathways and efficacy scores. Compared with the limitations of traditional methods that rely on a single indicator for optimization screening, this multi-dimensional screening strategy can better reflect the overall efficacy and therapeutic effect of Chinese medicine prescriptions, thereby enhancing the application value. BRIEF DESCRIPTION OF THE DRAWINGS

[0048] Figure 1 is a flow chart of a method for selecting a good method provided by an embodiment of this specification;

[0049] Figure 2 It is a schematic diagram of a framework of a method for selecting a good method provided by an embodiment of this specification;

[0050] Figure 3 It is an algorithm principle diagram of a good selection method provided by an embodiment of this specification;

[0051] Figure 4 This is a schematic diagram of the probability density distribution of a single cell line result of a good selection method provided in one embodiment of this specification;

[0052] Figure 5 This is a schematic diagram of the probability density distribution of the reversal rate of multiple cell line results of a good prescription selection method provided in one embodiment of this specification;

[0053] Figure 6 This is a schematic diagram of the probability density distribution of drug efficacy results of multiple cell lines of a method for selecting a good prescription provided in one embodiment of this specification;

[0054] Figure 7 It is a structural schematic diagram of a good method selection device provided by an embodiment of this specification;

[0055] Figure 8 It is a structural block diagram of a computing device provided by an embodiment of this specification. DETAILED DESCRIPTION

[0056] Many specific details are described in the following description to facilitate a full understanding of this specification. However, this specification can be implemented in many other ways than those described herein, and those skilled in the art can make similar generalizations without violating the connotation of this specification, so this specification is not limited to the specific implementation disclosed below.

[0057] The terms used in one or more embodiments of this specification are only for the purpose of describing specific embodiments, and are not intended to limit one or more embodiments of this specification. The singular forms of "a" and "the" used in one or more embodiments of this specification and the appended claims are also intended to include plural forms, unless the context clearly indicates other meanings. It should also be understood that the term "and / or" used in one or more embodiments of this specification refers to and includes any or all possible combinations of one or more associated listed items.

[0058] It should be understood that although the terms first, second, etc. may be used to describe various information in one or more embodiments of this specification, this information should not be limited to these terms. These terms are only used to distinguish the same type of information from each other. For example, without departing from the scope of one or more embodiments of this specification, the first may also be referred to as the second, and similarly, the second may also be referred to as the first. Depending on the context, the word "if" as used herein may be interpreted as "at the time of" or "when" or "in response to determining".

[0059] First, the terms involved in one or more embodiments of this specification are explained.

[0060] Gene Set Enrichment Analysis (GSEA): is a statistical method widely used in bioinformatics research, mainly used to test the enrichment of certain functions or features in a gene set. The main purpose of enrichment analysis is to find biologically meaningful patterns and functions from a large amount of genetic data.

[0061] NGS (Next Generation Sequencing): also known as high-throughput sequencing, is characterized by high output and high resolution. It can read the sequences of hundreds of thousands to millions of DNA molecules in parallel at one time. While providing rich genetic information, it can also greatly reduce sequencing costs and shorten sequencing time.

[0062] HTS 2 :high throughput sequencing-based high throughput screening, high-throughput drug screening based on high-throughput sequencing.

[0063] Pathway: A model that describes the interactions among genes, proteins or metabolites within cells, tissues or individuals.

[0064] FDR: false discovery rate, false discovery rate: in transcriptome analysis, it is mainly used in the analysis of differentially expressed genes to control the proportion of false positive results in the final analysis results.

[0065] In this specification, a good party selection method is provided. This specification also relates to a good party selection device, a computing device, and a computer-readable storage medium, which are described in detail one by one in the following embodiments.

[0066] See also Figure 1 , Figure 1 A flowchart of a good party selection method provided according to an embodiment of the present specification is shown, which specifically includes the following steps.

[0067] Step 101: Determine background pathways, drug sets, and patient data.

[0068] Among them, the background pathway can be a pathway related to the disease; the drug set can be a set of all selected Chinese medicines used to form a good prescription; the patient data can be the gene expression spectrum data of diseased patients and healthy people; the good prescription can be a combination of Chinese medicines or foods, which can also be called a prescription.

[0069] In practical applications, HTS 2 , NGS or gene chip platform to obtain the characteristic gene expression profile of each cell after TCM treatment and perform differential gene analysis to obtain drug expression profile data. Furthermore, disease-related databases and literature are searched to collect pathways as background pathways, and gene expression profile data of disease patients and healthy people can be downloaded from published public databases as patient data.

[0070] The basic data of the embodiments of this specification are accurate, and all of these basic data are obtained through experiments rather than relying on additional inferences, thereby ensuring the accuracy of the data.

[0071] Step 102: Determine a drug disorder pathway based on background pathways and a drug set, and determine a disease disorder pathway based on background pathways and patient data.

[0072] Among them, the significantly enriched pathways were screened out according to the FDR threshold and were called dysregulated pathways.

[0073] In a possible implementation, determining a drug dysregulation pathway based on a background pathway and a drug set includes: intervening in cells through a drug set to determine drug expression profile data; and performing gene enrichment analysis based on the drug expression profile data and the background pathway to determine the drug dysregulation pathway.

[0074] In practical applications, see Figure 2 ,Based on the background pathways screened in the above steps, gene enrichment analysis of drug expression profile data was performed to obtain drug disorder pathways, namely Figure 2 GSEA analysis outputs the standardized enrichment score (NES value) and significance test value (FDR value) of each Chinese medicine on the background pathway. When the FDR value is less than the set FDR threshold, the original NES is retained. When the FDR value exceeds the set threshold, the NES value is reset to 0. The NES value can be regarded as a quantitative value of the effect on the pathway.

[0075] In a possible implementation, determining a drug disorder pathway based on a background pathway and a drug set includes: determining disease expression profile data by performing differential gene analysis based on patient data; and determining a disease disorder pathway based on the disease expression profile data and the background pathway.

[0076] In practical applications, see Figure 2 , based on the gene expression profile data of disease patients and healthy people, differential gene analysis is performed, and then Gene Set Enrichment Analysis (GSEA) is performed based on the background pathways screened out in the above steps. The significantly enriched pathways are screened out according to the FDR threshold to obtain the disease disorder pathway, that is, Figure 2After obtaining the TCM pathway spectrum and the disordered pathway spectrum, the subsequent steps can be carried out to calculate the good prescription of a single cell line or multiple cell lines.

[0077] The embodiments of this specification combine drug gene expression data to improve the scientificity and pertinence of Chinese medicine prescriptions through scientific analysis, making the drug effects more accurate and effective. And with the reversal of disease signal pathways as the goal orientation, it effectively guides the screening and optimization process of Chinese medicine prescriptions and enhances the pertinence and accuracy of drug action mechanisms.

[0078] Step 103: Determine a molecular herbal database based on the disease disorder pathway and the drug disorder pathway.

[0079] In practical applications, after the above steps, a molecular herbal database can be established based on the obtained disease disorder pathways and drug disorder pathways, so as to select a good prescription based on the molecular herbal database. See the following steps for details.

[0080] Step 104: Determine an initial group based on the molecular herbal database; wherein the initial group includes multiple individuals, and the individuals include good prescriptions.

[0081] In one possible implementation, determining an initial population based on a molecular herbal database includes: determining the number of good prescriptions, and determining a population size parameter based on the number of good prescriptions; determining the number of compatible Chinese medicines, and determining an individual length parameter based on the number of compatible Chinese medicines; and determining the initial population based on the population size parameter and the individual length parameter.

[0082] The number of good prescriptions can be the number of good prescriptions to be selected, and the number of compatible Chinese medicines can be the number of types of Chinese medicines in each good prescription.

[0083] In practical applications, see Figure 3 The genetic algorithm simulates the selection, crossover and mutation mechanisms in natural evolution, and then uses the disease molecular signaling pathways obtained from the analysis of clinical patient omics data as targets, combined with the drug cell gene expression profile data, to screen and optimize the Chinese medicine formula with the reversal of the disease signaling pathway ratio as the core.

[0084] Specifically, the population needs to be initialized first. The population size and individual length need to be determined. The population size is the number of randomly generated prescriptions, which is an integer, denoted by N; the individual length is the number of compatible Chinese medicines contained in the generated prescription, which is an integer, denoted by L.

[0085] Step 105: determine the fitness function, perform a genetic algorithm based on the initial population and the fitness function, and determine the target solution, wherein the fitness function includes the quantity parameter and the multi-index parameter of the cell line.

[0086] The target prescription may be the prescription finally screened, the quantity parameter of the cell line may be the quantity value of the cell line, and the multi-index parameter may be the quantity value of the index, which may be the reversal rate or the efficacy score.

[0087] In practical applications, genetic algorithms simulate the selection, crossover and mutation mechanisms in natural evolution, and then use the disease molecular signaling pathways obtained from the analysis of clinical patient omics data as targets, combined with drug cell gene expression profile data, to screen and optimize multi-cell line Chinese medicine prescriptions with the reversal of signal pathway ratios and drug efficacy as the core.

[0088] Specifically, the fitness value is calculated for each formulation of each cell line. The fitness function calculation is performed for the first time on the initial formulation population. After that, the same calculation is performed for each new generation formulation by applying genetic operators such as non-dominated sorting, selection, crossover and mutation.

[0089] Fitness is defined as:

[0090] fitness=[RevR1,ES1,RevR2,ES2...RevR n ,ES n ]

[0091] Where n represents a total of n cell lines, RevR n Represents the ratio of disease signaling pathways reversed by traditional Chinese medicine on the nth cell line, ES n Pharmacodynamics of traditional Chinese medicine on nth cell line.

[0092]

[0093] Where m means there are m disordered pathways, Cnes i Indicates the NES value of Chinese medicine on the ith disorder pathway, Dnes i Represents the NES value of the patient on the i-th disordered pathway.

[0094] This specification embodiment considers multiple evaluation indicators in the prescription design for the prescription of a single cell line. Compared with the traditional method of relying only on a single indicator optimization, the present invention introduces a multi-objective optimization method, which comprehensively considers multiple indicators such as reversal rate and drug efficacy score. This method can avoid the local optimal problem under a single indicator, ensure that the selected prescriptions perform well on multiple evaluation indicators, and improve the comprehensiveness of the optimization results. For multi-cell line prescriptions, it is by considering the balance of different evaluation indicators (such as drug efficacy score and reversal rate) in each cell line at the same time, to screen out a good prescription with both a higher average reversal rate and drug efficacy score, avoiding the problem that the optimization results of a single cell line are not effective in other cell lines.

[0095] In a possible implementation, a genetic algorithm is performed based on an initial population and a fitness function to determine a target good solution, including: performing a non-dominated sort on the initial population to determine individual ranks; performing a selection operation based on individual ranks and a fitness function, and performing a crossover operation and a mutation operation to determine a first-generation population; merging the initial population and the first-generation population to determine a parent population; and performing a genetic algorithm based on the parent population, a termination condition, and a fitness function to determine a target good solution.

[0096] In practical applications, see Figure 3 First, we need to perform a genetic algorithm on the initial population to generate the first generation population. Then we merge the parent generation with the child generation to perform the subsequent genetic algorithm steps. Figure 4 ,In the multi-objective genetic algorithm, the merging of parents and offspring is mainly used to form ,candidate solutions for the next generation, and retain the best individuals based on ,elite strategy in evolution.

[0097] Specifically, non-dominated sorting is an important method for determining the superiority level of solutions in multi-objective optimization. Its main idea is to divide the population into different Pareto frontiers, and the individuals in the frontiers do not dominate each other. The formula is as follows:

[0098]

[0099] Where I represents an indicator function, n(x i ) represents the dominant solution x i The number of

[0100] The embodiments of this specification adopt a non-dominated sorting mechanism, which divides the population into different Pareto frontiers, so that the selected prescriptions can better balance multiple indicators in multi-objective optimization. The core of non-dominated sorting is to avoid the degradation of other indicators caused by over-optimization of a single objective by finding solutions that perform well on multiple objectives, thereby ensuring the comprehensive performance of the prescription.

[0101] Furthermore, the selection operation can use the tournament selection method. The basic idea of ​​the algorithm is to randomly select two individuals, compare their non-dominated ranks, and select the individual with the lower rank; if the non-dominated ranks are the same, select the individual closest to the reference plane; if they are the same, select randomly. Its mathematical expression is as follows:

[0102]

[0103] Where R(x) represents the non-dominated rank of individual x, and d(x) represents the distance between the individual and the reference plane.

[0104] Furthermore, the crossover operation introduces simulated binary crossover, which imitates the single-point crossover behavior under binary coding, thereby effectively generating new solutions with diversity and exploration capabilities. The calculation formula is as follows:

[0105]

[0106] Where u is a uniformly distributed random number, η c represents the crossover distribution index, β is the crossover distribution index, x represents the parent generation, and c represents the offspring generation.

[0107] Furthermore, the mutation operation introduces polynomial variation, which increases the diversity of the population by making small adjustments to individuals. The formula is as follows:

[0108]

[0109] where δ i is the coefficient of variation, u is a uniformly distributed random number, η m is the distribution index of variation, x i ′ is the individual after mutation, x i It is an individual before compilation.

[0110] The embodiments of this specification introduce a multi-objective collaborative evolution algorithm to construct a multi-objective multi-cell line evaluation system, which breaks through the limitations brought by a single cell line. When screening good prescriptions, the good prescriptions can be evaluated from different dimensions, thereby improving the scientificity and accuracy of the screening results.

[0111] In a possible implementation, the method further includes: merging the initial population and the first-generation population, and performing non-dominated sorting on the initial population and the first-generation population to determine individual ranks when no parent population is generated; determining a reference point, and selecting a target individual from the initial population and the first-generation population based on the individual rank, the reference point and the fitness function; and determining the parent population based on the target individual.

[0112] In practical applications, when a new parent population cannot be generated by merging offspring with parents, non-dominated sorting and reference point sorting are still required to select excellent individuals to form a new parent population.

[0113] The embodiments of this specification also introduce a reference point guidance mechanism, which guides the population to evolve towards the optimal solution area by setting a reference point. Its main function is to ensure that the generated solutions can be concentrated in the potential optimal area in multi-objective optimization, improve the optimization efficiency of the algorithm, reduce the randomness in the evolution process, and make the screened prescriptions more targeted and practical.

[0114] Specifically, reference points are introduced to guide the selection operation and maintain a uniformly distributed Pareto solution in the high-dimensional space. The uniform distribution of reference points in the target space is usually generated by the following formula:

[0115] i1+i2+...+i k =H

[0116]

[0117] where i k represents the discretization value of a target (i.e., the kth target) in the target space, representing the relative weight or proportion of the target; H represents the number of discretization steps, which is used to control the precision of discretization; r j represents the jth reference point, which is a k-dimensional vector, indicating the value of the reference point on each objective function.

[0118] The number of reference points is calculated using the combined formula:

[0119]

[0120] Where N r represents the total number of reference points, n represents the number of objective functions, and H represents the number of discretization steps.

[0121] In the prescription screening process, the embodiment of this specification adopts reversal rate and efficacy score as core screening indicators for multiple cell lines. These two indicators are the key criteria for measuring whether the candidate prescription is close to the ideal state. The calculation method of reversal rate and efficacy score is shown in the above fitness function. The screening method is based on a multi-objective genetic algorithm, which can efficiently select the best comprehensive performance solution from a large number of candidate prescriptions of multiple cell lines by simulating the process of natural selection and genetic variation. This multi-objective evolutionary algorithm not only greatly improves the screening efficiency, can quickly process a large number of candidate prescriptions, but also ensures the quality and practicality of the screening results, and ensures that the final screened prescription has excellent comprehensive performance in both reversal rate and efficacy score. The advantage of this method is that it can balance various indicators under the screening environment of complex multiple cell lines, making the screening process more accurate and reliable.

[0122] In one possible implementation, a genetic algorithm is performed based on an initial population and a fitness function to determine a target solution, including: determining a target generation, after performing the genetic algorithm based on the initial population and the fitness function to the target generation, determining at least two sub-evolution processes; and performing the genetic algorithm in parallel based on the at least two sub-evolution processes to determine the target solution.

[0123] In practical applications, as the number of evolutionary generations increases, even after multiple iterations and selections, some characteristics (such as reversal rate) may not change significantly, which may cause the algorithm to fall into a local optimal solution, limiting the diversity and innovation of the results. The present invention proposes a parallel evolution technology to overcome the above limitations by processing multiple independent evolutionary processes in parallel.

[0124] The embodiments of this specification introduce genetic algorithms and parallel evolution methods to accelerate the screening process of traditional Chinese medicine prescriptions, while ensuring comprehensiveness and accuracy, and greatly improving efficiency.

[0125] In one embodiment, a single-objective genetic algorithm and a multi-objective genetic algorithm based on a single cell line are compared, and a single cell line is used to match 500 single drugs with 9, 10, and 11 drugs. Single-objective genetic algorithms and multi-objective genetic algorithms are used for formula matching and good prescription screening. The single-objective genetic algorithm evolves only with one goal, the reversal rate, while the multi-objective genetic algorithm considers both the reversal rate and the efficacy score at the same time to achieve their co-evolution. The results of the top 100 reversal rates are selected for statistics. The running results are detailed in Table 1, and the probability density distribution is detailed in Table 1. Figure 4 , it can be seen from the probability density distribution diagram that the optimization results of the multi-objective genetic algorithm are better than those of the single-indicator genetic algorithm in all indicators. When evaluating the reversal rate of drug combinations, for prescriptions of different sizes (9 drugs, 10 drugs, and 11 drugs), the probability density distribution of the multi-objective genetic algorithm is significantly better than that of the single-indicator genetic algorithm, showing a higher reversal rate distribution. In addition, in terms of drug efficacy scores, the probability density distribution of the multi-objective genetic algorithm is also significantly better than that of the single-objective genetic algorithm, showing a higher drug efficacy score density distribution.

[0126] Table 1

[0127]

[0128] It can be seen from the above table that the average reversal rate and efficacy score of the multi-objective genetic algorithm are better than those of the single-objective genetic algorithm.

[0129] In another embodiment, based on the exhaustive algorithm and the algorithm of the present invention, prescription compatibility and good prescription screening are performed in multiple cell lines, and the evaluation index is the reversal rate. For two cell lines, 476 single-drug combinations of 6 flavors and 7 flavors were respectively matched. When using the same computing power resources, the multi-cell line single-target exhaustive method can only complete the exhaustive prescription of one cell line, and then evaluate its prescription results in other cell lines; the algorithm of the present invention is aimed at the reversal rate of two cell lines, and a total of two targets are simultaneously evolved. The operation results are shown in Table 2.

[0130] Table 2

[0131]

[0132] It can be clearly seen from the data in the above table that the multi-cell line single-objective genetic algorithm is superior to the multi-cell line single-objective exhaustive method in terms of computational efficiency and result quality. a. In terms of running time, the multi-cell line single-objective genetic algorithm significantly reduces the computing time, especially when the number of medicinal flavors is large, the speed increase is more obvious. This improvement is due to the intelligent search strategy of the multi-cell line single-objective genetic algorithm, which effectively reduces the search space by simulating natural selection and evolutionary processes, avoiding the inefficient operation of enumerating each possible combination one by one in the multi-cell line single-objective exhaustive method. b. In terms of result quality, the multi-cell line single-objective genetic algorithm also shows significant advantages. According to the average maximum reversal rate data, the algorithm can not only generate results quickly, but also obtain a higher average maximum reversal rate of prescription results, indicating that it has better performance in optimizing the reversal rate. In addition, when the number of medicinal flavors reaches a large scale, the limitations of the multi-cell line single-objective exhaustive method are more obvious.

[0133] For the study of multiple cell lines, the core goal of this program is to find a drug combination that has a high reversal rate and maintains a good drug efficacy score in as many cell lines as possible. In this embodiment, for two cell lines, 300 single-drug combinations of 6 herbs are matched, and the exhaustive method and the algorithm of the present invention are used to perform prescription matching and good prescription screening. When using the same computing power resources, the multi-cell line multi-objective exhaustive method only exhaustively formulates one cell line, and then evaluates the results that meet the requirements of the cell line; the multi-cell line multi-objective genetic algorithm is for the reversal rate and drug efficacy score of the two cell lines, a total of four goals for simultaneous evolutionary calculations, the operation results are shown in Table 3, probability density statistical analysis, and related results are as follows. Figure 6 , Figure 7 shown.

[0134] Table 3

[0135]

[0136] From the above table, we can see that the overall results of the multi-cell lineage multi-objective genetic algorithm are better than the multi-cell lineage multi-objective exhaustive method.

[0137] from Figure 5 , Figure 6It can be seen that in the results of the multi-objective optimization algorithm, whether it is the average reversal rate, the reversal rate of cell line 1, the reversal rate of cell line 2, or the average efficacy score, the efficacy score of cell line 1, and the efficacy score of cell line 2, the distributions are very close, indicating that the algorithm can effectively balance these objectives when dealing with multiple cell lines and provide stable results. In contrast, the drug combination using the exhaustive method has a large difference in the distribution of these indicators, especially there is a significant fluctuation between the reversal rate and efficacy score of cell line 1 and cell line 2. This shows that the exhaustive method is difficult to achieve a balance between different cell lines when dealing with complex multi-objective optimization problems, resulting in a strong discreteness of the results.

[0138] The embodiments of this specification provide a method and device for selecting a good prescription, wherein the method includes: determining background pathways, drug sets and patient data; determining drug disorder pathways based on background pathways and drug sets, and determining disease disorder pathways based on background pathways and patient data; determining a molecular herbal database based on disease disorder pathways and drug disorder pathways; determining an initial population based on the molecular herbal database; wherein the initial population includes multiple individuals, and the individuals include a good prescription; determining a fitness function, performing a genetic algorithm based on the initial population and the fitness function, and determining a target good prescription, wherein the fitness function includes a number parameter and a multi-index parameter of a cell line. Screening of Chinese medicine prescriptions from multiple perspectives of reversing disordered pathways and efficacy scores. Compared with the limitations of traditional methods that rely on a single indicator for optimization screening, this multi-dimensional screening strategy can better reflect the overall efficacy and therapeutic effect of Chinese medicine prescriptions, thereby enhancing the application value.

[0139] Corresponding to the above method embodiment, this specification also provides a method selection device embodiment. Figure 7 FIG. 2 shows a schematic diagram of a structure of a device for selecting a good method provided by an embodiment of the present specification. Figure 7 As shown, the device comprises:

[0140] A data determination module 701 is configured to determine background pathway, drug set, and patient data;

[0141] a pathway determination module 702 configured to determine a drug disorder pathway based on background pathways and a drug set, and to determine a disease disorder pathway based on background pathways and patient data;

[0142] A database determination module 703 is configured to determine a molecular herbal database based on disease disorder pathways and drug disorder pathways;

[0143] The group determination module 704 is configured to determine an initial group based on the molecular herbal database; wherein the initial group includes a plurality of individuals, and the individuals include a good prescription;

[0144] The recipe selection module 705 is configured to determine a fitness function, and to perform a genetic algorithm based on the initial population and the fitness function to determine a target recipe, wherein the fitness function includes a quantity parameter and a multi-index parameter of the cell line.

[0145] In a possible implementation, the path determination module 702 is further configured to:

[0146] Determine drug expression profile data by intervening cells with a drug collection;

[0147] Gene enrichment analysis based on drug expression profile data and background pathways was performed to identify drug dysregulation pathways.

[0148] In a possible implementation, the path determination module 702 is further configured to:

[0149] Perform differential gene analysis based on patient data to determine disease expression profile data;

[0150] Identify disease dysregulated pathways based on disease expression profiling data and background pathways.

[0151] In a possible implementation, the group determination module 704 is further configured to:

[0152] Determine the number of good parties, and determine the group size parameter based on the number of good parties;

[0153] Determine the number of compatible Chinese medicines, and determine the individual length parameters based on the number of compatible Chinese medicines;

[0154] The initial population is determined based on the population size parameter and the individual length parameter.

[0155] In a possible implementation, the good party selection module 705 is further configured to:

[0156] Perform non-dominated sorting on the initial population and determine the individual ranks;

[0157] Perform selection operations based on individual levels and fitness functions, and perform crossover and mutation operations to determine the first generation population;

[0158] Based on the initial population and the first generation population, the parent population is determined by merging;

[0159] Based on the parent population, termination conditions and fitness function, a genetic algorithm is used to determine the target solution.

[0160] In a possible implementation, the good party selection module 705 is further configured to:

[0161] Based on the merger of the initial population and the first generation population, when no parent population is generated, the initial population and the first generation population are non-dominated and sorted to determine the individual level;

[0162] Determine the reference point, and select target individuals from the initial population and the first generation population based on the individual level, reference point and fitness function;

[0163] The parent population is determined based on the target individual.

[0164] In a possible implementation, the good party selection module 705 is further configured to:

[0165] Determine a target generation number, and after performing a genetic algorithm to the target generation number based on an initial population and a fitness function, determine at least two sub-evolution processes;

[0166] A genetic algorithm is performed in parallel based on at least two sub-evolution processes to determine a target solution.

[0167] The embodiments of this specification provide a method and device for selecting a good prescription, wherein the device includes: determining background pathways, drug sets and patient data; determining drug disorder pathways based on background pathways and drug sets, and determining disease disorder pathways based on background pathways and patient data; determining a molecular herbal database based on disease disorder pathways and drug disorder pathways; determining an initial population based on the molecular herbal database; wherein the initial population includes multiple individuals, and the individuals include a good prescription; determining a fitness function, performing a genetic algorithm based on the initial population and the fitness function, and determining a target good prescription, wherein the fitness function includes a number parameter and a multi-index parameter of a cell line. Screening of Chinese medicine prescriptions from multiple perspectives of reversing disordered pathways and efficacy scores. Compared with the limitations of traditional methods that rely on a single indicator for optimization screening, this multi-dimensional screening strategy can better reflect the overall efficacy and therapeutic effect of Chinese medicine prescriptions, thereby enhancing the application value.

[0168] The above is a schematic scheme of a good recipe selection device of this embodiment. It should be noted that the technical scheme of the good recipe selection device and the technical scheme of the good recipe selection method mentioned above belong to the same concept, and the details of the technical scheme of the good recipe selection device that are not described in detail can all be referred to the description of the technical scheme of the good recipe selection method mentioned above.

[0169] Figure 8 The block diagram of a computing device 800 according to an embodiment of the present specification is shown. The components of the computing device 800 include but are not limited to a memory 810 and a processor 820. The processor 820 is connected to the memory 810 via a bus 830, and a database 850 is used to store data.

[0170] The computing device 800 also includes an access device 840 that enables the computing device 800 to communicate via one or more networks 860. Examples of these networks include a public switched telephone network (PSTN), a local area network (LAN), a wide area network (WAN), a personal area network (PAN), or a combination of communication networks such as the Internet. The access device 840 may include one or more of any type of network interface (e.g., a network interface card (NIC)) that is wired or wireless, such as an IEEE 802.11 wireless local area network (WLAN) wireless interface, a world-wide interoperability for microwave access (Wi-MAX) interface, an Ethernet interface, a universal serial bus (USB) interface, a cellular network interface, a Bluetooth interface, and a near field communication (NFC).

[0171] In one embodiment of the present specification, the above components of the computing device 800 and Figure 8 Other components not shown in the figure may also be connected to each other, for example, via a bus. It should be understood that Figure 8 The computing device structure block diagram shown is only for the purpose of illustration, and is not intended to limit the scope of this specification. Those skilled in the art can add or replace other components as needed.

[0172] The computing device 800 may be any type of stationary or mobile computing device, including a mobile computer or mobile computing device (e.g., a tablet computer, a personal digital assistant, a laptop computer, a notebook computer, a netbook, etc.), a mobile phone (e.g., a smart phone), a wearable computing device (e.g., a smart watch, smart glasses, etc.), or other types of mobile devices, or a stationary computing device such as a desktop computer or a personal computer (PC). The computing device 800 may also be a mobile or stationary server.

[0173] The processor 820 is used to execute the following computer executable instructions, which implement the steps of the above-mentioned good selection method when executed by the processor. The above is a schematic scheme of a computing device of this embodiment. It should be noted that the technical scheme of the computing device and the technical scheme of the above-mentioned good selection method belong to the same concept, and the details not described in detail in the technical scheme of the computing device can be referred to the description of the technical scheme of the above-mentioned good selection method.

[0174] An embodiment of the present specification also provides a computer-readable storage medium storing computer-executable instructions, which can implement the steps of the above-mentioned method for selecting a good method when executed by a processor.

[0175] The above is a schematic scheme of a computer-readable storage medium of this embodiment. It should be noted that the technical scheme of the storage medium and the technical scheme of the above-mentioned good party selection method belong to the same concept, and the details not described in detail in the technical scheme of the storage medium can be referred to the description of the technical scheme of the above-mentioned good party selection method.

[0176] An embodiment of the present specification further provides a computer program, wherein when the computer program is executed in a computer, the computer is caused to execute the steps of the above-mentioned good method for selecting a good solution.

[0177] The above is an illustrative solution of a computer program of this embodiment. It should be noted that the technical solution of the computer program and the technical solution of the above-mentioned good choice method belong to the same concept, and the details not described in detail in the technical solution of the computer program can be referred to the description of the technical solution of the above-mentioned good choice method.

[0178] The above is a description of a specific embodiment of the specification. Other embodiments are within the scope of the appended claims. In some cases, the actions or steps recorded in the claims can be performed in an order different from that in the embodiments and still achieve the desired results. In addition, the processes depicted in the drawings do not necessarily require the specific order or continuous order shown to achieve the desired results. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0179] The computer instructions include computer program codes, which may be in source code form, object code form, executable files or some intermediate forms, etc. The computer-readable medium may include: any entity or device capable of carrying the computer program code, recording medium, USB flash drive, mobile hard disk, magnetic disk, optical disk, computer memory, read-only memory (ROM), random access memory (RAM), electric carrier signal, telecommunication signal and software distribution medium, etc. It should be noted that the content contained in the computer-readable medium may be appropriately increased or decreased according to the requirements of legislation and patent practice in the jurisdiction. For example, in some jurisdictions, according to legislation and patent practice, computer-readable media do not include electric carrier signals and telecommunication signals.

[0180] It should be noted that, for the convenience of description, the aforementioned method embodiments are all described as a series of action combinations, but those skilled in the art should be aware that the embodiments of this specification are not limited by the order of the actions described, because according to the embodiments of this specification, some steps can be performed in other orders or simultaneously. Secondly, those skilled in the art should also be aware that the embodiments described in the specification are all preferred embodiments, and the actions and modules involved are not necessarily required for the embodiments of this specification.

[0181] In the above embodiments, the description of each embodiment has its own emphasis. For parts that are not described in detail in a certain embodiment, reference can be made to the relevant descriptions of other embodiments.

[0182] The preferred embodiments of this specification disclosed above are only used to help explain this specification. The optional embodiments do not describe all the details in detail, nor do they limit the invention to only the specific implementation methods described. Obviously, many modifications and changes can be made according to the content of the embodiments of this specification. This specification selects and specifically describes these embodiments in order to better explain the principles and practical applications of the embodiments of this specification, so that technicians in the relevant technical field can well understand and use this specification. This specification is only limited by the claims and their full scope and equivalents.

Claims

1. A method for selecting a good prescription, characterized in that: include: Identify background pathways, drug collections, and patient data; determining a drug disorder pathway based on the background pathway and the drug set, and determining a disease disorder pathway based on the background pathway and the patient data; Determining a molecular herbal database based on the disease disorder pathway and the drug disorder pathway; Determine an initial population based on the molecular herbal database; wherein the initial population includes a plurality of individuals, and the individuals include Shan Fang; Determine a fitness function, perform a genetic algorithm based on the initial population and the fitness function, and determine a target solution; wherein the fitness function includes a quantity parameter and a multi-index parameter of the cell line.

2. The method according to claim 1, characterized in that The determining of a drug disorder pathway based on the background pathway and the drug set comprises: Intervening cells with the drug set to determine drug expression profile data; Gene enrichment analysis is performed based on the drug expression profile data and the background pathway to determine the drug disorder pathway.

3. The method according to claim 1, characterized in that The determining of a drug disorder pathway based on the background pathway and the drug set comprises: Perform differential gene analysis based on patient data to determine disease expression profile data; A disease dysregulated pathway is determined based on the disease expression profile data and the background pathway.

4. The method according to claim 1, characterized in that: The determining of the initial population based on the molecular herbal database comprises: determining a number of good parties, and determining a population size parameter based on the number of good parties; Determine the number of Chinese medicines to be combined, and determine the individual length parameters based on the number of Chinese medicines to be combined; An initial population is determined based on the population size parameter and the individual length parameter.

5. The method according to claim 1, characterized in that The performing of a genetic algorithm based on the initial population and the fitness function to determine a target fitness method includes: Performing non-dominated sorting on the initial population to determine individual ranks; Performing a selection operation based on the individual level and the fitness function, and performing a crossover operation and a mutation operation to determine a first generation population; Determine a parent population based on merging the initial population and the first generation population; A genetic algorithm is performed based on the parent population, the termination condition and the fitness function to determine a target solution.

6. The method according to claim 5, characterized in that Also includes: Based on the merging of the initial population and the first generation population, when the parent population is not generated, performing non-dominated sorting on the initial population and the first generation population to determine the individual ranks; Determine a reference point, and select a target individual from the initial population and the first generation population based on the individual level, the reference point and the fitness function; A parent population is determined based on the target individual.

7. The method according to claim 1, characterized in that The performing of a genetic algorithm based on the initial population and the fitness function to determine a target fitness method includes: Determining a target generation number, after performing the genetic algorithm based on the initial population and the fitness function to the target generation number, determining at least two sub-evolution processes; A genetic algorithm is performed in parallel based on the at least two sub-evolution processes to determine a target solution.

8. A good prescription selection device, characterized in that: include: a data determination module configured to determine background pathway, drug set, and patient data; a pathway determination module configured to determine a drug disorder pathway based on the background pathway and the drug set, and to determine a disease disorder pathway based on the background pathway and the patient data; a database determination module, configured to determine a molecular herbal database based on the disease disorder pathway and the drug disorder pathway; A population determination module is configured to determine an initial population based on the molecular herbal database; wherein the initial population includes a plurality of individuals, and the individuals include Shan Fang; The good recipe selection module is configured to determine a fitness function, and to perform a genetic algorithm based on the initial population and the fitness function to determine a target good recipe; wherein the fitness function includes a quantity parameter and a multi-index parameter of a cell line.

9. A computing device, characterized in that include: Memory and processor; The memory is used to store computer-executable instructions, and the processor is used to execute the computer-executable instructions. When the computer-executable instructions are executed by the processor, the steps of the good party selection method described in any one of claims 1 to 7 are implemented.

10. A computer-readable storage medium storing computer-executable instructions, which, when executed by a processor, implement the steps of the method for selecting a better party as claimed in any one of claims 1 to 7.

Images