Personalized drug treatment scheme making system

Abstract

The invention discloses a personalized drug treatment scheme making system, which relates to the technical field of personalized treatment scheme making, and comprises a drug treatment data acquisition module, a drug treatment effect evaluation module, a drug treatment scheme making module and a drug treatment scheme output module, basic treatment data and gene detection data are collected, and the collected data are preprocessed; the drug curative effect evaluation module is used for evaluating drug metabolism and drug curative effect on the basis of basic treatment data and gene detection data; the drug treatment scheme output module is used for formulating a personalized drug treatment scheme table in combination with an output result of the drug treatment scheme formulating module. A data acquisition technology, a gene detection technology, a neural network algorithm, a linear regression algorithm and a convolutional neural network algorithm in the method are closely combined with a modern information technology; the problem that in the prior art, it is difficult to personally formulate a drug treatment scheme by combining multiple physiological factors is solved.

Description

Technical Field

[0001] The present invention relates to the technical field of personalized treatment plan formulation, and specifically relates to a system for formulating a personalized drug treatment plan. Background Art

[0002] Traditional drug treatments follow a unified standard plan, formulating drug treatment plans based on common disease manifestations and past experience. However, this general approach has significant drawbacks. Since patients vary widely in many physiological factors, their responses to the same drug are also different. With the development of big data analysis and artificial intelligence algorithms, it has become a reality to efficiently process massive amounts of medical data, integrate multi-dimensional data of patients, deeply mine and analyze these multi-dimensional data, and accurately predict the efficacy and safety of drugs in different individuals. Nowadays, the requirements for accuracy in clinical treatment are continuously rising, and the traditional model can no longer meet the needs. Against this background, a system for formulating a personalized drug treatment plan has emerged, providing strong support for realizing precision medicine;

[0003] Although the prior art has made great progress in the direction of formulating drug treatment plans, there are still some problems to be optimized. In the process of formulating drug treatment plans, it is difficult for the prior art to formulate drug treatment plans personalized by combining various factors, resulting in poor treatment effects of drug treatment plans. Therefore, how to combine basic treatment data and gene detection data to improve the precision of drug treatment plans is the problem we need to solve. For this purpose, a system for formulating a personalized drug treatment plan is proposed herein. Summary of the Invention

[0004] To achieve the above objectives, the present invention is realized through the following technical solutions: A system for formulating a personalized drug treatment plan includes a drug treatment data collection module, a drug efficacy evaluation module, a drug treatment plan formulation module, and a drug treatment plan output module, wherein each module is communicatively connected;

[0005] The drug treatment data collection module collects basic treatment data and gene detection data, and preprocesses the collected data, providing data preparation for the realization of the functions of subsequent modules;

[0006] The drug efficacy evaluation module evaluates drug metabolism and drug efficacy based on the basic treatment data and gene detection data;

[0007] The drug treatment plan formulation module formulates a drug treatment plan based on the evaluation results of drug metabolism and drug efficacy;

[0008] The drug treatment plan output module formulates a personalized drug treatment plan table in combination with the output result of the drug treatment plan formulation module, solving the problem that in the process of formulating a drug treatment plan, it is difficult for the prior art to formulate a drug treatment plan personalized by combining multiple factors, resulting in poor treatment effects of the drug treatment plan.

[0009] A further improvement of the technical solution of the present invention lies in that: the drug efficacy evaluation module and the drug treatment plan formulation module specifically include:

[0010] The drug efficacy evaluation module is divided into a basic evaluation unit, a gene evaluation unit and a comprehensive efficacy evaluation unit. Among them, the basic evaluation unit obtains a drug basic metabolism index through basic treatment data; the gene evaluation unit obtains a drug gene metabolism index and a drug efficacy index respectively through gene detection data; the comprehensive efficacy evaluation unit constructs a drug treatment index evaluation model by combining the drug basic metabolism index, the drug gene metabolism index and the drug efficacy index;

[0011] The drug treatment plan formulation module is divided into a drug dose allocation unit and a drug treatment plan formulation unit. Among them, the drug dose allocation unit obtains a drug dose allocation index by using the output result of the drug dose allocation model, and then constructs a drug dose allocation model; the drug treatment plan formulation unit obtains a drug treatment plan by combining the output result of the drug dose allocation model, and then constructs a drug treatment plan model.

[0012] A further improvement of the technical solution of the present invention lies in that: the drug treatment data acquisition module collects basic treatment data and gene detection data, and the process of preprocessing the collected data includes:

[0013] Deploy collection devices to collect basic treatment data and gene detection data. Among them, the collection devices include medical height and weight scales, gene sequencers, etc.;

[0014] The basic treatment data is the age, gender, height, weight and BMI value of the treatment object; the gene detection data is the metabolic enzyme gene and drug target gene of the treatment object;

[0015] Collect the age and gender of the treatment object through a questionnaire survey, collect the height, weight and BMI value of the treatment object by using a medical height and weight scale, and collect the metabolic enzyme gene and drug target gene of the treatment object by using a gene sequencer;

[0016] Data cleaning and calibration are performed on the age, gender, height, weight, and BMI values of the treated subjects. Using the FastQC tool, quality assessment is carried out on the metabolic enzyme genes and drug target genes of the treated subjects. Through the Cutadapt tool, low-quality bases in the metabolic enzyme genes and drug target genes are removed. The GATK BaseRecalibrator is used to correct the quality values of the bases in the metabolic enzyme genes and drug target genes.

[0017] A further improvement of the technical solution of the present invention lies in that: the process of the basic evaluation unit for obtaining the drug basic metabolic index includes:

[0018] When the age of the treated subject is between 0 and 18 years old, an age coefficient of 1 is assigned; when the age of the treated subject is between 18 and 44 years old, an age coefficient of 0.95 is assigned; when the age of the treated subject is between 45 and 64 years old, an age coefficient of 0.9 is assigned; when the age of the treated subject is greater than 65 years old, an age coefficient of 0.8 is assigned;

[0019] When the treated subject is male, a gender coefficient of 1 is assigned; when the treated subject is female, a gender coefficient of 0.9 is assigned;

[0020] When the BMI value of the treated subject is between 18.5 and 24.9, a BMI coefficient of 1 is assigned; when the BMI value of the treated subject is between 25 and 29.9, a BMI coefficient of 1.1 is assigned; when the BMI value of the treated subject is greater than 29.9, a BMI coefficient of 1.2 is assigned;

[0021] Using the age, height, and weight of the treated subject, according to the Mifflin-St Jeor equation, the BMR value of the treated subject is calculated. By calculating the product of the BMR value of the treated subject, the age coefficient of the treated subject, the gender coefficient of the treated subject, and the BMI coefficient of the treated subject, the drug basic metabolic index is obtained.

[0022] A further improvement of the technical solution of the present invention lies in that: the process of the gene evaluation unit for obtaining the drug gene metabolic index and the drug efficacy index includes:

[0023] Referring to the PharmGKB database, metabolic enzyme genes related to drug metabolism are screened out. According to the drug action mechanism, drug target genes of the treated subjects related to drug efficacy are screened out. Using a nucleic acid extractor and a magnetic bead kit, the magnetic bead method is used to extract the mRNA in the screened metabolic enzyme genes and drug target genes respectively;

[0024] By fluorescence quantitative PCR, detect the mRNA in the metabolic enzyme gene and the mRNA in the drug target gene, and record the cycle numbers corresponding to the inflection points where the fluorescence signals of the mRNA in the metabolic enzyme gene and the mRNA in the drug target gene enter the exponential growth phase from the background. These cycle numbers are the Ct values of the metabolic enzyme gene and the drug target gene respectively;

[0025] Extract the reference gene data related to the metabolic enzyme gene and the drug target gene from the ICG database. Using the extracted reference gene data, calculate the Ct value of the reference gene. According to the Ct value of the reference gene, calculate the ratio of the Ct value of the metabolic enzyme gene to the Ct value of the reference gene and the ratio of the Ct value of the drug target gene to the Ct value of the reference gene respectively, and obtain the relative expression levels of the metabolic enzyme gene and the drug target gene;

[0026] Set the metabolic baseline value. When the relative expression level of the metabolic enzyme gene is lower than 0.5 times the metabolic baseline value, the drug gene metabolism index is 0.8, indicating weak metabolism of the drug by the metabolic enzyme gene; when the relative expression level of the metabolic enzyme gene is between 0.5 times and 1.5 times the metabolic baseline value, the drug gene metabolism index is 1.5, indicating moderate metabolism of the drug by the metabolic enzyme gene; when the relative expression level of the metabolic enzyme gene is higher than 1.5 times the metabolic baseline value, the drug gene metabolism index is 2.5, indicating strong metabolism of the drug by the metabolic enzyme gene, and obtain the drug gene metabolism index;

[0027] Set the efficacy baseline value. When the relative expression level of the drug target gene is lower than 0.5 times the efficacy baseline value, the drug efficacy index is 1, indicating low efficacy of the drug against the drug target gene; when the relative expression level of the drug target gene is between 0.5 times and 2 times the efficacy baseline value, the drug efficacy index is 2, indicating general efficacy of the drug against the drug target gene; when the relative expression level of the drug target gene is higher than 2 times the efficacy baseline value, the drug efficacy index is 3, indicating high efficacy of the drug against the drug target gene, and obtain the drug efficacy index.

[0028] A further improvement of the technical solution of the present invention lies in: for the efficacy comprehensive evaluation unit, the construction process of the drug treatment index evaluation model includes:

[0029] According to the contributions of the drug basic metabolism index and the drug gene metabolism index to drug metabolism, assign weights to the drug basic metabolism index and the drug gene metabolism index respectively, and calculate the drug metabolism index by combining the weights. The calculation formula is , where U is the drug metabolism index, and are the weights of the drug basic metabolism index and the drug gene metabolism index respectively, and They are the drug basic metabolism index and the drug gene metabolism index respectively;

[0030] Take the basic treatment data, gene detection data, drug metabolism index and drug efficacy index as the first data set, and divide them into a training set and a test set according to the ratio of 7:3;

[0031] Combined with the training set data and the neural network algorithm, take the basic treatment data and gene detection data as inputs, and the drug metabolism index and drug efficacy index as outputs, learn the non-linear relationship between the basic treatment data, the metabolic enzyme genes of the treatment object and the drug metabolism index, and the non-linear relationship between the drug target genes and the drug efficacy index, and train the drug treatment index evaluation model;

[0032] Input the test set data into the drug treatment index evaluation model, use the MSE function to evaluate the error between the drug metabolism index output by the drug treatment index evaluation model and the actual drug metabolism index, and the error between the drug treatment index output by the drug treatment index evaluation model and the actual drug efficacy index. According to the evaluation results, adjust the parameters of the drug treatment index evaluation model, optimize the performance of the drug treatment index evaluation model, and deploy the optimized drug treatment index evaluation model into the system.

[0033] A further improvement of the technical solution of the present invention lies in that: the process of obtaining the drug dose allocation index by the drug dose allocation unit includes:

[0034] Input the basic treatment data and gene detection data into the drug treatment index evaluation model, and output the drug metabolism index through the drug treatment index evaluation model;

[0035] Take the drug metabolism index as the second data set, and divide it into a training set and a test set according to the ratio of 8:2;

[0036] Use the training set data and the linear regression algorithm, take the drug metabolism index as the input, and the drug dose allocation index as the output, adjust the intercept term and the regression coefficient of the drug metabolism index, learn the non-linear relationship between the drug metabolism index and the drug dose allocation index, and train the drug dose allocation model;

[0037] Input the test set data into the drug dose allocation model, evaluate the performance of the drug dose allocation model, adjust the intercept term and regression coefficient of the drug dose allocation model, optimize the drug dose allocation model, and deploy the optimized drug dose allocation model into the system. The expression of this drug dose allocation model is , where is the drug dose allocation index, is the intercept term, is the regression coefficient of the drug metabolism index, p is the drug metabolism index, is the error term.

[0038] A further improvement of the technical solution of the present invention lies in that: the process of the drug treatment plan formulation unit obtaining the drug treatment plan by combining the output result of the drug dose allocation model includes:

[0039] Input the drug metabolism index into the drug dose allocation model, and output the drug dose allocation index through the drug dose allocation model; input the gene detection data into the drug treatment index evaluation model, and output the drug efficacy index through the drug treatment index evaluation model;

[0040] When the drug dose allocation index is between 0 and 3, allocate low-dose drugs to the treatment object; when the drug dose allocation index is between 3 and 6, allocate conventional-dose drugs to the treatment object; when the drug dose allocation index is between 6 and 10, allocate high-dose drugs to the treatment object;

[0041] When the drug efficacy index is between 0 and 2, it indicates that the corresponding drug has poor efficacy, and drug treatment is replaced; when the drug efficacy index is between 2 and 6, it indicates that the corresponding drug has average efficacy, and combination of similar drugs and other drugs is used for treatment; when the drug efficacy index is between 6 and 10, it indicates that the corresponding drug has obvious efficacy, and continue to use similar drugs for treatment;

[0042] Combine the drug dose allocation plan and the drug type allocation plan to obtain the drug treatment plan, and encode the drug treatment plan.

[0043] A further improvement of the technical solution of the present invention lies in that: the process of constructing the drug treatment plan model by the drug treatment plan formulation unit includes:

[0044] Use the drug dose allocation index, the drug efficacy index, and the drug treatment plan code as the third data set, and divide it into a training set and a test set according to a ratio of 8:2;

[0045] Use the training set data and the neural network algorithm, take the drug dose allocation index and the drug efficacy index as inputs, and take the drug treatment plan code as the output, learn the non-linear relationship between the drug dose allocation index, the drug efficacy index and the drug treatment plan code, and train the drug treatment plan model;

[0046] Input the test set data into the drug treatment plan model, compare the actual drug treatment plan code with the drug treatment plan code output by the drug treatment plan model, evaluate the performance of the drug treatment plan model, use the Adam optimizer to update the parameters of the convolutional neural network model, optimize the drug treatment plan model, and deploy the optimized drug treatment plan model to the system.

[0047] A further improvement of the technical solution of the present invention lies in that: the process of formulating a personalized drug treatment plan table by the drug treatment plan formulating module in combination with the output result of the drug treatment plan formulating module includes:

[0048] Input the drug dose allocation index and the drug efficacy index into the drug treatment plan output model, output the drug treatment plan code through the drug treatment plan output model, match the corresponding drug treatment plan according to the drug treatment plan code, and formulate a personalized drug treatment plan table in combination with the drug dose allocation index and the drug efficacy index.

[0049] The beneficial effects of the present invention are as follows: In the present invention, a personalized drug treatment plan formulation system, compared with the traditional personalized drug treatment plan formulation system, the data acquisition technology, gene detection technology, neural network algorithm, linear regression algorithm and convolutional neural network algorithm in the method of the present invention are closely combined with modern information technology, accurately capture the basic treatment data and gene detection data, and then obtain the drug dose allocation index and the drug efficacy index, achieving real-time and comprehensive monitoring of the individual differences of patients, solving the problem that it is difficult for traditional technologies to formulate drug treatment plans personalized by combining various physiological factors, ensuring that the system in the present invention can refine the dynamic monitoring standard for a personalized drug treatment plan formulation system within a more accurate range, making the monitored data more accurate indicators under the same conditions. The research and application of this method significantly enhance the degree of intelligence in the process of formulating personalized drug treatment plans. Description of the Drawings

[0050] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, the following will briefly introduce the drawings required in the embodiments. Obviously, the drawings described below are only some embodiments recorded in the present invention. For those of ordinary skill in the art, other drawings can also be obtained based on these drawings.

[0051] Figure 1 It is a block diagram of a personalized drug treatment plan formulation system of the present invention. Detailed Embodiments

[0052] To make the objectives, technical solutions and advantages of the embodiments of the present invention clearer, the following will clearly and completely describe the technical solutions in the embodiments of the present invention with reference to the drawings in the embodiments of the present invention. Obviously, the described embodiments are some, but not all, of the embodiments of the present invention. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts fall within the scope of protection of the present invention.

[0053] Such as Figure 1As shown in the figure, the present invention provides a personalized drug treatment plan formulation system, including a drug treatment data collection module, a drug efficacy evaluation module, a drug treatment plan formulation module, and a drug treatment plan output module. Among them, each module is communicatively connected;

[0054] The drug treatment data collection module collects basic treatment data and gene detection data, preprocesses the collected data, and provides data preparation for the implementation of the subsequent module functions;

[0055] The drug efficacy evaluation module evaluates drug metabolism and drug efficacy based on the basic treatment data and gene detection data;

[0056] The drug treatment plan formulation module formulates a drug treatment plan based on the evaluation results of drug metabolism and drug efficacy;

[0057] The drug treatment plan output module combines the output results of the drug treatment plan formulation module to formulate a personalized drug treatment plan table, solving the problem that in the process of formulating a drug treatment plan, it is difficult for the prior art to formulate a drug treatment plan personalized by combining multiple factors, resulting in poor treatment effects of the drug treatment plan.

[0058] The drug efficacy evaluation module and the drug treatment plan formulation module specifically include:

[0059] The drug efficacy evaluation module is divided into a basic evaluation unit, a gene evaluation unit, and an efficacy comprehensive evaluation unit. Among them, the basic evaluation unit obtains the basic drug metabolism index through the basic treatment data; the gene evaluation unit obtains the drug gene metabolism index and the drug efficacy index respectively through the gene detection data; the efficacy comprehensive evaluation unit combines the basic drug metabolism index, the drug gene metabolism index, and the drug efficacy index to construct a drug treatment index evaluation model;

[0060] The drug treatment plan formulation module is divided into a drug dose allocation unit and a drug treatment plan formulation unit. Among them, the drug dose allocation unit uses the output result of the drug dose allocation model to obtain the drug dose allocation index, and then constructs a drug dose allocation model; the drug treatment plan formulation unit combines the output result of the drug dose allocation model to obtain the drug treatment plan, and then constructs a drug treatment plan model.

[0061] Preferably, the process of the drug treatment data collection module collecting basic treatment data and gene detection data and preprocessing the collected data includes:

[0062] Deploy collection devices to collect basic treatment data and gene detection data. Among them, the collection devices include medical height and weight scales, gene sequencers, etc.;

[0063] Among them, the basic treatment data are the age, gender, height, weight, and BMI value of the treatment subject; the gene detection data are the metabolic enzyme genes and drug target genes of the treatment subject;

[0064] Through a questionnaire survey, the age and gender of the treatment subject are collected. Using a medical height and weight scale, the height, weight, and BMI value of the treatment subject are collected. Through a gene sequencer, the metabolic enzyme genes and drug target genes of the treatment subject are collected;

[0065] Data cleaning and calibration are performed on the collected age, gender, height, weight, and BMI value of the treatment subject. Using the FastQC tool, quality assessment is performed on the metabolic enzyme genes and drug target genes of the treatment subject. Through the Cutadapt tool, low-quality bases in the metabolic enzyme genes and drug target genes are removed. Using GATK BaseRecalibrator, the quality values of the bases in the metabolic enzyme genes and drug target genes are corrected.

[0066] Preferably, the process of the basic evaluation unit for obtaining the drug basic metabolic index includes:

[0067] When the age of the treatment subject is between 0 and 18 years old, an age coefficient of 1 is assigned; when the age of the treatment subject is between 18 and 44 years old, an age coefficient of 0.95 is assigned; when the age of the treatment subject is between 45 and 64 years old, an age coefficient of 0.9 is assigned; when the age of the treatment subject is greater than 65 years old, an age coefficient of 0.8 is assigned;

[0068] When the treatment subject is male, a gender coefficient of 1 is assigned; when the treatment subject is female, a gender coefficient of 0.9 is assigned;

[0069] When the BMI value of the treatment subject is between 18.5 and 24.9, a BMI coefficient of 1 is assigned; when the BMI value of the treatment subject is between 25 and 29.9, a BMI coefficient of 1.1 is assigned; when the BMI value of the treatment subject is greater than 29.9, a BMI coefficient of 1.2 is assigned;

[0070] Using the age, height, and weight of the treatment subject, according to the Mifflin-St Jeor equation, the BMR value of the treatment subject is calculated. By calculating the product of the BMR value of the treatment subject, the age coefficient of the treatment subject, the gender coefficient of the treatment subject, and the BMI coefficient of the treatment subject, the drug basic metabolic index is obtained.

[0071] Preferably, the process of the gene evaluation unit for obtaining the drug gene metabolic index and the drug efficacy index includes:

[0072] Referring to the PharmGKB database, metabolic enzyme genes related to drug metabolism were screened out. According to the drug action mechanism, drug target genes of the treatment subjects related to drug efficacy were screened out. Using a nucleic acid extractor and a magnetic bead kit, the magnetic bead method was adopted to extract mRNA from the screened metabolic enzyme genes and drug target genes respectively;

[0073] By fluorescence quantitative PCR, the mRNA in the metabolic enzyme genes and the mRNA in the drug target genes were detected. The cycle numbers corresponding to the inflection points where the fluorescence signals of the mRNA in the metabolic enzyme genes and the mRNA in the drug target genes entered the exponential growth phase from the background were recorded respectively. These cycle numbers were the Ct values of the metabolic enzyme genes and the Ct values of the drug target genes;

[0074] Reference gene data related to the metabolic enzyme genes and drug target genes were extracted from the ICG database. Using the extracted reference gene data, the Ct values of the reference genes were calculated. According to the Ct values of the reference genes, the ratios of the Ct values of the metabolic enzyme genes to the Ct values of the reference genes and the ratios of the Ct values of the drug target genes to the Ct values of the reference genes were calculated respectively to obtain the relative expression levels of the metabolic enzyme genes and the relative expression levels of the drug target genes;

[0075] A metabolic baseline value was set. When the relative expression level of the metabolic enzyme gene was lower than 0.5 times the metabolic baseline value, the drug gene metabolism index was 0.8, indicating weak metabolism of the drug by the metabolic enzyme gene; when the relative expression level of the metabolic enzyme gene was between 0.5 times and 1.5 times the metabolic baseline value, the drug gene metabolism index was 1.5, indicating moderate metabolism of the drug by the metabolic enzyme gene; when the relative expression level of the metabolic enzyme gene was higher than 1.5 times the metabolic baseline value, the drug gene metabolism index was 2.5, indicating strong metabolism of the drug by the metabolic enzyme gene. The drug gene metabolism index was obtained;

[0076] A efficacy baseline value was set. When the relative expression level of the drug target gene was lower than 0.5 times the efficacy baseline value, the drug efficacy index was 1, indicating low efficacy of the drug on the drug target gene; when the relative expression level of the drug target gene was between 0.5 times and 2 times the efficacy baseline value, the drug efficacy index was 2, indicating general efficacy of the drug on the drug target gene; when the relative expression level of the drug target gene was higher than 2 times the efficacy baseline value, the drug efficacy index was 3, indicating high efficacy of the drug on the drug target gene. The drug efficacy index was obtained.

[0077] Preferably, for the efficacy comprehensive evaluation unit, the construction process of the drug treatment index evaluation model includes:

[0078] According to the contributions of the drug basic metabolism index and the drug gene metabolism index to drug metabolism, weights were assigned to the drug basic metabolism index and the drug gene metabolism index respectively, and the drug metabolism index was calculated by combining the weights. The calculation formula is , where U is the drug metabolism index, and are the weights of the drug basal metabolism index and the drug gene metabolism index respectively, and are the drug basal metabolism index and the drug gene metabolism index respectively;

[0079] Take the basic treatment data, gene detection data, drug metabolism index and drug efficacy index as the first data set, and divide it into a training set and a test set according to the ratio of 7:3;

[0080] Combined with the training set data and the neural network algorithm, take the basic treatment data and gene detection data as inputs, and the drug metabolism index and drug efficacy index as outputs, learn the non-linear relationship between the basic treatment data, the metabolic enzyme genes of the treatment object and the drug metabolism index, and the non-linear relationship between the drug target genes and the drug efficacy index, and train the drug treatment index evaluation model;

[0081] Input the test set data into the drug treatment index evaluation model, use the MSE function to evaluate the error between the drug metabolism index output by the drug treatment index evaluation model and the actual drug metabolism index, and the error between the drug treatment index output by the drug treatment index evaluation model and the actual drug efficacy index. According to the evaluation results, adjust the parameters of the drug treatment index evaluation model, optimize the performance of the drug treatment index evaluation model, and deploy the optimized drug treatment index evaluation model into the system.

[0082] Preferably, the process of the drug dose allocation unit obtaining the drug dose allocation index includes:

[0083] Input the basic treatment data and gene detection data into the drug treatment index evaluation model, and output the drug metabolism index through the drug treatment index evaluation model;

[0084] Take the drug metabolism index as the second data set, and divide it into a training set and a test set according to the ratio of 8:2;

[0085] Use the training set data and the linear regression algorithm, take the drug metabolism index as the input, and the drug dose allocation index as the output, adjust the intercept term and the regression coefficient of the drug metabolism index, learn the non-linear relationship between the drug metabolism index and the drug dose allocation index, and train the drug dose allocation model;

[0086] Input the test set data into the drug dose allocation model, evaluate the performance of the drug dose allocation model, adjust the intercept term and the regression coefficient of the drug dose allocation model, optimize the drug dose allocation model, and deploy the optimized drug dose allocation model into the system. The expression of this drug dose allocation model is , where, is the drug dose allocation index, is the intercept term, is the regression coefficient of the drug metabolism index, p is the drug metabolism index, is the error term.

[0087] Preferably, the process of the drug treatment plan formulation unit obtaining the drug treatment plan by combining the output result of the drug dose allocation model includes:

[0088] Input the drug metabolism index into the drug dose allocation model, and output the drug dose allocation index through the drug dose allocation model; input the gene detection data into the drug treatment index evaluation model, and output the drug efficacy index through the drug treatment index evaluation model;

[0089] When the drug dose allocation index is between 0 and 3, allocate low-dose drugs to the treatment object; when the drug dose allocation index is between 3 and 6, allocate conventional-dose drugs to the treatment object; when the drug dose allocation index is between 6 and 10, allocate high-dose drugs to the treatment object;

[0090] When the drug efficacy index is between 0 and 2, it indicates that the corresponding drug has poor efficacy, and the drug treatment should be replaced; when the drug efficacy index is between 2 and 6, it indicates that the corresponding drug has average efficacy, and the treatment should be combined with similar drugs and other drugs; when the drug efficacy index is between 6 and 10, it indicates that the corresponding drug has obvious efficacy, and continue to use the similar drug for treatment;

[0091] Combine the drug dose allocation plan and the drug type allocation plan to obtain the drug treatment plan, and encode the drug treatment plan.

[0092] Preferably, the construction process of the drug treatment plan model by the drug treatment plan formulation unit includes:

[0093] Take the drug dose allocation index, the drug efficacy index, and the drug treatment plan code as the third data set, and divide it into a training set and a test set according to a ratio of 8:2;

[0094] Using the training set data and the neural network algorithm, take the drug dose allocation index and the drug efficacy index as inputs, and the drug treatment plan code as the output, learn the non-linear relationship between the drug dose allocation index, the drug efficacy index, and the drug treatment plan code, and train the drug treatment plan model;

[0095] Input the test set data into the drug treatment plan model, compare the actual drug treatment plan code with the drug treatment plan code output by the drug treatment plan model, evaluate the performance of the drug treatment plan model, use the Adam optimizer to update the parameters of the convolutional neural network model, optimize the drug treatment plan model, and deploy the optimized drug treatment plan model to the system.

[0096] Preferably, the process of formulating a personalized drug treatment plan table by the drug treatment plan formulation module in combination with the output result of the drug treatment plan formulation module includes:

[0097] Input the drug dose allocation index and the drug efficacy index into the drug treatment plan output model. Output the drug treatment plan code through the drug treatment plan output model. According to the drug treatment plan code, match the corresponding drug treatment plan, and combine the drug dose allocation index and the drug efficacy index to formulate a personalized drug treatment plan table.

[0098] First, collect basic treatment data by combining a questionnaire survey and a medical height and weight scale; use a gene sequencer to collect gene detection data and preprocess the collected data. Secondly, according to the age, gender, and BMI value of the treatment object, assign an age coefficient, a gender coefficient, and a BMI coefficient to the treatment object respectively. Use the age, height, and weight of the treatment object to calculate the BMR value of the treatment object according to the Mifflin-St Jeor equation, and then obtain the drug basal metabolic index. Immediately afterwards, use a nucleic acid extractor and a magnetic bead kit, and adopt the magnetic bead method to extract the mRNA in the screened metabolic enzyme gene and drug target gene respectively, and obtain the relative expression level of the metabolic enzyme gene and the relative expression level of the drug target gene through real-time fluorescence quantitative PCR, and then obtain the drug gene metabolism index and the drug efficacy index. Then, use the drug basal metabolic index and the drug gene metabolism index to calculate the drug metabolism index, and construct a drug treatment index evaluation model through a neural network algorithm. Then, input the basic treatment data and the gene detection data into the drug treatment index evaluation model. The drug treatment index evaluation model outputs the drug metabolism index, uses a linear regression algorithm to obtain the drug dose allocation index, and obtains the drug treatment plan and its code according to the drug dose allocation index and the drug efficacy index. Immediately afterwards, use a convolutional neural network algorithm to construct a drug treatment plan model to realize inputting the drug dose allocation index and the drug efficacy index, and the drug treatment plan model automatically outputs the drug treatment plan. Finally, combine the personalized drug treatment plan output model to formulate a personalized drug treatment plan table.

[0099] As described above, it is only the specific implementation manner of the present application, but the protection scope of the present application is not limited thereto. Any person skilled in the art within the technical scope disclosed by the present application can easily think of changes or substitutions, which should all be covered within the protection scope of the present application. Therefore, the protection scope of the present application should be subject to the protection scope of the claims.

Claims

1. A personalized drug treatment program formulation system, comprising a drug treatment data acquisition module, a drug efficacy evaluation module, a drug treatment program formulation module and a drug treatment program output module, wherein: Each module is connected in communication, characterized by: The drug treatment data collection module collects basic treatment data and gene detection data, and pre-processes the collected data; The drug efficacy evaluation module evaluates drug metabolism and drug efficacy based on basic treatment data and genetic testing data; The drug treatment program formulation module formulates a drug treatment program based on the evaluation results of drug metabolism and drug efficacy; The drug treatment plan output module formulates a personalized drug treatment plan table in combination with the output results of the drug treatment plan formulation module.

2. A personalized drug treatment program formulation system according to claim 1, characterized in that: The drug efficacy evaluation module and the drug treatment plan formulation module specifically include: The drug efficacy evaluation module is divided into a basic evaluation unit, a gene evaluation unit and a comprehensive efficacy evaluation unit, wherein the basic evaluation unit obtains the basic drug metabolism index through basic treatment data; the gene evaluation unit obtains the drug gene metabolism index and the drug efficacy index respectively through gene detection data; the comprehensive efficacy evaluation unit constructs a drug treatment index evaluation model by combining the basic drug metabolism index, the drug gene metabolism index and the drug efficacy index; The drug treatment plan formulation module is divided into a drug dosage allocation unit and a drug treatment plan formulation unit, wherein the drug dosage allocation unit uses the output results of the drug dosage allocation model to obtain the drug dosage allocation index, and then constructs the drug dosage allocation model; the drug treatment plan formulation unit combines the output results of the drug dosage allocation model to obtain the drug treatment plan, and then constructs the drug treatment plan model.

3. A personalized drug treatment program formulation system according to claim 2, characterized in that: The drug treatment data collection module collects basic treatment data and gene detection data, and pre-processes the collected data, including: Deploy collection equipment to collect basic treatment data and genetic testing data, wherein the collection equipment includes a medical height and weight scale and a genetic sequencer; The basic treatment data are the age, gender, height, weight and BMI value of the treatment object; the genetic testing data are the metabolic enzyme genes and drug target genes of the treatment object; The collected age, gender, height, weight and BMI values ​​of the treatment subjects were cleaned and calibrated. The quality of the metabolic enzyme genes and drug target genes of the treatment subjects was assessed using the FastQC tool. The low-quality bases in the metabolic enzyme genes and drug target genes were removed using the Cutadapt tool. The quality values ​​of the bases in the metabolic enzyme genes and drug target genes were corrected using the GATK BaseRecalibrator.

4. A personalized drug treatment program formulation system according to claim 3, characterized in that: The process of obtaining the basic drug metabolism index by the basic evaluation unit includes: According to the age range of the treatment object, an age coefficient is assigned to the treatment object of the corresponding age range; according to the gender type of the treatment object, a gender coefficient is assigned to the treatment object of the corresponding gender type; according to the BMI value range of the treatment object, a BMI coefficient is assigned to the treatment object of the corresponding BMI value range; The BMR value of the treated subject is calculated according to the Mifflin-St Jeor equation using the age, height and weight of the treated subject. The basic drug metabolic index is obtained by calculating the product of the BMR value of the treated subject, the age coefficient of the treated subject, the gender coefficient of the treated subject and the BMI coefficient of the treated subject.

5. A personalized drug treatment program formulation system according to claim 4, characterized in that: The process of obtaining the drug gene metabolism index and the drug efficacy index by the gene evaluation unit includes: Refer to the PharmGKB database to screen out metabolic enzyme genes related to drug metabolism, and screen out drug target genes related to drug efficacy based on the drug action mechanism. Use a nucleic acid extractor and a magnetic bead kit to extract mRNA from the screened metabolic enzyme genes and drug target genes using the magnetic bead method. The mRNA in the metabolic enzyme gene and the mRNA in the drug target gene are detected by fluorescence quantitative PCR, and the cycle numbers corresponding to the inflection points of the fluorescence signals of the mRNA in the metabolic enzyme gene and the mRNA in the drug target gene from the background to the exponential growth phase are recorded respectively, and the cycle numbers are the Ct values ​​of the metabolic enzyme gene and the Ct values ​​of the drug target gene respectively; Extracting internal reference gene data related to metabolic enzyme genes and drug target genes from the ICG database, calculating the internal reference gene Ct value using the extracted internal reference gene data, and calculating the ratio of the Ct value of the metabolic enzyme gene to the Ct value of the internal reference gene and the ratio of the Ct value of the drug target gene to the Ct value of the internal reference gene according to the internal reference gene Ct value, respectively, to obtain the relative expression of the metabolic enzyme gene and the relative expression of the drug target gene; The metabolic baseline value was set. When the relative expression of the metabolic enzyme gene was lower than 0.5 times the metabolic baseline value, the drug gene metabolic index was 0.8; when the relative expression of the metabolic enzyme gene was between 0.5 times the metabolic baseline value and 1.5 times the metabolic baseline value, the drug gene metabolic index was 1.5; when the relative expression of the metabolic enzyme gene was higher than 1.5 times the metabolic baseline value, the drug gene metabolic index was 2.5; The efficacy baseline value is set. When the relative expression level of the drug target gene is lower than 0.5 times the efficacy baseline value, the drug efficacy index is 1; when the relative expression level of the drug target gene is between 0.5 times the efficacy baseline value and 2 times the efficacy baseline value, the drug efficacy index is 2; when the relative expression level of the drug target gene is higher than 2 times the efficacy baseline value, the drug efficacy index is 3.

6. A personalized drug treatment program formulation system according to claim 5, characterized in that: The construction process of the comprehensive efficacy evaluation unit and the drug therapeutic index evaluation model includes: According to the contribution of drug basic metabolism index and drug gene metabolism index to drug metabolism, weights are assigned to drug basic metabolism index and drug gene metabolism index respectively, and the drug metabolism index is calculated by combining the weights; The basic treatment data, genetic testing data, drug metabolism index and drug efficacy index are used as the first data set and divided into training set and test set in a ratio of 7:3; Combine the training set data with the neural network algorithm, take the basic treatment data and gene detection data as input, take the drug metabolism index and drug efficacy index as output, learn the nonlinear relationship between the basic treatment data, the metabolic enzyme gene of the treatment object and the drug metabolism index, the nonlinear relationship between the drug target gene and the drug efficacy index, and train the drug therapeutic index evaluation model; The test set data is input into the drug therapeutic index evaluation model. The MSE function is used to evaluate the error between the drug metabolism index output by the drug therapeutic index evaluation model and the actual drug metabolism index, as well as the error between the drug therapeutic index output by the drug therapeutic index evaluation model and the actual drug efficacy index. According to the evaluation results, the drug therapeutic index evaluation model parameters are adjusted to optimize the performance of the drug therapeutic index evaluation model, and the optimized drug therapeutic index evaluation model is deployed into the system.

7. A personalized drug treatment program formulation system according to claim 6, characterized in that: The process of obtaining the drug dosage distribution index by the drug dosage distribution unit includes: Inputting basic treatment data and genetic testing data into a drug therapeutic index evaluation model, and outputting a drug metabolism index through the drug therapeutic index evaluation model; The drug metabolism index was used as the second data set and divided into a training set and a test set in a ratio of 8:2; Using the training set data and the linear regression algorithm, the drug metabolism index is used as input, the drug dose distribution index is used as output, the intercept term and the regression coefficient of the drug metabolism index are adjusted, the nonlinear relationship between the drug metabolism index and the drug dose distribution index is learned, and the drug dose distribution model is trained; Input the test set data into the drug dose allocation model, evaluate the performance of the drug dose allocation model, adjust the intercept term and regression coefficient of the drug dose allocation model, optimize the drug dose allocation model, and deploy the optimized drug dose allocation model into the system.

8. A personalized drug treatment program formulation system according to claim 7, characterized in that: The process of obtaining the drug treatment plan by the drug treatment plan formulation unit in combination with the output result of the drug dosage distribution model includes: The drug metabolism index is input into the drug dosage distribution model, and the drug dosage distribution index is output through the drug dosage distribution model; the gene detection data is input into the drug therapeutic index evaluation model, and the drug efficacy index is output through the drug therapeutic index evaluation model; allocating a drug dose allocation plan for the treatment subject according to the drug dose allocation index, the drug dose allocation plan comprising a low dose allocation, a conventional dose allocation and a high dose allocation; According to the drug efficacy index, a drug type allocation plan is allocated to the treatment subject, and the drug type allocation plan includes changing drug treatment, combining the same drug with other drugs for treatment, and continuing to use the same drug for treatment; The drug dosage allocation plan is combined with the drug type allocation plan to obtain the drug treatment plan, and the drug treatment plan is coded.

9. A personalized drug treatment program formulation system according to claim 8, characterized in that: The drug treatment plan formulation unit, the process of constructing the drug treatment plan model includes: The drug dose distribution index, drug efficacy index and drug treatment regimen coding were used as the third data set and divided into training set and test set in a ratio of 8:2; Using training set data and neural network algorithm, taking drug dose distribution index and drug efficacy index as input and drug treatment regimen coding as output, learning the nonlinear relationship between drug dose distribution index and drug efficacy index and drug treatment regimen coding, and training drug treatment regimen model; Input the test set data into the drug treatment regimen model, compare the actual drug treatment regimen code with the drug treatment regimen code output by the drug treatment regimen model, evaluate the performance of the drug treatment regimen model, update the convolutional neural network model parameters, optimize the drug treatment regimen model, and deploy the optimized drug treatment regimen model to the system.

10. A personalized drug treatment program formulation system according to claim 9, characterized in that: The process of the drug treatment plan formulation module formulating a personalized drug treatment plan table in combination with the output result of the drug treatment plan formulation module includes: The drug dosage distribution index and the drug efficacy index are input into the drug treatment plan output model, the drug treatment plan code is output through the drug treatment plan output model, the corresponding drug treatment plan is matched according to the drug treatment plan code, and a personalized drug treatment plan table is formulated in combination with the drug dosage distribution index and the drug efficacy index.

Images