Drug risk assessment system and method based on gene polymorphism

Abstract

The invention belongs to the technical field of medical data processing, and provides a drug risk assessment system and method based on gene polymorphism, and the method comprises the steps: collecting historical medical data, carrying out the data preprocessing of the historical medical data, and carrying out the feature screening according to a specified screening rule; defining positive samples and negative samples based on the plurality of screened features, and establishing a training data set, specifically comprising: before establishing the training data set, performing oversampling on a small number of samples so that the number of the positive samples is equal to that of the negative samples; establishing a risk prediction model, and training the risk prediction model by adopting the training data set to obtain a trained risk prediction model; and inputting the drug dose and the test data of a to-be-evaluated object into the trained risk prediction model to obtain a risk evaluation result. According to the method, the risk condition of maintaining the dosage of the metoprolol is effectively evaluated.

Description

Technical Field

[0001] The present invention relates to the technical field of medical data processing, and in particular to a drug risk assessment system and method based on gene polymorphism. Background Art

[0002] In the prior art, the selective β1 receptor blocker metoprolol is a drug for cardiovascular diseases in the elderly. However, in the use of metoprolol in elderly patients with multiple diseases, multiple medications, and impaired liver and kidney function, there is a large inter-individual variability in the treatment response, which brings troubles to the clinical treatment of elderly patients with cardiovascular diseases. Cytochrome P450 2D6 (CYP2D6) gene polymorphism is an important reason for the inter-individual variability in metoprolol response.

[0003] However, the existing methods for processing medical data are too cumbersome. In addition, there is still much room for improvement in how to effectively assess the risk of metoprolol and effectively process medical data.

[0004] Therefore, it is necessary to provide a drug evaluation method to solve the above problems. Summary of the invention

[0005] The present invention is intended to provide a drug risk assessment system and method based on gene polymorphism to solve the technical problems in the prior art such as how to effectively conduct risk assessment on metoprolol drugs and effectively process medical data. The technical problems to be solved by the present invention are achieved through the following technical solutions.

[0006] The first aspect of the present invention proposes a drug risk assessment method based on gene polymorphism, comprising: collecting historical medical data, and performing data preprocessing on the collected historical medical data, and performing feature screening according to specified screening rules, wherein the feature screening comprises: using univariate analysis to calculate each feature for the first screening, and using the Spielman correlation method to calculate feature correlation for the second screening; based on the screened multiple features, defining positive samples and negative samples, and establishing a training data set, specifically comprising: before establishing the training data set, oversampling a small number of samples so that the number of positive samples and negative samples is equal; establishing a risk prediction model, and using the training data set to train the risk prediction model to obtain a trained risk prediction model; inputting the drug dosage and test data of the object to be evaluated into the trained risk prediction model to obtain a risk assessment result.

[0007] The second aspect of the present invention proposes a drug risk assessment system based on gene polymorphism, which adopts the drug risk assessment method based on gene polymorphism described in the first aspect of the present invention, including: a data collection module, which collects historical medical data, and performs data preprocessing on the collected historical medical data, and performs feature screening according to specified screening rules, wherein the feature screening includes: using univariate analysis to calculate each feature for the first screening, and using the Spierman correlation method to calculate the feature correlation for the second screening; a first establishment module, which defines positive samples and negative samples based on the screened multiple features, and establishes a training data set, specifically including: before establishing the training data set, oversampling a small number of samples so that the number of positive samples and negative samples is equal; a second establishment module, which establishes a risk prediction model, and uses the training data set to train the risk prediction model to obtain a trained risk prediction model; a result output module, which inputs the drug dosage and test data of the object to be evaluated into the trained risk prediction model to obtain a risk assessment result.

[0008] The third aspect of the present invention provides an electronic device, comprising: one or more processors; a storage device for storing one or more programs; when the one or more programs are executed by the one or more processors, the one or more processors implement the method described in the first aspect of the present invention.

[0009] A fourth aspect of the present invention provides a computer-readable medium having a computer program stored thereon, wherein the computer program, when executed by a processor, implements the method described in the first aspect of the present invention.

[0010] The embodiments of the present invention include the following advantages: Compared with the prior art, the present invention analyzes CYP2D6 and ADRB1 gene polymorphisms, records and collects relevant medical data on metoprolol tolerance, adverse reactions, blood pressure and heart rate responses, and long-term prognosis, performs data processing on the above-collected relevant medical data, screens out multiple features to establish a training data set, and establishes a risk prediction model, uses the training data set to train the risk prediction model to obtain a trained risk prediction model; inputs the drug dosage and test data of the object to be evaluated into the trained risk prediction model to obtain a risk assessment result, thereby providing a theoretical basis for establishing an individualized metoprolol medication regimen guided by gene polymorphism.

[0011] In addition, univariate analysis was used to calculate each feature for the first screening, and the Spierman correlation method was used to calculate the feature correlation for the second screening. The two screening processes can help build a more reliable and interpretable model, simplify the model, remove redundant features, and improve the interpretability of features. BRIEF DESCRIPTION OF THE DRAWINGS

[0012] Figure 1 is a flowchart of an example of the drug evaluation method of the present invention; Figure 2 is a partial flow chart of the drug evaluation method of the present invention; Figure 3 is another partial flow chart of the drug evaluation method of the present invention; Figure 4 is a structural block diagram of the drug evaluation system of the present invention; Figure 5 is a schematic structural diagram of an electronic device embodiment according to the present invention; Figure 6 It is a schematic diagram of the structure of a computer-readable medium embodiment according to the present invention. DETAILED DESCRIPTION

[0013] It should be noted that, in the absence of conflict, the embodiments and features in the embodiments of the present application can be combined with each other. The present invention will be described in detail below with reference to the accompanying drawings and in combination with the embodiments.

[0014] In view of the above problems, the present invention proposes a drug evaluation method, which analyzes CYP2D6 and ADRB1 gene polymorphisms, records and collects relevant medical data on metoprolol tolerance, adverse reactions, blood pressure and heart rate responses, and long-term prognosis, performs data processing on the above-collected relevant medical data, screens out multiple features to establish a training data set, and establishes a risk prediction model, and uses the training data set to train the risk prediction model to obtain a trained risk prediction model; the drug dosage and test data of the object to be evaluated are input into the trained risk prediction model to obtain a risk assessment result, thereby providing a theoretical basis for establishing an individualized metoprolol medication regimen guided by gene polymorphism.

[0015] Embodiment 1: Figure 1 The following is a flowchart of an example of the method for evaluating a drug according to the present invention.

[0016] Refer to the following Figure 1 , Figure 2 and Figure 3 , the contents of the present invention will be described in detail.

[0017] First, in step S101, historical medical data is collected, and the collected historical medical data includes condition data, drug dosage, physiological index data, and test data.

[0018] In a specific embodiment, for example, historical medical data on the use of metoprolol in a hospital within a specified historical period (e.g., April 2017 to July 2019) is collected.

[0019] For example, the users are preferably elderly people over 65 years old. Specifically, there are 1039 male users.

[0020] Specifically, historical medical data includes condition data, drug dosage, physiological index data, and test data. Condition data includes cardiovascular condition. Drug dosage includes metoprolol starting dose (for example, all metoprolol doses are converted to metoprolol tartrate doses, specifically 50 mg of metoprolol tartrate). Physiological index data include the age and weight of historical users, blood pressure, heart rate, and gene polymorphism data before and after drug use. The gene polymorphism data is obtained by, for example, measuring the polymorphism of CYP2D6*2 (rs16947), *10 (rs1065852), *14 (rs5030865) and ADRB1 (rs1801253) sites using Sanger sequencing.

[0021] In addition, the test data includes whether a clinical endpoint event occurs and the date of occurrence. In addition, the metoprolol dosage of the above-mentioned users and whether adverse reactions occur are also collected. The method of collecting data at equal intervals and for a specific length of time (e.g., 12 weeks, 24 weeks, or 36 weeks) is specifically used. For example, blood pressure, heart rate, metoprolol maintenance dose, and related adverse reactions are collected every two weeks, and the related adverse reactions specifically include cardiovascular adverse reactions and non-cardiovascular adverse reactions, wherein cardiovascular adverse reactions are defined as postural hypotension, bradycardia (heart rate ≤ 55bpm), arrest, II degree and above atrioventricular block, syncope, and cold limbs. Non-cardiovascular adverse reactions are defined as dyspnea, sleep disorders and fatigue, headache or dizziness, and depression.

[0022] Specifically, if metoprolol was not discontinued at the time of the last data collection, the metoprolol dose was defined as the metoprolol maintenance dose.

[0023] In another embodiment, for example, the medical record data disclosed in the electronic medical record system (EMR) is obtained. The privacy information in the obtained data is desensitized before use. In addition, because real-world data has problems such as missing patient individual information, incomplete relevant medical records, and serious differences between records, it is also necessary to process missing values ​​and outliers.

[0024] It should be noted that, for the above-mentioned method of collecting medical data, it is also possible to combine outpatient inquiries, query electronic medical record systems and telephone follow-up. In addition, a method of collecting data every six months can also be adopted. In addition, clinical endpoint events include recorded all-cause death, cardiovascular death, myocardial infarction, heart failure and acute stroke, etc. In addition, the method of the present invention is particularly suitable for elderly people aged 65 to 85 years old. The above is only explained as an optional example and cannot be understood as a limitation of the present invention.

[0025] Next, in step S102, the collected historical medical data is preprocessed and feature screening is performed according to specified screening rules. The feature screening includes: using univariate analysis to calculate each feature for the first screening, and using the Spearman correlation method to calculate feature correlation for the second screening.

[0026] Specifically, the collected historical medical data were preprocessed. The CYP2D6 activity score was calculated according to the latest consensus and converted into the CYP2D6 metabolic phenotype. For example, the site polymorphisms of CYP2D6*2 (rs16947), *10 (rs1065852), *14 (rs5030865) and ADRB1 (rs1801253) were determined by Sanger sequencing, and the site polymorphism of CYP2D6*5 (gene deletion) was detected by optimized large fragment polymerase chain reaction. The CYP2D6 activity score (AS) was calculated according to the latest consensus, and the CYP2D6 genotype was converted into a metabolic phenotype. In the data of this example, there were 651 cases (62.7%) of normal metabolizers (NMs), 385 cases (37.1%) of intermediate metabolizers (IMs), and 3 cases (0.3%) of poor metabolizers (PMs). PMs will be excluded in subsequent data analysis and processing due to their small number. All allele frequencies were in Hardy-Weinberg equilibrium. For example, the stable plasma trough concentration of metoprolol was determined by high performance liquid chromatography-mass spectrometry.

[0027] Furthermore, the "activity score" was converted into a categorical variable. Specifically, an activity score ≤ 1, corresponding to a category value of 0, refers to intermediate metabolizers (IMs). An activity score > 1, corresponding to a category value of 1, refers to normal metabolizers (NMs). And, it was named "activity score classification".

[0028] Convert "metoprolol maintenance dose" into a categorical variable. For example, use one-hot encoding to convert features. Specifically, if the metoprolol maintenance dose is ≤25 mg / d, the corresponding category value is 1, named "maintenance dose_category 1"; if the metoprolol maintenance dose is >25 mg / d, the corresponding category value is 2, named "maintenance dose_category 2" (it is classified according to the dose of metoprolol, which can be considered as low dose or normal dose) or use 50 mg / d to define the category label.

[0029] The following expression was used to perform index conversion on “brain natriuretic peptide (BNP), cardiac troponin I (cTnI), cardiac troponin T (cTnT)”;

[0030] in, Indicates the data of converting "brain natriuretic peptide, troponin I, troponin T" into index form. i Is a positive integer.

[0031] Then, all preprocessed medical data are subjected to feature screening according to the specified screening rules.

[0032] Univariate analysis was used to calculate the P value of significance between each feature and adverse reaction, and the calculated P was compared with the first specific value for the first screening.

[0033] After the first screening, the feature correlation is calculated using the Spearman correlation method, and the absolute value of the calculated correlation between the features is compared with the second specific value for the second screening.

[0034] Specifically, the specified rules include: deleting features whose calculated P values ​​are greater than a first specific value; and deleting selected features when the calculated absolute value of the correlation between the selected features is greater than a second specific value.

[0035] Optionally, the first specific value is 0.05 to 0.2, preferably 0.1, and the second specific value is 0.6 to 0.8, preferably 0.7.

[0036] In this example, 50% of the features passed the above screening process. The specific P values ​​are shown in Table 1 below.

[0037]

[0038] Specifically, in the second screening, the Spielman correlation is used to calculate the correlation between the features screened out in the first screening and "whether it is an adverse reaction". Features whose absolute value of correlation with "whether it is an adverse reaction" is greater than or equal to 0.1 are selected as candidate features. At the same time, ensure that there is no high correlation between the selected features. Features with a correlation coefficient greater than 0.7 (for example, a total of about 30% of the features are retained) will be excluded to avoid redundant information brought by multiple highly correlated features. Therefore, the above screening process can help build a more reliable and interpretable model, and can also simplify the model, remove redundant features, and improve interpretability.

[0039] Next, the candidate features obtained after the first and second screenings were input into the random forest model to determine the feature combination with the optimal AUC value, namely, the following nine features were obtained: activity score, age, brain natriuretic peptide, cerebrovascular disease, liver disease, troponin I, heart rate, diastolic blood pressure, and atrial fibrillation.

[0040] Specifically, the Gini parameter is used to train the random forest model, calculate the AUC value and evaluate the importance of different features (specifically, the features are arranged in descending order).

[0041] In a specific embodiment, the model hyperparameters of the random forest model are "n_estimators=100, max_depth=None, random_state=42, criterion='gini'". Subsequently, different numbers of features are gradually selected and input into the random forest model. The AUC value of the model is recalculated. By finding the feature combination with the optimal AUC value, the feature combination with the best predictive performance in the feature screening process is determined, and the feature dimension can be reduced, and the operating efficiency of the model can be improved. Through the above screening process, 9 features were selected for modeling, namely "activity score", "age", "brain natriuretic peptide", "cerebrovascular disease", "liver disease", "troponin I", "heart rate", "diastolic blood pressure", and "atrial fibrillation".

[0042] It should be noted that the above is only described as an optional example and should not be understood as a limitation to the present invention.

[0043] Next, in step S103, based on the screened multiple features, positive samples and negative samples are defined to establish a training data set, which specifically includes: before establishing the training data set, oversampling a small number of samples to make the number of positive samples and negative samples equal.

[0044] Specifically, the nine features determined are divided into continuous variables and discrete variables. Among them, the continuous variables include "age", "brain natriuretic peptide", "troponin I", "heart rate", and "diastolic blood pressure". The discrete variables include "cerebrovascular disease", "liver disease", and "atrial fibrillation".

[0045] Oversampling a small number of samples, such as Figure 3 As shown, the specific steps include: Step S301: Divide the determined nine features into continuous variables and discrete variables.

[0046] Step S302: For all continuous variables, use the SMOTE method to synthesize continuous variable negative samples.

[0047] Step S303: For each synthesized negative sample, find its k nearest neighbor negative samples for each discrete variable, and calculate the Hamming distance of these negative samples on the category feature, select the nearest neighbor sample with the smallest Hamming distance as a reference, and synthesize the category feature of the new sample by randomly selecting a category value different from the reference sample.

[0048] Step S304: combining the type characteristics of the synthesized discrete variable with the continuous variable value to obtain a complete synthesized sample.

[0049] For example, all continuous variables are synthesized with the SMOTE method as the negative sample.

[0050] Use the K-nearest neighbor algorithm (for example, k=5) to calculate the number of samples in each minority class. The Euclidean distance between a point and its k neighbors is calculated, and one of the neighbors (i.e., the nearest neighbor sample) is randomly selected. Random interpolation is performed on the line segment between the sample point and the neighbor. The interpolation position is determined by a random number between 0 and 1. The calculation formula is as follows:

[0051] Among them, x new is the newly synthesized sample point, x i is the original minority class sample, x j is x i Neighbors. rand(0,1) represents a random number between 0 and 1, such as 0.6, 0.7, 0.8, and preferably 0.6.

[0052] For each synthesized negative sample x new , for each discrete variable, find the negative samples of its k nearest neighbors 𝑁 new (x new ), and calculate the Hamming distance H of these negative samples on the category features d, select the nearest neighbor sample X with the minimum Hamming distance ref As a reference, the category feature of the new sample is synthesized by randomly selecting a category value different from the reference sample. The specific formula is as follows:

[0053] in, and The samples are and The value of the jth discrete variable; is the number of possible categories of the dth discrete variable; Represents a bitwise exclusive-or operation.

[0054] Combining the type features of the synthesized discrete variables with the continuous variable values, a complete synthetic sample is obtained. The synthetic sample consists of 9 modeled features.

[0055] Okay, let's use a specific example to describe how to combine the type features of the synthesized discrete variables with the continuous variable values ​​to get a complete synthetic sample. This synthetic sample will contain 9 features for modeling.

[0056] For example, there are 9 features below, of which the first 5 are continuous variables and the last 4 are discrete variables: 1. Activity score (continuous), 2. Age (continuous), 3. Brain natriuretic peptide (continuous), 4. Troponin I (continuous), 5. Heart rate (continuous), 6. Cerebrovascular disease (discrete), 7. Liver disease (discrete), 8. Troponin T (discrete), 9. Atrial fibrillation (discrete).

[0057] Synthesize a new sample, where the values ​​of the continuous variables have been generated by the SMOTE algorithm, and the values ​​of the discrete variables need to be synthesized through the following steps: Step 1: Synthesis of continuous variable values Suppose that the following continuous variable values ​​are synthesized for a negative sample (a patient who did not experience an adverse reaction) through the SMOTE algorithm: - Activity score: 0.5 - Age: 68 - Brain natriuretic peptide: 300 - Troponin I: 0.2 - Heart rate: 72 Step 2: Synthesis of discrete variable values For discrete variables, find the positive sample (patients with adverse reactions) that is most similar to the original negative sample, and synthesize new category values ​​according to certain rules. Suppose we have the following discrete variable values ​​for positive samples: - Cerebrovascular disease: Yes - Liver disease: No - Troponin T: Yes - Atrial fibrillation: No Use the K-nearest neighbor algorithm to find the positive sample that is most similar to the synthesized continuous variable value and calculate the Hamming distance. The following positive samples are selected as references: - Cerebrovascular disease: No - Liver disease: Yes - Troponin T: No - Atrial fibrillation: Yes Step 3: Combine continuous and discrete variables Combine the synthesized continuous variable values ​​with the discrete variable values ​​to get a complete synthetic sample. In this example, you might decide to change the status of "cerebrovascular disease" and "atrial fibrillation" to increase the diversity of the sample. Therefore, the discrete variable values ​​of the synthetic sample might be as follows: - Cerebrovascular disease: Yes (changed from "No" to "Yes") - Liver disease: Yes (remains unchanged) - Troponin T: Yes (changed from "No" to "Yes") - Atrial Fibrillation: Yes (changed from No to Yes) Complete synthetic sample Combining continuous and discrete variables, we get the following complete synthetic sample: - Activity score: 0.5 - Age: 68 - Brain natriuretic peptide: 300 - Troponin I: 0.2 - Heart rate: 72 - Cerebrovascular disease: Yes - Liver disease: Yes - Troponin T: Yes - Atrial fibrillation: Yes This synthetic sample can now be added to the training dataset for training the risk prediction model. In this way, not only the number of positive samples is increased, but also the model's ability to identify different patient characteristics is improved by introducing different combinations of discrete variables.

[0058] Based on the selected features, the above oversampling step (specifically, oversampling a small number of samples so that the number of positive samples and negative samples is equal) is used to sample the positive samples to negative samples in a ratio of 1:1 to establish a training data set.

[0059] Therefore, through the above process of oversampling a small number of samples, the data imbalance problem can be effectively improved, the model performance can be improved, and the generalization ability of the model can be enhanced.

[0060] Next, in step S104, a risk prediction model is established, and the risk prediction model is trained using a training data set to obtain a trained risk prediction model.

[0061] For example, a naive Bayes method is used to establish a risk prediction model. The risk prediction model is trained using the training set established in step S103.

[0062] Optionally, the data set is also randomly split into a training set and a test set in the form of training set: test set = 8:2. After sampling the training set using S103, a risk prediction model is established using 8 different machine learning models and nomograms, and the model effect is evaluated using the test set to determine the selection of the naive Bayes method to establish the risk prediction model.

[0063] In the above steps, it is mentioned that 8 different machine learning models are used to build risk prediction models, and the model effects are evaluated through test sets. In order to enhance training, this process can be expanded, including adding more algorithm models and performing more detailed model evaluation and selection. The following are the expanded and supplemented steps: 1. Dataset Splitting: First, the dataset is randomly split into 80% training set and 20% test set. This ensures that there is enough data to train the model and that there is an independent dataset to evaluate the generalization ability of the model.

[0064] 2. Feature Engineering: Before training the model, feature engineering is performed, including feature selection, feature extraction, and feature transformation. This may include standardizing continuous variables, encoding discrete variables, handling missing values, and transforming nonlinear features.

[0065] 3. Model training: Use the training set data to train multiple machine learning models. In addition to Naive Bayes, the following models can also be considered: - Logistic Regression - Support Vector Machine (SVM) - Decision Tree - Random Forest - Gradient Boosting Machine (GBM) - Extra Trees - Artificial Neural Networks 4. Model tuning: Tune each model, including adjusting hyperparameters and selecting the best model configuration. This can be achieved through methods such as grid search or random search.

[0066] 5. Cross-validation: To more accurately evaluate model performance, use cross-validation (such as k-fold cross-validation) to evaluate the stability and accuracy of each model.

[0067] 6. Model evaluation: Use the test set to evaluate each model. Evaluation indicators may include accuracy, sensitivity, specificity, precision, F1 score, AUC value, etc.

[0068] 7. Model comparison: Compare the performance of different models and select the best performing model. If a model performs well on multiple evaluation metrics, it may be the best choice.

[0069] 8. Model fusion: Consider using model fusion techniques such as stacking or voting to combine the prediction results of multiple models to improve the overall performance.

[0070] 9. Model interpretation: For the best model selected, use model interpretation tools (such as SHAP, LIME) to explain the model’s decision-making process, which is especially important for models in the medical field.

[0071] 10. Model deployment: Deploy the trained model to the production environment to perform risk assessment on new patient data.

[0072] 11. Continuous monitoring: After the model is deployed, continuously monitor the performance of the model to ensure that it remains accurate and reliable in real-world applications.

[0073] These expansions and supplementary steps enable a more comprehensive evaluation and selection of the best machine learning models, as well as enhanced training to improve model performance and robustness.

[0074] In an optional implementation, the above nine features are divided into two or more categories of labeled data, and the risk prediction model is enhanced and trained to obtain a trained risk prediction model.

[0075] In another optional implementation, the above nine features are divided into data with multiple labels, and the risk prediction model is trained or enhanced to obtain a trained risk prediction model.

[0076] Next, in step S105, the drug dosage and test data of the subject to be evaluated are input into the trained risk prediction model to obtain a risk assessment result.

[0077] When applying the risk prediction model, the drug dosage and test data of the subject to be evaluated are input into the trained risk prediction model, and the risk assessment value is output to determine the adverse reaction or risk situation. The risk assessment value is, for example, the probability of whether the risk occurs, and further based on the probability, the occurrence of adverse reactions or other risks is determined.

[0078] It should be noted that the above is only described as an optional example and should not be understood as a limitation to the present invention.

[0079] In order to verify the effect of the present invention, a comparative test was conducted, and the comparative effect is shown in Table 1 below.

[0080]

[0081] Through the above comparative test, the method of the present invention is compared with the methods of 'K nearest neighbor', 'support vector machine', 'logistic regression', 'decision tree', 'AdaBoost', 'XGBoost', and 'random forest' in the prior art to perform risk assessment calculations and obtain the evaluation indicators in the above Table 1.

[0082] Specifically, the confusion matrix can be calculated based on the actual results and the predicted results, so as to calculate various evaluation indicators, as follows: Confusion Matrix

[0083]

[0084] AUROC refers to the area under the curve between sensitivity and (1-specificity), and AUPRC refers to the area under the curve between precision and sensitivity.

[0085] From Table 1, it can be seen that the method of the present invention is superior to other existing methods in various evaluation indicators.

[0086] In addition, the drawings are only schematic illustrations of the processes included in the method according to an exemplary embodiment of the present invention, and are not intended to be limiting. It is easy to understand that the processes shown in the drawings do not indicate or limit the time sequence of these processes. In addition, it is also easy to understand that these processes can be performed synchronously or asynchronously, for example, in multiple modules.

[0087] Compared with the prior art, the present invention analyzes CYP2D6 and ADRB1 gene polymorphisms, records and collects relevant medical data on metoprolol tolerance, adverse reactions, blood pressure and heart rate responses, and long-term prognosis, performs data processing on the above-collected relevant medical data, screens out multiple features to establish a training data set, and establishes a risk prediction model, uses the training data set to train the risk prediction model to obtain a trained risk prediction model; inputs the drug dosage and test data of the object to be evaluated into the trained risk prediction model to obtain a risk assessment result, thereby providing a theoretical basis for establishing an individualized metoprolol medication regimen guided by gene polymorphism.

[0088] In addition, univariate analysis was used to calculate each feature for the first screening, and the Spierman correlation method was used to calculate the feature correlation for the second screening. The two screening processes can help build a more reliable and interpretable model, simplify the model, remove redundant features, and improve the interpretability of features.

[0089] The following is a system embodiment of the present invention, which can be used to implement the method embodiment of the present invention. For details not disclosed in the system embodiment of the present invention, please refer to the method embodiment of the present invention.

[0090] Figure 4 Schematic diagram of the structure of an example of a drug evaluation system according to the present invention.

[0091] Reference Figure 4 The second aspect of the present disclosure provides a drug evaluation system 400, which adopts the drug evaluation method described in the first aspect of the present disclosure. The drug evaluation system includes a data collection module 410, a feature screening module 420, a first establishment module 430, a second establishment module 440 and a result output result 450.

[0092] Specifically, the data collection module 410 is used to collect historical medical data, and the collected historical medical data includes condition data, drug dosage, physiological index data, and test data. The feature screening module 420 performs data preprocessing on the collected historical medical data, and performs feature screening according to the specified screening rules. The feature screening includes: using univariate analysis to calculate each feature for the first screening, and using the Spierman correlation method to calculate the feature correlation for the second screening. The first establishment module 430 defines positive samples and negative samples based on the multiple features screened out, and establishes a training data set, specifically including: before establishing the training data set, a small number of samples are oversampled so that the number of positive samples and negative samples is equal. The second establishment module 440 is used to establish a risk prediction model, and the risk prediction model is trained using the training data set to obtain a trained risk prediction model. The result output module 450 inputs the drug dosage and test data of the object to be evaluated into the trained risk prediction model to obtain a risk assessment result.

[0093] Compared with the prior art, the present invention analyzes CYP2D6 and ADRB1 gene polymorphisms, records and collects relevant medical data on metoprolol tolerance, adverse reactions, blood pressure and heart rate responses, and long-term prognosis, performs data processing on the above-collected relevant medical data, screens out multiple features to establish a training data set, and establishes a risk prediction model, uses the training data set to train the risk prediction model to obtain a trained risk prediction model; inputs the drug dosage and test data of the object to be evaluated into the trained risk prediction model to obtain a risk assessment result, thereby providing a theoretical basis for establishing an individualized metoprolol medication regimen guided by gene polymorphism.

[0094] Figure 5 is a schematic structural diagram of an electronic device according to an embodiment of the present invention.

[0095] like Figure 5 As shown, the electronic device is presented in the form of a general computing device. The processor may be one or more and work in coordination. The present invention does not exclude distributed processing, that is, the processor may be dispersed in different physical devices. The electronic device of the present invention is not limited to a single entity, but may also be the sum of multiple physical devices.

[0096] The memory stores a computer executable program, which is usually a machine-readable code. The computer-readable program can be executed by the processor to enable the electronic device to perform the method of the present invention, or at least part of the steps in the method.

[0097] The memory includes a volatile memory, such as a random access memory unit (RAM) and / or a cache memory unit, and may also be a non-volatile memory, such as a read-only memory unit (ROM).

[0098] Optionally, in this embodiment, the electronic device further includes an I / O interface, which is used for the electronic device to exchange data with an external device. The I / O interface can represent one or more of several types of bus structures, including a storage unit bus or a storage unit controller, a peripheral bus, a graphics acceleration port, a processing unit, or a local bus using any of a variety of bus structures.

[0099] It should be understood that Figure 5 The electronic device shown is only an example of the present invention, and the electronic device of the present invention may also include elements or components not shown in the above examples. For example, some electronic devices also include display units such as display screens, and some electronic devices also include human-computer interaction elements such as buttons, keyboards, etc. As long as the electronic device can execute the computer-readable program in the memory to implement the method of the present invention or at least part of the steps of the method, it can be considered as an electronic device covered by the present invention.

[0100] Through the above description of the implementation, it is easy for those skilled in the art to understand that the example implementation described here can be implemented by software, or by combining software with necessary hardware. Figure 6 As shown, the technical solution according to the embodiment of the present invention can be embodied in the form of a software product, which can be stored in a non-volatile storage medium (which can be a CD-ROM, a USB flash drive, a mobile hard disk, etc.) or on a network, and includes a number of commands to enable a computing device (which can be a personal computer, a server, or a network device, etc.) to execute the above method according to the embodiment of the present invention.

[0101] The software product may be any combination of one or more readable media. The readable medium may be a readable signal medium or a readable storage medium. The readable storage medium may be, for example, but not limited to, a system, device or component of electricity, magnetism, light, electromagnetic, infrared, or semiconductor, or any combination thereof. More specific examples (non-exhaustive list) of readable storage media include: an electrical connection with one or more wires, a portable disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination thereof.

[0102] The computer readable storage medium may include a data signal propagated in a baseband or as part of a carrier wave, wherein a readable program code is carried. This propagated data signal may take a variety of forms, including but not limited to an electromagnetic signal, an optical signal, or any suitable combination of the above. The readable storage medium may also be any readable medium other than a readable storage medium, which may send, propagate, or transmit a program for use by a command execution system, an apparatus, or a device or for use in combination with the program. The program code contained on the readable storage medium may be transmitted using any suitable medium, including but not limited to wireless, wired, optical cable, RF, etc., or any suitable combination of the above.

[0103] Program code for performing the operations of the present invention may be written in any combination of one or more programming languages, including object-oriented programming languages ​​such as Java, C++, etc., and conventional procedural programming languages ​​such as "C" or similar programming languages. The program code may be executed entirely on the user computing device, partially on the user device, as a separate software package, partially on the user computing device and partially on a remote computing device, or entirely on a remote computing device or server. In the case of a remote computing device, the remote computing device may be connected to the user computing device through any type of network, including a local area network (LAN) or a wide area network (WAN), or may be connected to an external computing device (e.g., via the Internet using an Internet service provider).

[0104] The computer-readable medium carries one or more programs (eg, computer-executable programs). When the one or more programs are executed by a device, the computer-readable medium implements the method of the present disclosure.

[0105] Those skilled in the art will appreciate that the above modules can be distributed in the device according to the description of the embodiment, or can be changed accordingly and only used in one or more devices different from the embodiment. The modules of the above embodiments can be combined into one module, or further divided into multiple sub-modules.

[0106] Through the description of the above embodiments, it is easy for those skilled in the art to understand that the exemplary embodiments described here can be implemented by software, or by software combined with necessary hardware. Therefore, the technical solution according to the embodiment of the present invention can be embodied in the form of a software product, which can be stored in a non-volatile storage medium (which can be a CD-ROM, a USB flash drive, a mobile hard disk, etc.) or on a network, and includes several commands to enable a computing device (which can be a personal computer, a server, a mobile terminal, or a network device, etc.) to execute the method according to the embodiment of the present invention.

[0107] The exemplary embodiments of the present invention are specifically shown and described above. It should be understood that the present invention is not limited to the detailed structure, configuration or implementation method described herein; on the contrary, the present invention is intended to cover various modifications and equivalent configurations included in the spirit and scope of the appended claims.

Claims

1. A drug risk assessment method based on gene polymorphism, characterized in that: include: Collect historical medical data, and analyze the collected historical medical data, including condition data, drug dosage, physiological index data, and test data; Preprocessing the collected historical medical data, and performing feature screening according to specified screening rules, wherein the feature screening includes: calculating each feature by single factor analysis for the first screening, and calculating feature correlation by the Spielman correlation method for the second screening; Based on the screened multiple features, positive samples and negative samples are defined, and a training data set is established, specifically including: before establishing the training data set, a small number of samples are oversampled so that the number of positive samples and negative samples is equal; Establishing a risk prediction model, using a training data set to train the risk prediction model, and obtaining a trained risk prediction model; The drug dosage and test data of the subject to be evaluated are input into the trained risk prediction model to obtain the risk assessment results.

2. The drug risk assessment method based on gene polymorphism according to claim 1, characterized in that: The feature screening according to the specified screening rules includes: Univariate analysis was used to calculate the P value of significance between each feature and the adverse reaction, and the calculated P was compared with the first specific value for the first screening; The specified rule includes: deleting features whose calculated P values ​​are greater than a first specific value.

3. The drug risk assessment method based on gene polymorphism according to claim 2, characterized in that: After the first screening, the feature correlation is calculated using the Spielman correlation method, and the absolute value of the calculated correlation between the features is compared with the second specific value for the second screening; The specified screening rule includes: when the calculated absolute value of the correlation between the selected features is greater than a second specific value, the selected features are deleted.

4. The drug risk assessment method based on gene polymorphism according to claim 3, characterized in that: The first specific value is 0.1, and the second specific value is 0.

7.

5. The drug risk assessment method based on gene polymorphism according to claim 3, characterized in that: The candidate features obtained after the first and second screenings are input into the random forest model to determine the feature combination with the optimal AUC value, that is, the following nine features are obtained: Activity score, age, brain natriuretic peptide, cerebrovascular disease, liver disease, troponin I, heart rate, diastolic blood pressure, atrial fibrillation.

6. The drug risk assessment method based on gene polymorphism according to claim 5, characterized in that: The nine characteristics identified were divided into continuous and discrete variables; For all continuous variables, use the SMOTE method to synthesize negative samples of continuous variables; For each synthesized negative sample, find its k nearest neighbor negative samples for each discrete variable, and calculate the Hamming distance of these negative samples on the category features. Select the nearest neighbor sample with the smallest Hamming distance as a reference, and synthesize the category features of the new sample by randomly selecting a category value different from the reference sample. Combining the type characteristics of the synthesized discrete variables with the continuous variable values, a complete synthetic sample is obtained.

7. The drug risk assessment method based on gene polymorphism according to claim 1, characterized in that: The data preprocessing of the collected historical medical data includes: The CYP2D6 activity score was calculated according to the latest consensus and converted into the CYP2D6 metabolic phenotype; Convert "activity score" into a categorical variable; Convert "maintenance dose" into a categorical variable; The following expression was used to perform exponential transformation on "brain natriuretic peptide, troponin I, troponin T"; ； in, Indicates the data converted into index form for "brain natriuretic peptide, troponin I, troponin T". i Is a positive integer.

8. The drug risk assessment method based on gene polymorphism according to claim 1, characterized in that: include: When applying the risk prediction model, the drug dosage and test data of the subject to be evaluated are input into the trained risk prediction model, and the risk assessment value is output to determine the adverse reaction situation or risk situation.

9. A drug risk assessment system based on gene polymorphism, using the drug risk assessment method based on gene polymorphism according to claim 1, characterized in that: include: A data collection module, which collects historical medical data and processes the collected historical medical data, including condition data, drug dosage, physiological index data, and test data; A feature screening module performs data preprocessing on the collected historical medical data and performs feature screening according to specified screening rules, wherein the feature screening includes: using univariate analysis to calculate each feature for the first screening, and using the Spierman correlation method to calculate feature correlation for the second screening; The first establishment module defines positive samples and negative samples based on the screened multiple features and establishes a training data set, specifically including: before establishing the training data set, oversampling a small number of samples so that the number of positive samples and negative samples is equal; The second module is to establish a risk prediction model, and the risk prediction model is trained using a training data set to obtain a trained risk prediction model; The result output module inputs the drug dosage and test data of the subject to be evaluated into the trained risk prediction model to obtain the risk assessment result.

10. The drug risk assessment system based on gene polymorphism according to claim 9, characterized in that: Further including: Univariate analysis was used to calculate the P value of significance between each feature and the adverse reaction, and the calculated P was compared with the first specific value for the first screening; The specified rule includes: deleting features whose calculated P values ​​are greater than a first specific value.

Images