Method, system and equipment for predicting asthma

Abstract

The invention provides a method, a system and equipment for predicting asthma. The method for predicting asthma is carried out based on a computer and specifically comprises the following steps: acquiring data: acquiring expression level data of biomarkers ATP7B, CDKN2A, GLS, LIAS, LIPT1, MTF1, PDHA1 and PDHB in a sample to be detected; data processing: inputting expression level data of the biomarkers into the constructed asthma prediction model to calculate a risk score; predicting the risk of the to-be-detected sample suffering from asthma based on the risk score, and if the risk score is higher than a threshold value, obtaining a classification result that the risk of the to-be-detected sample suffering from asthma is low; if the risk score is lower than a threshold value, obtaining a classification result of high asthma risk of the to-be-detected sample; and outputting a result.

Description

Technical Field

[0001] This invention belongs to the field of bioinformatics, specifically, it relates to a method, system, and device for predicting asthma. Background Technology

[0002] Asthma is a common, chronic, heterogeneous respiratory disease characterized by chronic airway inflammation, hyperresponsiveness, and airway remodeling. However, asthma is a vague term describing a range of clinical symptoms associated with reversible airflow limitation or bronchial hyperresponsiveness. It is considered a general diagnosis encompassing several diseases, including TH2-asthma, early-onset allergic TH2 asthma, late-onset persistent eosinophilic asthma, exercise-induced asthma, non-TH2 asthma, obesity-related asthma, and neutrophilic asthma, among others, exhibiting diverse phenotypes with different mechanisms. Furthermore, patients with different phenotypes show specific diversity in natural disease course, risk factors, disease severity, and treatment response. Therefore, asthma is highly specific and varies greatly from person to person, and current clinical treatments remain inadequate.

[0003] Therefore, given the highly heterogeneous nature of asthma, it is particularly urgent to develop a method that can accurately distinguish different asthma phenotypes. This can help doctors improve asthma management and achieve a shift from "generalized treatment" to "precision intervention," which has important clinical significance and social value. Summary of the Invention

[0004] In view of this, in order to overcome the shortcomings of the prior art, the present invention is proposed.

[0005] The first aspect of this invention provides a method for predicting asthma, the method being computer-based and specifically including: Data Acquisition: Obtain the expression level data of biomarkers ATP7B, CDKN2A, GLS, LIAS, LIPT1, MTF1, PDHA1, and PDHB in the test samples; Data processing: Input the expression level data of biomarkers into the constructed asthma prediction model to calculate the risk score; The risk score is used to predict the risk of a test sample having asthma. If the risk score is higher than the threshold, the test sample is classified as having a low risk of asthma; if the risk score is lower than the threshold, the test sample is classified as having a high risk of asthma. Output results.

[0006] In some implementations, the asthma prediction model is constructed using machine learning methods.

[0007] In some implementations, the machine learning method includes an algorithmic model developed using development tools.

[0008] In some implementations, the development tools include, but are not limited to, TensorFlow, Scikit-Learn, PyTorch, OpenNN, RapidMiner, Azure Machine Learning, Apache Mahout, Shogun, KNIME, VertexAI, H2Oai, Anaconda, Keras, Tableau, Fast.ai, Catalyst, Amazon ML, MLJAR, and Spell.

[0009] In some implementations, the algorithm model includes, but is not limited to, linear regression model, logistic regression model, Lasso regression model, Ridge regression model, linear discriminant analysis model, nearest neighbor model, decision tree model, perceptron model, neural network model, support vector machine model, Naive Bayes model, AdaBoost model, GBDT model, XGBoost model, LightGBM model, CatBoost model, and random forest model.

[0010] In some implementations, the formula for the asthma prediction model is: riskscore = -0.5504791*ATP7B - 0.1914906*CDKN2A + 1.7046933*GLS + 0.4785051*LIAS - 1.032255*LIPT1 - 0.397675*MTF1 - 1.2189608*PDHA1 + 0.8047971*PDHB.

[0011] A second aspect of the present invention provides a method for distinguishing between moderate and severe asthma, wherein the method for distinguishing between moderate and severe asthma is computer-based and specifically includes: Data Acquisition: Obtain the expression level data of biomarkers ATP7B, CDKN2A, DLAT, DLD, FDX1, MTF1, PDHB, and SLC31A1 in the asthma samples to be tested; The data processing involves inputting biomarker expression level data into the constructed asthma discrimination model to calculate risk scores; Based on the risk score, the classification result is obtained. If the risk score is higher than the threshold, the asthma sample to be tested is classified as having severe asthma; if the risk score is lower than the threshold, the asthma sample to be tested is classified as having moderate asthma. Output results.

[0012] In some implementations, the asthma differentiation model is constructed using machine learning methods.

[0013] In some implementations, the machine learning method includes an algorithmic model developed using development tools.

[0014] In some implementations, the development tools include, but are not limited to, TensorFlow, Scikit-Learn, PyTorch, OpenNN, RapidMiner, Azure Machine Learning, Apache Mahout, Shogun, KNIME, VertexAI, H2Oai, Anaconda, Keras, Tableau, Fast.ai, Catalyst, Amazon ML, MLJAR, and Spell.

[0015] In some implementations, the algorithm model includes, but is not limited to, linear regression model, logistic regression model, Lasso regression model, Ridge regression model, linear discriminant analysis model, nearest neighbor model, decision tree model, perceptron model, neural network model, support vector machine model, Naive Bayes model, AdaBoost model, GBDT model, XGBoost model, LightGBM model, CatBoost model, and random forest model.

[0016] In some implementations, the formula for the asthma discrimination model is: riskscore = 0.41676458*ATP7B + 1.58534551*CDKN2A - 0.86943323*DLAT + 0.14545929*DLD + 0.59052775*FDX1 - 0.42259671*MTF1 - 2.36497694*PDHB - 0.05850206*SLC31A1.

[0017] A third aspect of the present invention provides a method for distinguishing asthma subtypes, the method being computer-based and specifically comprising: Data Acquisition: Acquire the expression level data of one or more biomarkers among ATP7B, CDKN2A, DLAT, DLD, FDX1, GLS, LIAS, LIPT1, MTF1, PDHA1 and / or PDHB in the asthma samples to be tested; Data processing: Input the expression level data of biomarkers into the constructed asthma subtype differentiation model to calculate the risk score; Based on the risk score, the classification result is obtained. If the risk score is higher than the threshold, the test sample is classified as having asthma subtype B; if the risk score is lower than the threshold, the test sample is classified as having asthma subtype A. Output results; The copper mortality level in the cluster A subtype of asthma was significantly higher than that in the cluster B subtype of asthma, and the immune score of the cluster B subtype of asthma was significantly higher than that of the cluster A subtype of asthma.

[0018] In some implementations, the data acquisition refers to acquiring DLD expression level data in the asthma sample to be tested.

[0019] In some implementations, the data acquisition refers to acquiring the expression levels of ATP7B, CDKN2A, DLAT, DLD, FDX1, GLS, LIAS, LIPT1, MTF1, PDHA1, and PDHB in the asthma sample to be tested.

[0020] In some implementations, the asthma subtype differentiation model is constructed using machine learning methods.

[0021] In some implementations, the machine learning method includes an algorithmic model developed using development tools.

[0022] In some implementations, the development tools include, but are not limited to, TensorFlow, Scikit-Learn, PyTorch, OpenNN, RapidMiner, Azure Machine Learning, Apache Mahout, Shogun, KNIME, VertexAI, H2Oai, Anaconda, Keras, Tableau, Fast.ai, Catalyst, Amazon ML, MLJAR, and Spell.

[0023] In some implementations, the algorithm model includes, but is not limited to, linear regression model, logistic regression model, Lasso regression model, Ridge regression model, linear discriminant analysis model, nearest neighbor model, decision tree model, perceptron model, neural network model, support vector machine model, Naive Bayes model, AdaBoost model, GBDT model, XGBoost model, LightGBM model, CatBoost model, and random forest model.

[0024] In some implementations, the formula for the asthma subtype differentiation model is: riskscore = -5.226028*ATP7B + 0.6953184*CDKN2A - 10.892986*DLAT - 6.2257913*DLD - 7.4345691*FDX1 - 2.9635735*GLS - 9.0143665*LIAS - 13.5622949*LIPT1 + 3.1714137*MTF1 + 3.7540659*PDHA1 - 2.1202445*PDHB.

[0025] The fourth aspect of the present invention provides any of the following models: (1) Asthma prediction model, the formula of the asthma prediction model is riskscore=-0.5504791*ATP7B-0.1914906*CDKN2A+1.7046933*GLS+0.4785051*LIAS-1.032255*LIPT1-0.397675*MTF1-1.2189608*PDHA1+0.8047971*PDHB; (2) The model for distinguishing between moderate and severe asthma, wherein the formula for the model for distinguishing between moderate and severe asthma is riskscore=0.41676458*ATP7B+1.58534551*CDKN2A-0.86943323*DLAT+0.14545929*DLD+0.59052775*FDX1-0.42259671*MTF1-2.36497694*PDHB-0.05850206*SLC31A1; (3) Asthma subtype differentiation model, the formula of the asthma subtype differentiation model is riskscore=-5.226028*ATP7B+0.6953184*CDKN2A-10.892986*DLAT-6.2257913*DLD-7.4345691*FDX1-2.9635735*GLS-9.0143665*LIAS-13.5622949*LIPT1+3.1714137*MTF1+3.7540659*PDHA1-2.1202445*PDHB.

[0026] The fifth aspect of the present invention provides any of the following systems: (1) A system for predicting asthma, the system specifically comprising: Data acquisition unit: used to acquire expression level data of biomarkers ATP7B, CDKN2A, GLS, LIAS, LIPT1, MTF1, PDHA1, and PDHB in the sample to be tested; Data classification unit: Used to input biomarker expression level data into the constructed asthma prediction model to calculate risk score; based on the risk score, predict the risk level of the test sample to have asthma. If the risk score is higher than the threshold, the test sample is classified as having a low risk of asthma; if the risk score is lower than the threshold, the test sample is classified as having a high risk of asthma. Result output unit: Used to output classification results; In some implementations, the formula for the asthma prediction model is: riskscore = -0.5504791*ATP7B - 0.1914906*CDKN2A + 1.7046933*GLS + 0.4785051*LIAS - 1.032255*LIPT1 - 0.397675*MTF1 - 1.2189608*PDHA1 + 0.8047971*PDHB; (2) A system for distinguishing between moderate and severe asthma, the system specifically comprising: Data acquisition unit: used to acquire expression level data of biomarkers ATP7B, CDKN2A, DLAT, DLD, FDX1, MTF1, PDHB, and SLC31A1 in the asthma samples to be tested; Data classification unit: used to input biomarker expression level data into the constructed asthma differentiation model to calculate risk score; based on the risk score, the differentiation result is obtained. If the risk score is higher than the threshold, the asthma sample to be tested is classified as having severe asthma; if the risk score is lower than the threshold, the asthma sample to be tested is classified as having moderate asthma. Result output unit: Used to output classification results; In some implementations, the formula for the asthma differentiation model is: riskscore = 0.41676458*ATP7B + 1.58534551*CDKN2A - 0.86943323*DLAT + 0.14545929*DLD + 0.59052775*FDX1 - 0.42259671*MTF1 - 2.36497694*PDHB - 0.05850206*SLC31A1; (3) A system for distinguishing asthma subtypes, the system specifically comprising: Data acquisition unit: used to acquire expression level data of one or more biomarkers among ATP7B, CDKN2A, DLAT, DLD, FDX1, GLS, LIAS, LIPT1, MTF1, PDHA1 and / or PDHB in the asthma sample to be tested; Data classification unit: used to input biomarker expression level data into the constructed asthma subtype differentiation model to calculate risk score; based on the risk score, the differentiation result is obtained. If the risk score is higher than the threshold, the classification result of the test sample having asthma subtype is cluster B; if the risk score is lower than the threshold, the classification result of the test sample having asthma subtype is cluster A. Result output unit: Used to output classification results; The copper mortality level in the cluster A subtype of asthma was significantly higher than that in the cluster B subtype of asthma, and the immune score of the cluster B subtype of asthma was significantly higher than that of the cluster A subtype of asthma.

[0027] In some implementations, the formula for the asthma subtype differentiation model is: riskscore = -5.226028*ATP7B + 0.6953184*CDKN2A - 10.892986*DLAT - 6.2257913*DLD - 7.4345691*FDX1 - 2.9635735*GLS - 9.0143665*LIAS - 13.5622949*LIPT1 + 3.1714137*MTF1 + 3.7540659*PDHA1 - 2.1202445*PDHB.

[0028] The sixth aspect of the present invention provides any of the following products: (1) A computer device / apparatus, the device / apparatus comprising a memory and a processor, the memory being used to store program instructions; the processor being used to invoke the program instructions, which, when executed, implement the method for predicting asthma as described in the first aspect of the present invention, the method for distinguishing between moderate and severe asthma as described in the second aspect of the present invention, or the method for distinguishing asthma subtypes as described in the third aspect of the present invention; (2) A computer-readable storage medium storing a computer program that, when executed by a processor, implements the method for predicting asthma according to the first aspect of the present invention, the method for distinguishing between moderate and severe asthma according to the second aspect of the present invention, or the method for distinguishing asthma subtypes according to the third aspect of the present invention.

[0029] The seventh aspect of the present invention provides any of the following applications: (1) Application of ATP7B, CDKN2A, GLS, LIAS, LIPT1, MTF1, PDHA1, and PDHB in constructing asthma prediction models; (2) Application of ATP7B, CDKN2A, DLAT, DLD, FDX1, MTF1, PDHB, and SLC31A1 in constructing a differential asthma model; (3) Application of ATP7B, CDKN2A, DLAT, DLD, FDX1, GLS, LIAS, LIPT1, MTF1, PDHA1, and PDHB in constructing a model for differentiating asthma subtypes.

[0030] The advantages and beneficial effects of this invention are as follows: This invention provides a method, system, and device for predicting asthma. The method for predicting asthma is computer-based and specifically includes: acquiring data: acquiring expression level data of biomarkers ATP7B, CDKN2A, GLS, LIAS, LIPT1, MTF1, PDHA1, and PDHB in the sample to be tested; processing data: inputting the biomarker expression level data into a constructed asthma prediction model to calculate a risk score; predicting the risk of the sample having asthma based on the risk score; if the risk score is higher than a threshold, a classification result of low risk of asthma is obtained; if the risk score is lower than the threshold, a classification result of high risk of asthma is obtained; and outputting the results. This invention can help doctors improve asthma management and achieve a shift from "generalized treatment" to "precision intervention," and has significant clinical and social value. Attached Figure Description

[0031] Figure 1 This is a schematic flowchart of a method for predicting asthma provided in an embodiment of the present invention; Figure 2 This is a schematic flowchart of a method for distinguishing between moderate and severe asthma provided in an embodiment of the present invention; Figure 3 This is a schematic flowchart of a method for differentiating asthma subtypes provided in an embodiment of the present invention; Figure 4 This is a schematic diagram of a system for predicting asthma provided in an embodiment of the present invention; Figure 5 This is a schematic diagram of a system for distinguishing between moderate and severe asthma, provided by an embodiment of the present invention; Figure 6 This is a schematic diagram of a system for distinguishing asthma subtypes provided in an embodiment of the present invention; Figure 7 This is a schematic diagram of the structure of the computer device / apparatus provided in the embodiments of the present invention; Figure 8 The following are changes in copper death-related genes in asthma: A: Expression of copper death genes in the experimental group; B: Genomic location of copper death genes; C: Correlation between copper death genes in asthma samples; D: Heatmap of copper death genes in control and moderate and severe asthma; E: PPI network of copper death genes. Figure 9 These are diagnostic models for normal, moderate, and severe asthma; A, B: LOSSA screening results; C: ROC curves of the diagnostic models for normal and asthma during training; D, E: LOSSA screening results; F: ROC curves of the diagnostic models for moderate and severe asthma during training. Figure 10These are the results of drug prediction and molecular docking; A: Drug prediction results; B: Conformation with the lowest binding energy when IMATINIB and CDKN2A molecules dock; C: Conformation with the lowest binding energy when DEXAMETHASONE and GLS molecules dock. Figure 11 The following are the immune characteristics in asthma: A: Differences in 23 immune cell types in the experimental group between the control and asthma; B: Differences in 23 immune cell types in the experimental group in moderate to severe asthma; C: Differences in immune activity in the experimental group between the control and asthma; D: Differences in immune activity in the experimental group in moderate to severe asthma; E: Correlation results between 23 immune cell types and copper death genes; F: Differences in 23 immune cell types in the validation group between the two groups. Figure 12 This shows the asthma classification in the experimental and validation groups; AD: consensus matrix heatmap results in the experimental group; EF: consensus clustering results in the experimental group; GJ: consensus matrix heatmap results in the validation group; EF: consensus clustering results in the validation group. Figure 13 This refers to copper death and hypoxia status in two copper death-related subtypes of asthma; A: Differences in copper death genes between the two subtypes in the experimental group; B: Differences in copper death genes between the two subtypes in the validation group; C: Differences in immune scores between the two subtypes; DF: Hypoxia-related gene set GSEA typing; Figure 14 The immune status in two copper mortality-related subtypes of asthma; A: Differences in 23 immune cells between the two subtypes in the experimental group; B: Differences in 23 immune cells between the two subtypes in the validation group; Figure 15 The difference lies in the pathways associated with copper death in the two subtypes of asthma. Figure 16 Volcano plots and heatmaps show the differences between two copper mortality-related subtypes in asthma. Figure 17 These are the WGCNA results for two copper mortality-related subtypes of asthma; A: Sample screening dendrogram; B: Sample dendrogram and feature heatmap; C: Scale-free fit index and average connectivity for different soft threshold powers (β); D: Module merging; E: Genes are divided into different modules through hierarchical clustering, with different colors representing different modules; F: Heatmap showing the correlation between module feature genes and the two asthma models; Figure 18 The results show the core genes and their enrichment; A, B: Scatter plots of Blue and green modules; C: Core genes; D: GO enrichment results; Figure 19 These are diagnostic models for two different subtypes of asthma; A, B: LOSSA screening results; C: ROC curve of the model in the experimental group; D: ROC curve of the model in the validation group; Figure 20 ROC curves for each gene in the experimental group; Figure 21 To verify the ROC curves of each gene in the group. Detailed Implementation

[0032] To enable those skilled in the art to better understand the present invention, the technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings.

[0033] In some of the processes described in the specification, claims, and accompanying drawings of this invention, multiple operations appearing in a specific order are included. However, it should be clearly understood that these operations may not be executed in the order they appear herein, or may be executed in parallel. The operation numbers, such as 101, 102, etc., are merely used to distinguish different operations and do not represent any execution order. Furthermore, these processes may include more or fewer operations, and these operations may be executed sequentially or in parallel. It should be noted that the descriptions such as "first," "second," etc., in this document are used to distinguish different messages, devices, modules, etc., and do not represent a sequential order, nor do they limit "first" and "second" to different types.

[0034] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0035] Figure 1 This is a schematic flowchart of a method for predicting asthma provided by an embodiment of the present invention, which specifically includes the following steps: 101: Data Acquisition: Acquire the expression level data of biomarkers ATP7B, CDKN2A, GLS, LIAS, LIPT1, MTF1, PDHA1, and PDHB in the test samples; In some implementations, the database is processed using R. The search keyword is "asthma," and dataset GSE69683 is included after excluding cell line or animal level studies and single-sample studies. GSE69683 contains blood samples from 87 healthy controls, 77 patients with moderate asthma, and 334 patients with severe asthma.

[0036] In some implementations, analysis is performed by downloading gene expression matrix files submitted by the authors, annotating the gene expression profiles using GPL platform annotation files, and converting gene probes into gene symbols, where the average value is taken for multiple probes corresponding to the same gene.

[0037] In some implementations, the distribution of 12 copper death-related genes on chromosomes is shown, such as Figure 8 As shown in B. Then, the functions of these 12 copper death-related genes were represented as a protein-protein network, and we observed complex interactions between them, such as... Figure 8 As shown in E. A subsequent correlation study of these 12 copper death-related genes revealed a close correlation among them. FDX1 and DLAT showed the highest positive correlation (0.77), while MTF1 and LIPT1 showed the highest negative correlation (-0.46). Figure 8 As shown in C. Finally, differential expression analysis identified altered expression of seven copper death-related genes (CDKN2A, DLAT, GLS, LIAS, LIPT1, PDHA1, PDHB) in control, moderate asthma, and severe asthma, as shown in Figure C. Figure 8 As shown in A and D.

[0038] In some implementations, the expression of these genes is validated in the validation group, where DLAT, PDHA1, and the expression of PDHA1 are consistent in the training group and show significant differences in asthma patients.

[0039] In some implementations, the validation group can choose samples from the database or clinical samples.

[0040] In some implementations, the database includes, but is not limited to, GSE182503 and GSE207751.

[0041] In some implementations, the validation group includes 50 healthy control samples and 92 asthma patient samples.

[0042] In some implementations, the validation group includes 36 healthy control samples and 50 asthma patient samples.

[0043] In some implementations, the validation group includes 36 healthy control samples, 19 samples from patients with severe asthma, and 24 samples from patients with severe asthma.

[0044] In some implementation schemes, to investigate the contribution of copper death-related genes to the pathogenesis of asthma, LASSO regression analysis was performed on 12 copper death genes for feature selection and dimensionality reduction, thereby excluding relatively unimportant copper death genes. Ultimately, we screened out 8 genes: ATP7B, CDKN2A, GLS, LIAS, LIPT1, MTF1, PDHA1, and PDHB. Figure 9 As shown in A and B.

[0045] In the context of this invention, the term "sample" as used refers to a composition obtained from or derived from a patient / subject that contains cells and / or other molecular entities to be characterized and / or identified based on, for example, physical, biochemical, chemical, and / or physiological characteristics. For example, a sample refers to any sample derived from a patient / subject that is expected or known to contain cells and / or molecular entities to be characterized. Samples include, but are not limited to, tissue samples, primary or cultured cells or cell lines, cell cultures, cell supernatants, cell lysates, platelets, serum, plasma, vitreous fluid, lymph, synovial fluid, follicular fluid, semen, pancreatic juice, amniotic fluid, milk, whole blood, blood-derived cells, urine, cerebrospinal fluid, saliva, sputum, tears, sweat, mucus, tissue culture fluid, tissue extracts, homogenized tissue, cell extracts, and combinations thereof.

[0046] In some implementations, the patient may be human or non-human and may include, for example, animal strains or species used as a "model system" for research purposes. Furthermore, the patient may preferably be a mammal (e.g., human or non-human). Examples of mammals include, but are not limited to, any member of the mammalian class: humans, non-human primates (e.g., chimpanzees) and other apes and monkeys; livestock such as cattle, horses, sheep, goats, and pigs; domestic animals such as rabbits, dogs, and cats; and laboratory animals including rodents such as rats, mice, and guinea pigs.

[0047] 102: Data Processing: Input the expression level data of biomarkers into the constructed asthma prediction model to calculate the risk score; The risk score is used to predict the risk of a test sample having asthma. If the risk score is higher than the threshold, the test sample is classified as having a low risk of asthma; if the risk score is lower than the threshold, the test sample is classified as having a high risk of asthma. In some embodiments of the present invention, the model construction method is known to those skilled in the art and can be implemented in different ways to achieve the steps of associating the expression levels of ATP7B, CDKN2A, GLS, LIAS, LIPT1, MTF1, PDHA1, and PDHB with a certain probability or risk.

[0048] In some implementations, the model is constructed using the following steps: Expression data of ATP7B, CDKN2A, GLS, LIAS, LIPT1, MTF1, PDHA1, and PDHB were obtained; the expression data were obtained from asthma patients and healthy controls, and the expression data were input into a machine learning algorithm to construct a model.

[0049] In the context of this invention, the term "machine learning" refers to the use of computers to simulate or implement human learning activities. Technicians typically use various development tools to build algorithmic models for machine learning. These development tools include, but are not limited to, TensorFlow, Scikit-Learn, PyTorch, OpenNN, RapidMiner, Azure Machine Learning, Apache Mahout, Shogun, KNIME, Vertex AI, H2Oai, Anaconda, Keras, Tableau, Fast.ai, Catalyst, Amazon ML, MLJAR, and Spell.

[0050] In the context of this invention, the algorithm model includes, but is not limited to, linear regression model, logistic regression model, Lasso regression model, Ridge regression model, linear discriminant analysis model, nearest neighbor model, decision tree model, perceptron model, neural network model, support vector machine model, Naive Bayes model, AdaBoost model, GBDT model, XGBoost model, LightGBM model, CatBoost model, or random forest model.

[0051] In the context of this invention, the construction step of the evaluation model further includes verifying the model's effectiveness.

[0052] In the context of this invention, the method for verifying model performance includes ROC curves.

[0053] The ROC curve represents the true positive rate (sensitivity) of the test relative to the false positive rate (100%). A plot of specific characteristics. This is useful for depicting the performance of a particular characteristic when distinguishing between two populations. Typically, characteristic data are selected across the entire population in ascending order based on the values ​​of a single characteristic. Then, for each value of that characteristic, the true positive and false positive rates of the data are calculated. The true positive rate is determined by counting the number of cases with values ​​higher than the characteristic and dividing by the total number of cases. The false positive rate is determined by counting the number of controls with values ​​higher than the characteristic and dividing by the total number of controls. While this definition refers to cases where the characteristic is higher in cases compared to controls, it also applies to cases where the characteristic is lower in cases compared to controls (in which case samples with values ​​lower than the characteristic are counted). ROC curves can be generated with respect to individual characteristics and can also be generated with respect to other individual outputs. For example, combinations of two or more characteristics can be mathematically combined (e.g., addition, subtraction, multiplication, etc.) to provide a single sum value, which can be plotted on the ROC curve. Additionally, any combination of multiple characteristics derived from individual output values ​​can be plotted on the ROC curve.

[0054] In some implementation schemes, diagnostic models for normal individuals and asthma were constructed based on the corresponding gene coefficients and expression levels. The RiskScore AUC value was 0.718, which effectively distinguished between healthy and asthmatic samples based on the risk score. Figure 9 As shown in C, the AUC of each gene is shown in Table 1.

[0055] Table 1. AUC values ​​of diagnostic models for normal individuals and asthma.

[0056] 103: Output results.

[0057] Figure 2 This is a schematic flowchart of a method for distinguishing between moderate and severe asthma provided in an embodiment of the present invention, specifically including: 201: Data Acquisition: Acquire the expression level data of biomarkers ATP7B, CDKN2A, DLAT, DLD, FDX1, MTF1, PDHB, and SLC31A1 in the asthma samples to be tested; 202: Data Processing: Input the expression level data of biomarkers into the constructed asthma discrimination model to calculate the risk score; Based on the risk score, the classification result is obtained. If the risk score is higher than the threshold, the asthma sample to be tested is classified as having severe asthma; if the risk score is lower than the threshold, the asthma sample to be tested is classified as having moderate asthma. In some implementation schemes, the contribution of copper death-related genes in moderate and severe asthma was also explored. LASSO regression analysis was performed on 12 copper death genes for feature selection and dimensionality reduction, thereby excluding relatively unimportant copper death genes. Ultimately, we screened out 8 genes: ATP7B, CDKN2A, DLAT, DLD, FDX1, MTF1, PDHB, and SLC31A1. Figure 9 As shown in D and E. Subsequently, diagnostic models for moderate and severe asthma were constructed based on the corresponding gene coefficients and expression levels. The RiskScore2 AUC value was 0.721, which effectively distinguishes between moderate and severe asthma samples based on the risk score. Figure 9 As shown in F, copper death does indeed play a key role in the development and progression of asthma. The AUC of each gene is shown in Table 2.

[0058] Table 2. AUC values ​​of diagnostic models for moderate and severe asthma.

[0059] 203: Output result.

[0060] In some implementation schemes, seven copper death-related genes whose expression was altered in controls, moderate asthma, and severe asthma were also entered into the DGIdb database (https: / / dgidb.org / ) for drug lookup. Ultimately, two genes and their corresponding 32 drugs were found, such as... Figure 10 As shown in Figure A. After reviewing the literature, we found that IMATINIB has been shown in clinical trials in patients with severe asthma to improve forced expiratory volume (FEV1) within 1 second, while DEXAMETHASONE is one of the widely used first-line drugs for treating asthma. We also performed molecular docking between IMATINIB and its target gene CDKN2A, and DEXAMETHASONE and its target gene GLS. The optimal conformation was observed when the binding energy between the drug and the target protein was lowest, revealing the interaction mode between the drug molecule and the target protein. The lowest binding energy for IMATINIB when docking with CDKN2A was -7.01 kcal / mol, forming three hydrogen bonds with ALA-148 and GLU-149 residues. The results are as follows... Figure 10 As shown in B, the drug molecule DEXAMETHASONE exhibits a minimum binding energy of -6.52 kcal / mol during molecular docking with GLS, forming two hydrogen bonds with GLU-325 and GLU-312 residues, as shown in Figure B. Figure 10 As shown in C.

[0061] Figure 3 This is a schematic flowchart of a method for distinguishing asthma subtypes provided in an embodiment of the present invention, specifically including: 301: Data Acquisition: Acquire the expression level data of biomarkers ATP7B, CDKN2A, DLAT, DLD, FDX1, GLS, LIAS, LIPT1, MTF1, PDHA1, and PDHB in the asthma samples to be tested; In some implementations, to explore the biological behavior between copper death and the immune characteristics of asthma, we analyzed the differences in the populations of specific infiltrating immune cells and the activity of specific immune responses in control and asthma samples using ssGSEA. The correlation between copper death-related genes and immune cells was then analyzed. The results showed that, compared to the control group, Activated B cells and Immature B cells were significantly reduced in asthma samples, while Activated dendritic cells, Eosinophil, Gamma delta T cells, Immature dendritic cells, Macrophage, Mast cells, MDSCs, Neutrophil, Regulatory T cells, and Type 17 T helper cells were significantly increased in asthma samples. The MDSC levels in the validation group were completely consistent with those in the experimental group; detailed results can be found in [link to relevant documentation]. Figure 11 A, F. The immune-related activities included interleukins receptors, antimicrobials, BCR signaling pathway, TNF family members, chemokines, cytokines, interferons, and interleukins. Significant differences were observed between the control and asthma groups, as shown in the results below. Figure 11 C.

[0062] Furthermore, compared to moderate asthma, the levels of activated B cells, CD56bright natural killer cells, CD56dim natural killer cells, immature B cells, immature dendritic cells, and Tfollicular helper cells were significantly reduced in severe asthma samples, while macrophage, Mast cells, MDSCs, Neutrophil, and Type 17 T helper cells were significantly increased in severe asthma samples. (See attached figures). Figure 11 B. Significant differences were observed in immune-related factors such as antimicrobials, antigen processing and presentation, BCRS indicating pathway, chemokine receptors, chemokines, cytokines, interferon receptors, and interferons in moderate to severe asthma. (See attached results). Figure 11 D.

[0063] Eight immune cell types—Activated B cell, Immature B cell, Immature dendritic cell, Macrophage, Mast cell, MDSC, Neutrophil, and Type 17 T helper cell—showed significant differences between normal individuals and those with asthma, as well as those with moderate to severe asthma. Subsequently, we analyzed the correlation between these eight immune cell types and 12 copper death genes. The results showed that MDSC and Neutrophil were significantly correlated with all 12 copper death genes. Furthermore, MTF1 showed the strongest positive correlation with Macrophage, while PDHB showed the strongest negative correlation with Neutrophil. (See attached figures). Figure 11 E.

[0064] In some implementations, based on the expression values ​​of copper death-related genes in each asthma sample of the experimental group dataset, consensus clustering is performed according to PAM and Euclidean. When k = 2, the consensus matrix heatmap remains clear and has well-defined boundaries, indicating that the samples have stable and robust clusters. The optimal total number of clusters is chosen as k = 2 (the two subgroups are referred to as clusterA and clusterB, respectively). Subgroups A and B contain 238 and 173 asthma patients, respectively. Figure 12 ).

[0065] Furthermore, using the same method, two asthma subtypes with different copper mortality patterns were validated in the validation group, yielding similarly good classification results. The cluster A and culture B subgroups contained 38 and 54 asthma patients, respectively. Consensus clustering revealed significant differences in molecular characteristics between the two asthma cluster subtypes.

[0066] In some implementation schemes, to investigate the relationship between the two asthma subtypes and copper death, we compared the expression differences of 12 copper death-related genes between the cluster A and culture B subgroups. ATP7B, DLAT, DLD, FDX1, GLS, LIAS, LIPT1, and PDHB were significantly underexpressed in the asthma cluster B sample, while CDKN2A, MTF1, and SLC31A1 were significantly overexpressed. A total of 6 pro-copper death genes were significantly underexpressed in the cluster B sample, and 2 anti-copper death genes were significantly overexpressed. Furthermore, the expression levels of ATP7B, DLAT, DLD, FDX1, GLS, LIAS, LIPT1, and PDHB in the validation group were completely consistent with those in the experimental group. Detailed results can be found in [link to results]. Figure 13 In conclusion, copper death is more pronounced in cluster A, B, and A.

[0067] Furthermore, mounting evidence suggests that elevated hypoxia increases the severity of asthma inflammation in patients and model systems, and that copper-induced cell death is attenuated under hypoxic conditions, indicating that copper death also requires mitochondrial respiration. Therefore, we selected three hypoxia-related gene sets (HALLMARK HYPOXIA, WINTER HYPOXIA DN, and MANALO HYPOXIA DN) for GSEA analysis to assess hypoxia status among different subtypes of asthma copper death. HALLMARK HYPOXIA showed an upregulation trend in cluster B, while WINTER HYPOXIA DN and MANALO HYPOXIA DN were significantly upregulated in cluster A. In conclusion, hypoxia was more pronounced in cluster B; detailed results can be found in [link to results]. Figure 13 D, E, and F. Furthermore, we calculated the immune score for each asthma patient using the ESTIMATE algorithm, finding a more significant immune response in the culture B subtype. Detailed results can be found in [link to results]. Figure 13 C.

[0068] In some implementations, to explore the biological behavior between the immune characteristics of different copper death pattern subtypes in asthma, we analyzed the differences in specific infiltrating immune cell populations of the two subtypes in asthma samples using ssGSEA. The results showed that, compared to cluster A, Activated CD4 T cells, Activated CD8 T cells, Gammadelta T cells, Immature B cells, Immature dendritic cells, and Regulatory T cells were significantly reduced in asthma cluster B samples, while Activated dendritic cells, CD56dim natural killer cells, Macrophage, Mast cells, MDSCs, Monocytes, Natural killer cells, Natural killer T cells, Neutrophil, Plasmacytoid dendritic cells, T follicular helper cells, and Type 2 Thelper cells were significantly increased in asthma cluster B samples. The levels of Activated CD4 T cells, Naturalalkiller cells, and Neutrophil in the validation group were completely consistent with those in the experimental group, as shown in the results below. Figure 14 As shown.

[0069] In some implementations, GSVA analysis was also performed to reveal differences in metabolic pathways between asthma patients with two different copper mortality patterns. Differentially expressed pathways were calculated using the R package "limma" and screened with an FDR < 0.05. A total of 146 metabolic pathways were screened, of which 47 pathways were more active in cluster A and 99 pathways were more active in cluster B. The top 10 pathways that were more active in each of the two asthma subtypes were selected for presentation. The pathways PROTEINEXPORT, GLYCOSYLPHOSPHATIDYLINOSITOL GPI ANCHOR BIOSYNTHESIS, RNA DEGRADATION, AMINOACYL tRNA BIOSYNTHESIS, and MISMATCH REPAIR were more active in cluster A, while the pathways TAURINE AND HYPOTAURINE METABOLISM, ECM RECEPTOR INTERACTION, COMPLEMENT AND COAGULATION CASCADES, OLFACTORY TRANSDUCTION, and GLYCOSAMINOGLYCAN DEGRADATION were more active in cluster B. The results are as follows: Figure 15 As shown.

[0070] In some implementations, the limma package in R software was used to group individuals according to two non-copper mortality-related subtypes of asthma and perform differential gene expression analysis. The screening thresholds for differential gene expression were set as FDR < 0.05 and |log2(FC)| > 0.7. A total of 481 differentially expressed genes were identified, including 38 upregulated genes and 443 downregulated genes. Figure 16 As shown.

[0071] In some implementation schemes, genes with the top 25% coefficients of variation in 411 asthma samples were selected for constructing a weighted gene co-expression network. Given WGCNA's sensitivity to outliers, a rigorous quality control procedure was implemented to ensure the highest quality level. First, the samples were clustered using the average linkage method in WGCNA, with 30 selected as the cleavage tree height, resulting in the removal of 31 samples. The results are as follows... Figure 17 As shown in Figure A. The dendrogram of the remaining 380 samples and the feature heatmaps of subtypes A and B are shown below. Figure 17 As shown in B, we choose β=24 ( Figure 17C) Using a soft threshold as a basis for constructing a scale-free network, the minimum number of genes per module was set to 50. Modules with dissimilarity <25% were merged using a dynamic tree cutting method, ultimately resulting in 10 modules. Figure 17 As shown in D and E. Two subtypes were extracted as phenotypic data, and their association with the WGCNA module was analyzed. Genes from the Blue (1252 genes) and Green (292 genes) modules were most correlated with the two subtypes, as shown in the results. Figure 17 As shown in F.

[0072] In some implementations, genes were extracted from the Blue and green modules. Based on MM > 0.8 and GS > 0.5, we identified 844 genes (672 in Blue and 172 in Green) and two copper-related mortality subtypes associated with asthma. The results are as follows: Figure 18 As shown in A and B. Taking the intersection of these 844 genes with the differentially expressed genes (FDR < 0.05 and |log2(FC)| > 0.7) between the two copper mortality-related subtypes of asthma yielded 329 core genes, as shown in Figure 1. Figure 18 As shown in Figure C. Enrichment analysis of the core genes was performed using the David database to investigate their potential biochemical processes and pathways. Significance was determined by the p-value; p < 0.05 was considered meaningful. The results showed that the core genes were mainly distributed in cellular components such as nucleus, cytosol, cytoplasm, and nucleoplasm. Through molecular functions such as protein binding, metal ion binding, ubiquitin-protein transferase activity, and histone binding, they participate in biological processes such as regulation of transcription from RNA polymerase II promoter, regulation of transcription, DNA-templated, protein ubiquitination, DNA repair, and chromatin remodeling. The results are shown in Figure C. Figure 18 As shown in D.

[0073] In some implementations, based on the grouping of asthma samples into two subtypes, we performed LASSO regression analysis on 12 copper death genes for feature selection and dimensionality reduction, thereby excluding relatively unimportant copper death genes. Ultimately, we screened out 11 genes: ATP7B, CDKN2A, DLAT, DLD, FDX1, GLS, LIAS, LIPT1, MTF1, PDHA1, and PDHB. Figure 19As shown in A and B.

[0074] 302: Data Processing: Input the expression level data of biomarkers into the constructed asthma subtype differentiation model to calculate the risk score; Based on the risk score, the classification result is obtained. If the risk score is higher than the threshold, the test sample is classified as having asthma subtype B; if the risk score is lower than the threshold, the test sample is classified as having asthma subtype A. In some implementation schemes, a diagnostic model for copper death-related subtypes in asthma was subsequently constructed based on the corresponding gene coefficients and expression levels. This model consists of 11 copper death genes. In the experimental group, the AUC value of RiskScore3 was 1, and in the validation group, the AUC value of RiskScore3 was 0.906. RiskScore3 can effectively distinguish between two different copper death-related subtypes in asthma samples, such as... Figure 19 As shown in C and D, the AUC values ​​of the experimental groups are shown in Table 3. Figure 20 The AUCs in the validation group are shown in Table 4. Figure 21 ).

[0075] Table 3 AUC values ​​in the experimental group

[0076] Table 4 AUC values ​​of the validation group

[0077] 303: Output result; The copper mortality level in the cluster A subtype of asthma was significantly higher than that in the cluster B subtype of asthma, and the immune score of the cluster B subtype of asthma was significantly higher than that of the cluster A subtype of asthma.

[0078] Figure 4 This is a schematic diagram of a system for predicting asthma provided in an embodiment of the present invention; The system is programmed or otherwise configured to include a data acquisition unit 401, a data classification unit 402, and a result output unit 403. 401: Data Acquisition Unit: Used to acquire expression level data of biomarkers ATP7B, CDKN2A, GLS, LIAS, LIPT1, MTF1, PDHA1, and PDHB in the sample to be tested; 402: Data Classification Unit: Used to input biomarker expression level data into the constructed asthma prediction model to calculate risk score; based on the risk score, it predicts the risk level of the test sample to have asthma. If the risk score is higher than the threshold, the test sample is classified as having a low risk of asthma; if the risk score is lower than the threshold, the test sample is classified as having a high risk of asthma. 403: Result Output Unit: Used to output classification results.

[0079] In some implementations, the formula for the asthma prediction model is: riskscore = -0.5504791*ATP7B - 0.1914906*CDKN2A + 1.7046933*GLS + 0.4785051*LIAS - 1.032255*LIPT1 - 0.397675*MTF1 - 1.2189608*PDHA1 + 0.8047971*PDHB.

[0080] The system may be a user's electronic device or a computer system remotely located relative to that electronic device.

[0081] Figure 5 This is a schematic diagram of a system for distinguishing between moderate and severe asthma, provided by an embodiment of the present invention; The system is programmed or otherwise configured to include a data acquisition unit 501, a data classification unit 502, and a result output unit 503. 501: Data Acquisition Unit: Used to acquire the expression level data of biomarkers ATP7B, CDKN2A, DLAT, DLD, FDX1, MTF1, PDHB, and SLC31A1 in the asthma sample to be tested; 502: Data Classification Unit: Used to input biomarker expression level data into the constructed asthma discrimination model to calculate risk scores; based on the risk scores, the discrimination results are obtained. If the risk score is higher than the threshold, the asthma sample to be tested is classified as having severe asthma; if the risk score is lower than the threshold, the asthma sample to be tested is classified as having moderate asthma. 503: Result Output Unit: Used to output classification results.

[0082] In some implementations, the formula for the asthma discrimination model is: riskscore = 0.41676458*ATP7B + 1.58534551*CDKN2A - 0.86943323*DLAT + 0.14545929*DLD + 0.59052775*FDX1 - 0.42259671*MTF1 - 2.36497694*PDHB - 0.05850206*SLC31A1.

[0083] The system may be a user's electronic device or a computer system remotely located relative to that electronic device.

[0084] Figure 6 This is a schematic diagram of a system for distinguishing asthma subtypes provided in an embodiment of the present invention; The system is programmed or otherwise configured to include a data acquisition unit 601, a data classification unit 602, and a result output unit 603. 601: Data Acquisition Unit: Used to acquire expression level data of biomarkers ATP7B, CDKN2A, DLAT, DLD, FDX1, GLS, LIAS, LIPT1, MTF1, PDHA1, and PDHB in the asthma sample to be tested; 602: Data Classification Unit: Used to input biomarker expression level data into the constructed asthma subtype differentiation model to calculate risk scores; based on the risk scores, the differentiation results are obtained. If the risk score is higher than the threshold, the classification result of the test sample having asthma subtype B is obtained; if the risk score is lower than the threshold, the classification result of the test sample having asthma subtype A is obtained. 603: Result Output Unit: Used to output classification results; The copper mortality level in the cluster A subtype of asthma was significantly higher than that in the cluster B subtype of asthma, and the immune score of the cluster B subtype of asthma was significantly higher than that of the cluster A subtype of asthma.

[0085] In some implementations, the formula for the asthma subtype differentiation model is: riskscore = -5.226028*ATP7B + 0.6953184*CDKN2A - 10.892986*DLAT - 6.2257913*DLD - 7.4345691*FDX1 - 2.9635735*GLS - 9.0143665*LIAS - 13.5622949*LIPT1 + 3.1714137*MTF1 + 3.7540659*PDHA1 - 2.1202445*PDHB.

[0086] Figure 7 This is a schematic diagram of the structure of the computer device / appliance provided in the embodiments of the present invention.

[0087] The computer device 700 includes a processor 701 and a memory 702 coupled to the processor 701. The memory 702 stores program instructions. When the program instructions are executed by the processor 701, the processor 701 performs the method for predicting asthma as described in the first aspect of the present invention, the method for distinguishing between moderate and severe asthma as described in the second aspect of the present invention, or the method for distinguishing asthma subtypes as described in the third aspect of the present invention.

[0088] The processor 701 can also be referred to as a CPU (Central Processing Unit). The processor 701 may be an integrated circuit chip with signal processing capabilities. The processor 701 can also be a general-purpose processor, a digital signal processor (DSP), an application-specific integrated circuit (ASIC), a field-programmable gate array (FPGA), or other programmable logic devices, discrete gate or transistor logic devices, or discrete hardware components. A general-purpose processor can be a microprocessor or any conventional processor.

[0089] Computer device 700 can be a mobile electronic device.

[0090] It should be understood that the systems, apparatuses, and methods described in this invention can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative; for instance, the division of modules is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple modules or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the couplings or direct couplings or communication connections shown or discussed may be indirect couplings or communication connections through some interfaces, apparatuses, or modules, and may be electrical, mechanical, or other forms.

[0091] The modules described as separate components may or may not be physically separate. The components shown as modules may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0092] Furthermore, the functional modules in the various embodiments of the present invention can be integrated into one processing module, or each module can exist physically separately, or two or more modules can be integrated into one module. The integrated modules described above can be implemented in hardware or as software functional modules.

[0093] The above are merely embodiments of this application and do not limit the scope of this patent application. Any equivalent structural or procedural changes made using the content of this application's specification and drawings, or direct or indirect applications in other related technical fields, are similarly included within the scope of patent protection of this application.

Claims

1. A method for predicting asthma, characterized in that, The method for predicting asthma is computer-based and specifically includes: Data Acquisition: Obtain the expression level data of biomarkers ATP7B, CDKN2A, GLS, LIAS, LIPT1, MTF1, PDHA1, and PDHB in the test samples; Data processing: Input the expression level data of biomarkers into the constructed asthma prediction model to calculate the risk score; The risk score is used to predict the risk of a test sample having asthma. If the risk score is higher than the threshold, the test sample is classified as having a low risk of asthma; if the risk score is lower than the threshold, the test sample is classified as having a high risk of asthma. Output results.

2. The method for predicting asthma according to claim 1, characterized in that, The asthma prediction model is constructed using machine learning methods; Optionally, the machine learning method includes an algorithmic model developed using development tools; Optionally, the development tools include TensorFlow, Scikit-Learn, PyTorch, OpenNN, RapidMiner, Azure Machine Learning, Apache Mahout, Shogun, KNIME, Vertex AI, H2Oai, Anaconda, Keras, Tableau, Fast.ai, Catalyst, Amazon ML, MLJAR, and Spell. Optionally, the algorithm model includes linear regression model, logistic regression model, Lasso regression model, Ridge regression model, linear discriminant analysis model, nearest neighbor model, decision tree model, perceptron model, neural network model, support vector machine model, Naive Bayes model, AdaBoost model, GBDT model, XGBoost model, LightGBM model, CatBoost model, and random forest model; Optionally, the formula for the asthma prediction model is: riskscore = -0.5504791*ATP7B - 0.1914906*CDKN2A + 1.7046933*GLS + 0.4785051*LIAS - 1.032255*LIPT1 - 0.397675*MTF1 - 1.2189608*PDHA1 + 0.8047971*PDHB.

3. A method for distinguishing between moderate and severe asthma, characterized in that, The method for distinguishing between moderate and severe asthma is computer-based and specifically includes: Data Acquisition: Obtain the expression level data of biomarkers ATP7B, CDKN2A, DLAT, DLD, FDX1, MTF1, PDHB, and SLC31A1 in the asthma samples to be tested; Data processing: Input the expression level data of biomarkers into the constructed asthma discrimination model to calculate the risk score; The classification results are obtained based on the risk score. If the risk score is higher than the threshold, the asthma sample to be tested is classified as having severe asthma; if the risk score is lower than the threshold, the asthma sample to be tested is classified as having moderate asthma. Output results.

4. The method for distinguishing between moderate and severe asthma according to claim 2, characterized in that, The asthma differentiation model was constructed using machine learning methods. Optionally, the machine learning method includes an algorithmic model developed using development tools; Optionally, the development tools include TensorFlow, Scikit-Learn, PyTorch, OpenNN, RapidMiner, Azure Machine Learning, Apache Mahout, Shogun, KNIME, Vertex AI, H2Oai, Anaconda, Keras, Tableau, Fast.ai, Catalyst, Amazon ML, MLJAR, and Spell. Optionally, the algorithm model includes linear regression model, logistic regression model, Lasso regression model, Ridge regression model, linear discriminant analysis model, nearest neighbor model, decision tree model, perceptron model, neural network model, support vector machine model, Naive Bayes model, AdaBoost model, GBDT model, XGBoost model, LightGBM model, CatBoost model, and random forest model; Optionally, the formula for the asthma differentiation model is: riskscore = 0.41676458*ATP7B + 1.58534551*CDKN2A - 0.86943323*DLAT + 0.14545929*DLD + 0.59052775*FDX1 - 0.42259671*MTF1 - 2.36497694*PDHB - 0.05850206*SLC31A1.

5. A method for differentiating asthma subtypes, characterized in that, The method for distinguishing asthma subtypes is computer-based and specifically includes: Data Acquisition: Acquire the expression level data of one or more biomarkers among ATP7B, CDKN2A, DLAT, DLD, FDX1, GLS, LIAS, LIPT1, MTF1, PDHA1 and / or PDHB in the asthma samples to be tested; Data processing: Input the expression level data of biomarkers into the constructed asthma subtype differentiation model to calculate the risk score; Based on the risk score, the classification result is obtained. If the risk score is higher than the threshold, the test sample is classified as having asthma subtype B; if the risk score is lower than the threshold, the test sample is classified as having asthma subtype A. Output results; The copper mortality level in the cluster A subtype of asthma was significantly higher than that in the cluster B subtype of asthma, and the immune score of the cluster B subtype of asthma was significantly higher than that of the cluster A subtype of asthma.

6. The method for distinguishing asthma subtypes according to claim 5, characterized in that, The asthma subtype differentiation model was constructed using machine learning methods. Optionally, the machine learning method includes an algorithmic model developed using development tools; Optionally, the development tools include TensorFlow, Scikit-Learn, PyTorch, OpenNN, RapidMiner, Azure Machine Learning, Apache Mahout, Shogun, KNIME, Vertex AI, H2Oai, Anaconda, Keras, Tableau, Fast.ai, Catalyst, Amazon ML, MLJAR, and Spell. Optionally, the algorithm model includes linear regression model, logistic regression model, Lasso regression model, Ridge regression model, linear discriminant analysis model, nearest neighbor model, decision tree model, perceptron model, neural network model, support vector machine model, Naive Bayes model, AdaBoost model, GBDT model, XGBoost model, LightGBM model, CatBoost model, and random forest model; Optionally, the formula for the asthma subtype differentiation model is: riskscore = -5.226028*ATP7B + 0.6953184*CDKN2A - 10.892986*DLAT - 6.2257913*DLD - 7.4345691*FDX1 - 2.9635735*GLS - 9.0143665*LIAS - 13.5622949*LIPT1 + 3.1714137*MTF1 + 3.7540659*PDHA1 - 2.1202445*PDHB.

7. Any of the following models: (1) An asthma prediction model, characterized in that, The formula for the asthma prediction model is: riskscore = -0.5504791*ATP7B - 0.1914906*CDKN2A + 1.7046933*GLS + 0.4785051*LIAS - 1.032255*LIPT1 - 0.397675*MTF1 - 1.2189608*PDHA1 + 0.8047971*PDHB; (2) A model for distinguishing between moderate and severe asthma, characterized in that the formula for the model is riskscore=0.41676458*ATP7B+1.58534551*CDKN2A-0.86943323*DLAT+0.14545929*DLD+0.59052775*FDX1-0.42259671*MTF1-2.36497694*PDHB-0.05850206*SLC31A1; (3) Asthma subtype differentiation model, characterized in that the formula of the asthma subtype differentiation model is riskscore=-5.226028*ATP7B+0.6953184*CDKN2A-10.892986*DLAT-6.2257913*DLD-7.4345691*FDX1-2.9635735*GLS-9.0143665*LIAS-13.5622949*LIPT1+3.1714137*MTF1+3.7540659*PDHA1-2.1202445*PDHB.

8. Any of the following systems: (1) A system for predicting asthma, characterized in that, The system specifically includes: Data acquisition unit: used to acquire expression level data of biomarkers ATP7B, CDKN2A, GLS, LIAS, LIPT1, MTF1, PDHA1, and PDHB in the sample to be tested; Data classification unit: Used to input biomarker expression level data into the constructed asthma prediction model to calculate risk score; based on the risk score, predict the risk level of the test sample to have asthma. If the risk score is higher than the threshold, the test sample is classified as having a low risk of asthma; if the risk score is lower than the threshold, the test sample is classified as having a high risk of asthma. Result output unit: Used to output classification results; Optionally, the formula for the asthma prediction model is: riskscore = -0.5504791*ATP7B - 0.1914906*CDKN2A + 1.7046933*GLS + 0.4785051*LIAS - 1.032255*LIPT1 - 0.397675*MTF1 - 1.2189608*PDHA1 + 0.8047971*PDHB; (2) A system for distinguishing between moderate and severe asthma, characterized in that the system specifically comprises: Data acquisition unit: used to acquire expression level data of biomarkers ATP7B, CDKN2A, DLAT, DLD, FDX1, MTF1, PDHB, and SLC31A1 in the asthma samples to be tested; Data classification unit: used to input biomarker expression level data into the constructed asthma differentiation model to calculate risk score; based on the risk score, the differentiation result is obtained. If the risk score is higher than the threshold, the asthma sample to be tested is classified as having severe asthma; if the risk score is lower than the threshold, the asthma sample to be tested is classified as having moderate asthma. Result output unit: Used to output classification results; Optionally, the formula for the asthma differentiation model is: riskscore = 0.41676458*ATP7B + 1.58534551*CDKN2A - 0.86943323*DLAT + 0.14545929*DLD + 0.59052775*FDX1 - 0.42259671*MTF1 - 2.36497694*PDHB - 0.05850206*SLC31A1; (3) A system for distinguishing asthma subtypes, characterized in that the system specifically includes: Data acquisition unit: used to acquire expression level data of one or more biomarkers among ATP7B, CDKN2A, DLAT, DLD, FDX1, GLS, LIAS, LIPT1, MTF1, PDHA1 and / or PDHB in the asthma sample to be tested; Data classification unit: used to input biomarker expression level data into the constructed asthma subtype differentiation model to calculate risk score; based on the risk score, the differentiation result is obtained. If the risk score is higher than the threshold, the classification result of the test sample having asthma subtype is cluster B; if the risk score is lower than the threshold, the classification result of the test sample having asthma subtype is cluster A. Result output unit: Used to output classification results; The copper mortality level in the cluster A subtype of asthma was significantly higher than that in the cluster B subtype of asthma, and the immune score of the cluster B subtype of asthma was significantly higher than that of the cluster A subtype of asthma. Optionally, the formula for the asthma subtype differentiation model is: riskscore = -5.226028*ATP7B + 0.6953184*CDKN2A - 10.892986*DLAT - 6.2257913*DLD - 7.4345691*FDX1 - 2.9635735*GLS - 9.0143665*LIAS - 13.5622949*LIPT1 + 3.1714137*MTF1 + 3.7540659*PDHA1 - 2.1202445*PDHB.

9. Any one of the following products: (1) A computer device / apparatus, characterized in that, The device / apparatus includes a memory and a processor, the memory being used to store program instructions; the processor being used to invoke the program instructions, which, when executed, implement the method for predicting asthma as described in any one of claims 1-2, the method for distinguishing between moderate and severe asthma as described in any one of claims 3-4, or the method for distinguishing asthma subtypes as described in any one of claims 5-6. (2) A computer-readable storage medium, characterized in that the computer-readable storage medium stores a computer program, which, when executed by a processor, implements the method for predicting asthma as described in any one of claims 1-2, the method for distinguishing between moderate and severe asthma as described in any one of claims 3-4, or the method for distinguishing asthma subtypes as described in any one of claims 5-6.

10. Any of the following applications: (1) Application of ATP7B, CDKN2A, GLS, LIAS, LIPT1, MTF1, PDHA1, and PDHB in constructing asthma prediction models; (2) Application of ATP7B, CDKN2A, DLAT, DLD, FDX1, MTF1, PDHB, and SLC31A1 in constructing a differential asthma model; (3) Application of ATP7B, CDKN2A, DLAT, DLD, FDX1, GLS, LIAS, LIPT1, MTF1, PDHA1, and PDHB in constructing a model for differentiating asthma subtypes.

Images