A respiratory pathogen detection kit and method

Abstract

The present invention relates to the technical field of respiratory pathogen detection, and specifically to a respiratory pathogen detection kit and method. In the present invention, pathogen sample data is acquired and multiple detections are performed through a data collection module to generate PCR data, immune data and microarray data; a detection target analysis module fuses PCR data, immune data and microarray data to generate real-time target fusion data, and a decision tree algorithm is used to predict the type of pathogen and analyze the importance of features; a customized detection module adopts a targeted data fusion strategy according to the prediction results, and selects a corresponding detection method. This method realizes full-process personalization from data processing to detection method selection, significantly improves the accuracy, efficiency and pertinence of detection, is particularly suitable for rapid diagnosis of complex and emerging pathogens, and provides a reliable basis for clinical treatment.

Description

Technical Field

[0001] The present invention relates to the technical field of respiratory pathogen detection, and in particular to a respiratory pathogen detection kit and method. Background Art

[0002] Traditional respiratory pathogen detection methods have multiple technical problems, resulting in low detection efficiency and insufficient accuracy.

[0003] First, traditional methods often rely solely on a certain detection technology, such as using PCR or immunological methods alone, which cannot fully capture the multi-dimensional characteristics of pathogens and can easily lead to missed detection or misdiagnosis. Secondly, traditional methods lack targeted strategies for different types of pathogens and use a unified detection process, which is difficult to adapt to the particularities of different pathogens, affecting the sensitivity and specificity of detection. Furthermore, traditional methods are relatively simple in data processing and analysis, and fail to fully utilize the comprehensive advantages of multi-source data, resulting in insufficient information utilization and limited reliability of test results. In addition, traditional methods lack the ability to adjust dynamically and cannot optimize subsequent detection strategies based on preliminary analysis results, resulting in waste of resources and extended detection time.

[0004] Finally, traditional methods lack systematicity and interpretability in pathogen classification and feature importance analysis, making it difficult to provide in-depth pathogen feature information for clinical diagnosis and personalized treatment. These technical problems combined make it difficult for traditional methods to meet the needs of fast, accurate, and efficient detection in complex respiratory pathogen detection scenarios. Summary of the invention

[0005] The purpose of the present invention is to provide a respiratory pathogen detection kit and method to solve the problems raised in the above background technology.

[0006] To achieve the above-mentioned object, one of the objects of the present invention is to provide a respiratory pathogen detection kit, comprising a data collection module, a detection target analysis module and a customized detection module, wherein:

[0007] The data collection module acquires pathogen sample data, and performs multiple tests on the pathogen sample data to obtain PCR data, immune data, and microarray data;

[0008] The detection target analysis module fuses the PCR data, immune data and microarray data to obtain real-time target fusion data; predicts the type of pathogen sample based on the real-time target fusion data using a decision tree algorithm; and obtains feature importance scores of the PCR data, immune data and microarray data in the real-time target fusion data based on the path of the decision tree algorithm;

[0009] The customized detection module performs data fusion strategies of different combinations of PCR data, immune data and microarray data according to different pathogen sample types, and generates detection fusion data according to corresponding feature importance scores; and selects the corresponding pathogen detection method according to the detection fusion data and pathogen sample type.

[0010] As a further improvement of the present technical solution, the data collection module obtains pathogen sample data through an automated sample collection device, wherein the pathogen sample data includes sample ID, pathogen DNA and RNA, pathogen-specific proteins and antigens, and antibodies produced by the human body against the pathogen.

[0011] As a further improvement of the technical solution, the multiple detection in the data collection module specifically includes:

[0012] Use PCR technology to detect DNA and RNA sequences of pathogen sample data and generate PCR data, where PCR stands for polymerase chain reaction, and PCR data includes amplification curve, cycle threshold, melting curve, and gene copy number;

[0013] Use immunoassay technology to detect specific antigens and antibodies in pathogen sample data and generate immune data, which includes reaction intensity, quantitative results, positive / negative judgment and cross-reaction information;

[0014] Gene chips or protein chips are used, and microarray detection technology is used to conduct large-scale, multi-target rapid screening of pathogen sample data to generate microarray data. The microarray data includes fluorescence intensity graphs, hybridization signal intensity, differential expression profiles, and cluster analysis results.

[0015] As a further improvement of the present technical solution, the detection target analysis module includes a target data fusion unit, which fuses PCR data, immune data and microarray data to obtain real-time target fusion data, specifically including:

[0016] The PCR data, immune data and microarray data were standardized using the Z-Score standardization method, which is a standardization method used to convert data of different scales to the same standard.

[0017] The cycle threshold and gene copy number were extracted from PCR data as PCR key features; the antibody titer and antigen concentration were extracted from immune data as immune key features; the expression intensity of differentially expressed genes associated with known pathogens was extracted from microarray data as microarray key features;

[0018] Use variance analysis to evaluate the ability of various key features to distinguish sample categories. According to the F statistic score of variance analysis, select the first N features with higher F values. Use sample ID as index to correspond the data in the three data sources to the same sample. The F value is a statistic in variance analysis, which is used to evaluate the size of the difference between groups relative to the difference within groups. N represents the number of features finally retained. The methods for determining N include fixed ratio method and threshold method.

[0019] By adopting feature-level fusion, the first N selected features are merged into one vector, and principal component analysis is used for dimensionality reduction, retaining 95% of the variance information to obtain real-time target fusion data.

[0020] As a further improvement of the present technical solution, the detection target analysis module includes a sample type prediction unit, which predicts the type of pathogen sample using a decision tree algorithm model based on real-time target fusion data, specifically including:

[0021] Collect historical pathogen sample data from a public database, wherein the historical pathogen sample data includes the pathogen type of the sample, and perform multiple tests and data fusion on the data to generate historical target fusion data;

[0022] The historical target fusion data is divided into a training set and a test set, with 80% used as the training set and 20% used as the test set. Both the training set and the test set contain their unique sample IDs.

[0023] The decision tree algorithm model is constructed using the training set data. The decision tree algorithm will gradually perform feature segmentation based on the key features in the historical target fusion data to form a tree structure. During the training process, the decision tree algorithm will automatically select the best features and segmentation points to maximize the distinction between different pathogen sample types.

[0024] Use the test set data to evaluate the performance of the trained decision tree algorithm model. The evaluation indicators include accuracy, precision, recall and F1 score.

[0025] The real-time target fusion data is input into the trained decision tree algorithm model. The decision tree algorithm model will follow the path of the decision tree, make judgments based on the characteristic values ​​of the samples, and finally reach the leaf node to predict the pathogen type of the real-time target fusion data.

[0026] As a further improvement of the present technical solution, the detection target analysis module includes a feature importance analysis unit, which obtains feature importance scores of PCR data, immune data and microarray data in real-time target fusion data according to the path of the decision tree algorithm model, specifically including:

[0027] For real-time target fusion data, track its complete path in the decision tree algorithm model and record the features used by each decision node on the path;

[0028] Set a counter for each feature with an initial value of 0, and traverse the decision path of the real-time target fusion data. Whenever a feature is used at a segmentation point, its corresponding counter is incremented by 1.

[0029] The features were divided into three categories according to their sources: PCR data, immune data, and microarray data. The total number of times all features were used was calculated, and the number of times each feature was used was divided by the total to obtain the normalized feature importance scores of PCR data, immune data, and microarray data.

[0030] As a further improvement of the technical solution, the customized detection module includes a customized fusion unit, which performs data fusion strategies of different combinations of PCR data, immune data and microarray data according to different pathogen sample types of the predicted real-time target fusion data, and generates detection fusion data according to the corresponding feature importance scores, specifically including:

[0031] If the predicted pathogen type is a viral pathogen, the PCR data and immune data are fused;

[0032] If the predicted pathogen type is a bacterial pathogen, the PCR data and microarray data are fused;

[0033] If the predicted pathogen type is autoimmune disease-related, the immune data and microarray data are fused;

[0034] If the predicted pathogen type is an emerging infectious disease or an unknown pathogen, all data types are fused;

[0035] When performing the corresponding data fusion strategy, the corresponding feature importance score is selected and weighted summed to produce detection fusion data.

[0036] As a further improvement of the present technical solution, the customized detection module includes a detection method determination unit, which selects a corresponding pathogen detection method according to the detection fusion data and the pathogen sample type, specifically including:

[0037] receiving the detection fusion data from the customized fusion unit and the pathogen sample type predicted by the sample type prediction unit;

[0038] According to the predicted pathogen sample type, select the corresponding pathogen detection method, including:

[0039] When the predicted pathogen type is a viral pathogen, viral nucleic acid detection and viral antigen / antibody detection methods are selected;

[0040] When the predicted pathogen type is a bacterial pathogen, bacterial culture, identification, and susceptibility testing methods are selected;

[0041] When the predicted pathogen type is autoimmune disease-related, autoantibody detection and gene expression analysis methods are selected;

[0042] When the predicted pathogen type is an emerging infectious disease or an unknown pathogen, broad-spectrum nucleic acid sequencing and proteomic analysis methods are selected.

[0043] The second object of the present invention is to provide a method for a respiratory pathogen detection kit, comprising the following steps:

[0044] S1, obtaining pathogen sample data, and performing multiple tests on the pathogen sample data to obtain PCR data, immune data, and microarray data;

[0045] S2, fusing the PCR data, immune data and microarray data to obtain real-time target fusion data; using the decision tree algorithm to predict the type of pathogen samples based on the real-time target fusion data; and obtaining the feature importance scores of the PCR data, immune data and microarray data in the real-time target fusion data based on the path of the decision tree algorithm;

[0046] S3. According to the different types of pathogen samples, different combinations of PCR data, immune data and microarray data are used for data fusion strategies, and detection fusion data is generated according to the corresponding feature importance scores; and the corresponding pathogen detection method is selected according to the detection fusion data and the type of pathogen samples.

[0047] Compared with the prior art, the present invention has the following beneficial effects:

[0048] 1. The respiratory pathogen detection kit and method significantly improve the accuracy and efficiency of pathogen detection through a multi-level data fusion strategy; first, PCR, immune and microarray data are fused into real-time target fusion data. This initial fusion provides a comprehensive information basis for subsequent analysis; then, the pathogen sample type is predicted based on the real-time target fusion data. This step utilizes the comprehensive advantages of the fused data and improves the accuracy of the prediction; finally, different data fusion strategies are adopted according to the predicted pathogen type, such as fusion of PCR and immune data for viral pathogens, fusion of PCR and microarray data for bacterial pathogens, etc.; this dynamically adjusted fusion method can optimize data combinations for different types of pathogens, highlight key information, and effectively reduce interference factors, thereby providing more accurate detection results; this multi-level, dynamically adjusted data fusion process not only improves the sensitivity and specificity of detection, but also adapts to the characteristics of different pathogens, providing a reliable basis for the subsequent selection of detection methods.

[0049] 2. The multiple applications of the decision tree algorithm in the respiratory pathogen detection kit and method in this method demonstrate its powerful classification ability and interpretability. First, the decision tree algorithm is used to predict the pathogen sample type of real-time target fusion data. This step can quickly and accurately classify the samples into viral, bacterial, autoimmune-related or emerging unknown pathogens; the hierarchical structure of the decision tree makes the classification process clear and traceable, which helps to understand the classification basis; secondly, by analyzing the path of the decision tree, the importance score of each feature in the PCR, immune and microarray data can be obtained; this process not only reveals which features play a key role in pathogen classification, but also quantifies their importance; this information is crucial for understanding the characteristic manifestations of different types of pathogens, and can guide subsequent detection priorities and method selection; this dual application of the decision tree algorithm not only achieves efficient sample classification, but also provides in-depth feature analysis, greatly enhancing the scientificity and reliability of the entire detection process, and providing strong support for precision medicine and personalized diagnosis.

[0050] 3. The multiple applications of pathogen sample types in the respiratory pathogen detection kit and method in this method fully reflect the concept of personalized and precise detection. First, different data fusion strategies are adopted according to the predicted pathogen sample type, such as fusion of PCR and immune data for viral pathogens, fusion of PCR and microarray data for bacterial pathogens, etc.; this targeted data fusion can highlight the key features of different types of pathogens and improve the accuracy and efficiency of detection; secondly, based on the fused detection data and the predicted pathogen type, the most suitable detection method is selected, such as nucleic acid detection and antigen / antibody detection for viral pathogens, and culture and drug sensitivity testing for bacterial pathogens; the personalized strategy selected by this method can not only improve the specificity and sensitivity of detection, but also save time and resources and avoid unnecessary detection; through this multiple application, this method realizes the personalization of the whole process from data processing to the final detection method selection, which greatly improves the accuracy, efficiency and pertinence of respiratory pathogen detection, and provides strong support for rapid diagnosis and timely treatment. BRIEF DESCRIPTION OF THE DRAWINGS

[0051] Figure 1 It is a schematic diagram of the overall module unit of the present invention;

[0052] Figure 2 It is a schematic diagram of the overall method flow of the present invention.

[0053] In the figure: 100, data collection module; 200, detection target analysis module; 201, target data fusion unit; 202, sample type prediction unit; 203, feature importance analysis unit; 300, customized detection module; 301, customized fusion unit; 302, detection method determination unit. DETAILED DESCRIPTION

[0054] The following will be combined with the accompanying drawings in the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0055] Unless otherwise specified, the terms (including technical terms) used herein have the commonly understood meanings to those skilled in the art. In addition, it is understood that the terms defined in commonly used dictionaries should be understood to have the same meanings as those in the context of the relevant fields, and should not be understood as idealized or overly formal meanings.

[0056] According to the document, the following are some definitions of terms:

[0057] Pathogen sample data: sample information obtained through automated collection devices, including sample ID, pathogen DNA and RNA, unique proteins and antigens, human antibodies, and other relevant biomarkers;

[0058] PCR data: Data obtained by detecting pathogen DNA and RNA sequences using polymerase chain reaction technology, including information such as amplification curve, cycle threshold, melting curve, and gene copy number;

[0059] Immunological data: Data obtained through immunoassay technology, including reaction intensity, quantitative results, positive / negative determination and cross-reaction information, used to detect specific antigens and antibodies;

[0060] Microarray data: data obtained after large-scale multi-target screening using gene chips or protein chips, including fluorescence intensity graphs, hybridization signal intensity, differential expression profiles, and cluster analysis results;

[0061] Real-time target fusion data: comprehensive data obtained by fusing PCR data, immune data and microarray data through standardization and feature selection;

[0062] Pathogen sample type: Pathogen classification predicted based on sample characteristics, such as viral pathogens, bacterial pathogens, autoimmune disease-related or emerging infectious diseases / unknown pathogens;

[0063] Feature importance score: the contribution of each data type (PCR, immune, microarray) to the judgment of pathogen type based on the analysis of the decision tree algorithm model;

[0064] Data fusion strategy: a method to specifically combine and weight different data types (PCR, immune, microarray) according to the predicted pathogen sample type;

[0065] Pathogen detection methods: Specific detection technologies selected based on the predicted pathogen type and fusion data, such as viral nucleic acid detection, bacterial culture identification, autoantibody detection, or broad-spectrum nucleic acid sequencing.

[0066] Next, see Figure 1-Figure 2 One of the purposes of this embodiment is to provide a respiratory pathogen detection kit, including a data collection module 100, a detection target analysis module 200 and a customized detection module 300.

[0067] The data collection module 100 obtains pathogen sample data through an automated sample collection device, such as a manipulator or aerosol collector for nasopharyngeal swabs; wherein the pathogen sample data includes sample ID, DNA and RNA of the pathogen, proteins and antigens specific to the pathogen, antibodies produced by the human body against the pathogen, and other biomarkers that may be related to the pathogen; multiple tests are performed on the pathogen sample data to obtain PCR data, immune data, and microarray data; the multiple tests specifically include:

[0068] Use PCR technology, including variants such as real-time PCR and digital PCR, to detect DNA and RNA sequences in pathogen sample data and generate PCR data, where PCR stands for polymerase chain reaction, and PCR data includes amplification curves (showing the change in fluorescence signal with cycle number during DNA amplification), cycle threshold (cycle number when the fluorescence signal reaches the detection threshold), melting curves (used to verify the specificity of PCR products), and gene copy numbers (especially in digital PCR, the number of target sequences can be accurately determined);

[0069] Use immunoassay techniques, including immunofluorescence, immunochromatography and other methods, to detect specific antigens and antibodies in pathogen sample data and generate immune data, which includes reaction intensity (such as absorbance value or chemiluminescence intensity in ELISA, ELISA stands for enzyme-linked immunosorbent assay, which is a commonly used immunological detection method), quantitative results (such as antibody titer or antigen concentration), positive / negative judgment and cross-reaction information (data related to detection specificity);

[0070] Gene chips or protein chips are used, and microarray detection technology is used to conduct large-scale, multi-target rapid screening of pathogen sample data to generate microarray data. The microarray data includes fluorescence intensity maps (fluorescence intensity at each probe position, where the probe represents a molecule of known sequence that detects specific DNA, RNA or protein), hybridization signal intensity (a numerical value reflecting the degree of binding between the target molecule and the probe), differential expression profiles (comparison of expression differences between different samples), and cluster analysis results (sample or gene grouping information based on expression patterns).

[0071] The target data fusion unit 201 in the detection target analysis module 200 fuses the PCR data, immune data and microarray data to obtain real-time target fusion data, which specifically includes:

[0072] Standardize PCR data, immune data, and microarray data to ensure that data from different sources are comparable. Standardization uses the Z-Score standardization method, which is a standardization method used to convert data of different scales to the same standard.

[0073] The cycle threshold and gene copy number were extracted from PCR data as PCR key features; the antibody titer and antigen concentration were extracted from immune data as immune key features; the expression intensity of differentially expressed genes associated with known pathogens was extracted from microarray data as microarray key features;

[0074] Use variance analysis to evaluate the ability of various key features to distinguish sample categories. According to the F statistic score of variance analysis, select the top N features with higher F values. Use sample ID as index to correspond the data in the three data sources to the same sample. The F value is a statistic in variance analysis, which is used to evaluate the size of the difference between groups relative to the difference within groups. The larger the F value, the more effective the feature is in distinguishing samples of different categories. N represents the number of features that are finally retained. Methods for determining N include fixed ratio method (for example, selecting the top 10% or 20% of features with the highest F value) and threshold method (for example, selecting all features with F values ​​higher than a preset threshold).

[0075] By adopting feature-level fusion, the first N selected features are merged into one vector, and principal component analysis is used for dimensionality reduction, retaining 95% of the variance information to obtain real-time target fusion data.

[0076] The sample type prediction unit 202 in the detection target analysis module 200 predicts the type of pathogen sample using a decision tree algorithm model based on the real-time target fusion data, specifically including:

[0077] Collect historical pathogen sample data from a public database, where the historical pathogen sample data includes the pathogen type of the sample, such as "viral pathogen" or "bacterial pathogen", etc., and perform multiple tests and data fusion on it to generate historical target fusion data;

[0078] The historical target fusion data is divided into a training set and a test set, with 80% used as the training set and 20% used as the test set. Both the training set and the test set contain their unique sample IDs.

[0079] The decision tree algorithm model is constructed using the training set data. The decision tree algorithm will gradually perform feature segmentation based on the key features in the historical target fusion data, such as the PCR key features (cycle threshold, gene copy number), immune key features (antibody titer, antigen concentration) and microarray key features (expression intensity of differentially expressed genes) in the historical target fusion data to form a tree structure. During the training process, the decision tree algorithm will automatically select the best features and segmentation points to maximize the distinction between different pathogen sample types.

[0080] Use the test set data to evaluate the performance of the trained decision tree algorithm model. The evaluation indicators include accuracy, precision, recall, and F1 score.

[0081] The real-time target fusion data is input into the trained decision tree algorithm model. The decision tree algorithm model will follow the path of the decision tree, make judgments based on the characteristic values ​​of the samples, and finally reach the leaf node to predict the pathogen type of the real-time target fusion data.

[0082] The feature importance analysis unit 203 in the detection target analysis module 200 obtains the feature importance scores of the PCR data, immune data and microarray data in the real-time target fusion data according to the path of the decision tree algorithm model, specifically including:

[0083] For real-time target fusion data, track its complete path in the decision tree algorithm model and record the features used by each decision node on the path;

[0084] Set a counter for each feature with an initial value of 0, and traverse the decision path of the real-time target fusion data. Whenever a feature is used at a segmentation point, its corresponding counter is incremented by 1.

[0085] The features were divided into three categories according to their sources: PCR data, immune data, and microarray data. The total number of times all features were used was calculated, and the number of times each feature was used was divided by the total to obtain the normalized feature importance scores of PCR data, immune data, and microarray data.

[0086] The customized fusion unit 301 in the customized detection module 300 performs data fusion strategies in different combinations of PCR data, immune data and microarray data according to different pathogen sample types of the predicted real-time target fusion data, and generates detection fusion data according to the corresponding feature importance scores, specifically including:

[0087] If the predicted pathogen type is a viral pathogen, the PCR data and immune data are fused. In the case of a viral pathogen, the PCR data can detect extremely low concentrations of viral nucleic acid, while the immune data reflects the host's immune response to the virus. The combination of the two can improve the sensitivity of the test and reduce false negative results. In addition, the PCR data reflects the viral load, while the immune data shows the antibody level. The combination of the two can accurately determine the different stages of viral infection (such as acute phase and recovery phase).

[0088] If the predicted pathogen type is a bacterial pathogen, the PCR data and microarray data are fused. In this case, PCR data can quickly identify specific bacterial species, while microarray data can provide a broader spectrum of bacterial gene information. The combination of the two can quickly and accurately identify bacterial species; microarray data can detect drug-resistant genes, and combined with PCR data, it can more comprehensively evaluate bacterial resistance and provide important guidance for clinical drug use;

[0089] If the predicted pathogen type is related to autoimmune disease, the immune data and microarray data are fused. In this case, the immune data can detect specific autoantibodies, while the microarray data can analyze the gene expression profile related to immunity. The combination of the two can comprehensively evaluate the patient's immune status. By combining immune and gene expression data, new biomarkers for autoimmune diseases may be discovered.

[0090] If the predicted pathogen type is an emerging infectious disease or an unknown pathogen, all data types are integrated. In this case, integrating all data types can provide the most comprehensive information and help identify unknown pathogens. Combining multiple data types and detection methods can significantly improve the ability to discover and identify new pathogens.

[0091] When performing the corresponding data fusion strategy, the corresponding feature importance score is selected and weighted summed to produce detection fusion data.

[0092] The detection method determination unit 302 in the customized detection module 300 selects a corresponding pathogen detection method according to the detection fusion data and the pathogen sample type, specifically including:

[0093] Receiving the detection fusion data from the custom fusion unit 301 and the pathogen sample type predicted by the sample type prediction unit 202;

[0094] According to the predicted pathogen sample type, select the corresponding pathogen detection method, including:

[0095] When the predicted pathogen type is a viral pathogen, select viral nucleic acid detection and viral antigen / antibody detection methods; selecting viral nucleic acid detection and viral antigen / antibody detection methods can achieve rapid and accurate viral diagnosis. Nucleic acid detection can detect infection early and quantify viral load, while antigen / antibody detection provides information on the infection stage and immune response. This combination not only improves the sensitivity and specificity of the test, but also can comprehensively evaluate the viral infection status, providing key support for clinical diagnosis and treatment decisions;

[0096] When the predicted pathogen type is a bacterial pathogen, bacterial culture, identification and drug sensitivity test methods are selected; bacterial culture, identification and drug sensitivity test methods can accurately identify bacterial species and evaluate their antibiotic sensitivity. This method can not only confirm the presence of live bacteria, but also directly conduct drug sensitivity tests, providing accurate guidance for clinical drug use. In addition, culture methods are helpful in discovering new or variant bacterial strains, which is of great significance for monitoring bacterial evolution and preventing emerging infections.

[0097] When the predicted pathogen type is related to autoimmune disease, autoantibody detection and gene expression analysis methods are selected; selecting autoantibody detection and gene expression analysis methods can achieve early diagnosis and accurate classification of autoimmune diseases. Autoantibody detection can detect the disease before clinical symptoms appear, while gene expression analysis can reveal the molecular mechanism of the disease. This combination not only helps to formulate individualized treatment plans, but also monitors disease activity and evaluates treatment effects, promoting the progress of autoimmune disease research;

[0098] When the predicted pathogen type is an emerging infectious disease or an unknown pathogen, wide-spectrum nucleic acid sequencing and proteomic analysis methods are selected; wide-spectrum nucleic acid sequencing and proteomic analysis methods can quickly identify unknown pathogens and comprehensively analyze their characteristics. This method can obtain the whole genome sequence of the new pathogen in a short time and evaluate the host's response to the pathogen at the same time. Through this comprehensive analysis, not only can the genetic variation of the pathogen be monitored in real time, but also potential therapeutic targets can be quickly identified, providing key information support for epidemic prevention and control and drug development.

[0099] The second purpose of this embodiment is to provide a method for a respiratory pathogen detection kit, including the following method steps:

[0100] S1, obtaining pathogen sample data, and performing multiple tests on the pathogen sample data to obtain PCR data, immune data, and microarray data;

[0101] S2, fusing the PCR data, immune data and microarray data to obtain real-time target fusion data; using the decision tree algorithm to predict the type of pathogen samples based on the real-time target fusion data; and obtaining the feature importance scores of the PCR data, immune data and microarray data in the real-time target fusion data based on the path of the decision tree algorithm;

[0102] S3. According to the different types of pathogen samples, different combinations of PCR data, immune data and microarray data are used for data fusion strategies, and detection fusion data is generated according to the corresponding feature importance scores; and the corresponding pathogen detection method is selected according to the detection fusion data and the type of pathogen samples.

[0103] The above shows and describes the basic principles, main features and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited by the above embodiments. The above embodiments and descriptions are only preferred examples of the present invention and are not intended to limit the present invention. Without departing from the spirit and scope of the present invention, the present invention may have various changes and improvements, which fall within the scope of the present invention. The scope of protection of the present invention is defined by the attached claims and their equivalents.

Claims

1. A respiratory pathogen detection kit, characterized in that: The method comprises a data collection module (100), a detection target analysis module (200) and a customized detection module (300), wherein: The data collection module (100) acquires pathogen sample data, and performs multiple tests on the pathogen sample data to obtain PCR data, immune data and microarray data; The detection target analysis module (200) performs data fusion on PCR data, immune data and microarray data to obtain real-time target fusion data; predicts the type of pathogen sample using a decision tree algorithm based on the real-time target fusion data; and obtains feature importance scores of PCR data, immune data and microarray data in the real-time target fusion data based on the path of the decision tree algorithm; The customized detection module (300) performs data fusion strategies in different combinations of PCR data, immune data and microarray data according to different pathogen sample types, and generates detection fusion data according to corresponding feature importance scores; and selects a corresponding pathogen detection method according to the detection fusion data and the pathogen sample type.

2. The respiratory pathogen detection kit according to claim 1, characterized in that: The data collection module (100) acquires pathogen sample data through an automated sample collection device, wherein the pathogen sample data includes a sample ID, DNA and RNA of the pathogen, proteins and antigens specific to the pathogen, and antibodies produced by the human body against the pathogen.

3. The respiratory pathogen detection kit according to claim 2, characterized in that: The multiple detection in the data collection module (100) specifically includes: Use PCR technology to detect DNA and RNA sequences of pathogen sample data and generate PCR data, where PCR stands for polymerase chain reaction, and PCR data includes amplification curve, cycle threshold, melting curve, and gene copy number; Use immunoassay technology to detect specific antigens and antibodies in pathogen sample data and generate immune data, which includes reaction intensity, quantitative results, positive / negative judgment and cross-reaction information; Gene chips or protein chips are used, and microarray detection technology is used to conduct large-scale, multi-target rapid screening of pathogen sample data to generate microarray data. The microarray data includes fluorescence intensity graphs, hybridization signal intensity, differential expression profiles, and cluster analysis results.

4. The respiratory pathogen detection kit according to claim 1, characterized in that: The detection target analysis module (200) comprises a target data fusion unit (201), wherein the target data fusion unit (201) fuses PCR data, immune data and microarray data to obtain real-time target fusion data, specifically comprising: The PCR data, immune data and microarray data were standardized using the Z-Score standardization method, which is a standardization method used to convert data of different scales to the same standard. The cycle threshold and gene copy number were extracted from PCR data as PCR key features; the antibody titer and antigen concentration were extracted from immune data as immune key features; the expression intensity of differentially expressed genes associated with known pathogens was extracted from microarray data as microarray key features; Use variance analysis to evaluate the ability of various key features to distinguish sample categories. According to the F statistic score of variance analysis, select the first N features with higher F values. Use sample ID as index to correspond the data in the three data sources to the same sample. The F value is a statistic in variance analysis, which is used to evaluate the size of the difference between groups relative to the difference within groups. N represents the number of features finally retained. The methods for determining N include fixed ratio method and threshold method. By adopting feature-level fusion, the first N selected features are merged into one vector, and principal component analysis is used for dimensionality reduction, retaining 95% of the variance information to obtain real-time target fusion data.

5. The respiratory pathogen detection kit according to claim 4, characterized in that: The detection target analysis module (200) comprises a sample type prediction unit (202), wherein the sample type prediction unit (202) predicts the type of pathogen sample using a decision tree algorithm model according to real-time target fusion data, specifically comprising: Collect historical pathogen sample data from a public database, wherein the historical pathogen sample data includes the pathogen type of the sample, and perform multiple tests and data fusion on the data to generate historical target fusion data; The historical target fusion data is divided into a training set and a test set, with 80% used as the training set and 20% used as the test set. Both the training set and the test set contain their unique sample IDs. The decision tree algorithm model is constructed using the training set data. The decision tree algorithm will gradually perform feature segmentation based on the key features in the historical target fusion data to form a tree structure. During the training process, the decision tree algorithm will automatically select the best features and segmentation points to maximize the distinction between different pathogen sample types. Use the test set data to evaluate the performance of the trained decision tree algorithm model. The evaluation indicators include accuracy, precision, recall and F1 score. The real-time target fusion data is input into the trained decision tree algorithm model. The decision tree algorithm model will follow the path of the decision tree, make judgments based on the characteristic values ​​of the samples, and finally reach the leaf node to predict the pathogen type of the real-time target fusion data.

6. The respiratory pathogen detection kit according to claim 5, characterized in that: The detection target analysis module (200) comprises a feature importance analysis unit (203), wherein the feature importance analysis unit (203) obtains feature importance scores of PCR data, immune data and microarray data in the real-time target fusion data according to the path of the decision tree algorithm model, specifically comprising: For real-time target fusion data, track its complete path in the decision tree algorithm model and record the features used by each decision node on the path; Set a counter for each feature with an initial value of 0, and traverse the decision path of the real-time target fusion data. Whenever a feature is used at a segmentation point, its corresponding counter is incremented by 1. The features were divided into three categories according to their sources: PCR data, immune data, and microarray data. The total number of times all features were used was calculated, and the number of times each feature was used was divided by the total to obtain the normalized feature importance scores of PCR data, immune data, and microarray data.

7. The respiratory pathogen detection kit according to claim 1, characterized in that: The customized detection module (300) includes a customized fusion unit (301), wherein the customized fusion unit (301) performs data fusion strategies in different combinations of PCR data, immune data, and microarray data according to different pathogen sample types of the predicted real-time target fusion data, and generates detection fusion data according to corresponding feature importance scores, specifically including: If the predicted pathogen type is a viral pathogen, the PCR data and immune data are fused; If the predicted pathogen type is a bacterial pathogen, the PCR data and microarray data are fused; If the predicted pathogen type is autoimmune disease-related, the immune data and microarray data are fused; If the predicted pathogen type is an emerging infectious disease or an unknown pathogen, all data types are fused; When performing the corresponding data fusion strategy, the corresponding feature importance score is selected and weighted summed to produce detection fusion data.

8. The respiratory pathogen detection kit according to claim 7, characterized in that: The customized detection module (300) includes a detection method determination unit (302), and the detection method determination unit (302) selects a corresponding pathogen detection method according to the detection fusion data and the pathogen sample type, specifically including: Receiving detection fusion data from a custom fusion unit (301) and a pathogen sample type predicted by a sample type prediction unit (202); According to the predicted pathogen sample type, select the corresponding pathogen detection method, including: When the predicted pathogen type is a viral pathogen, viral nucleic acid detection and viral antigen / antibody detection methods are selected; When the predicted pathogen type is a bacterial pathogen, bacterial culture, identification, and susceptibility testing methods are selected; When the predicted pathogen type is autoimmune disease-related, autoantibody detection and gene expression analysis methods are selected; When the predicted pathogen type is an emerging infectious disease or an unknown pathogen, wide-spectrum nucleic acid sequencing and proteomic analysis methods are selected.

9. A method for using the respiratory pathogen detection kit according to any one of claims 1 to 8, characterized in that: The method comprises the following steps: S1, obtaining pathogen sample data, and performing multiple tests on the pathogen sample data to obtain PCR data, immune data, and microarray data; S2, fusing PCR data, immune data and microarray data to obtain real-time target fusion data; According to the real-time target fusion data, the pathogen sample type is predicted using the decision tree algorithm; and according to the path of the decision tree algorithm, the feature importance scores of the PCR data, immune data and microarray data in the real-time target fusion data are obtained; S3. According to the different types of pathogen samples, different combinations of PCR data, immune data and microarray data are used for data fusion strategies, and detection fusion data is generated according to the corresponding feature importance scores; And select the corresponding pathogen detection method based on the detection fusion data and pathogen sample type.

Images