A data analysis method and system for multiplex PCR multi-target parallel amplification

Abstract

The present invention relates to a data analysis method and system for multiplex PCR multi-target parallel amplification, which relates to the field of biological detection. By collecting user characteristics and respiratory characteristics of target users and analyzing them to obtain feature confidence, these features are used to predict multiplex PCR detection results, and then error calculation is performed with actual detection spectra. In combination with the feature confidence, abnormal patterns are detected and the actual detection spectra are labeled. This solves the technical problems that the analysis of multiplex PCR multi-target parallel amplification data is complex, prone to errors, and difficult to automatically and accurately analyze. It realizes automated and precise analysis of multiplex PCR multi-target parallel amplification data, effectively improves the accuracy and efficiency of data analysis, helps to timely discover detection anomalies, and enhances the reliability and accuracy of clinical research.

Description

Technical Field

[0001] The present invention relates to the field of biological detection, and in particular to a data analysis method and system for multiplex PCR multi-target parallel amplification. Background Art

[0002] In the field of molecular biology, multiplex polymerase chain reaction (MPCR) technology has become an important tool for the simultaneous detection of multiple bacterial, viral, or genetic targets. By incorporating multiple pairs of specific primers into a single PCR reaction, MPCR enables simultaneous amplification of multiple targets, significantly improving detection efficiency and throughput while reducing testing costs and time. In particular, MPCR technology can simultaneously detect multiple respiratory viruses and bacteria in the diagnosis of respiratory infections, providing important support for biological research.

[0003] However, with the widespread application of MPCR technology, the complexity and challenges of its data analysis have become increasingly prominent. The data generated by multiplex PCR reactions contain amplification information of multiple targets. How to accurately and efficiently parse this data and extract useful biological information has become a hot topic and difficulty in current research. Existing data analysis methods often rely on manual interpretation or simple threshold setting, which makes it difficult to cope with complex multiplex PCR data and is prone to misjudgment or omission. In addition, factors such as physiological differences between different users, sample quality, and changes in experimental conditions can also affect the MPCR test results, further increasing the difficulty of data analysis. Summary of the Invention

[0004] The present invention aims to solve the technical problems in the prior art that data analysis of multiplex PCR multi-target parallel amplification is complex, error-prone and difficult to automatically and accurately analyze, and provides a data analysis method and system for multiplex PCR multi-target parallel amplification to solve the problems.

[0005] The technical solution of the present invention to solve the above technical problems is as follows:

[0006] In a first aspect, the present invention provides a data analysis method for multiplex PCR multi-target parallel amplification, the method comprising: collecting user characteristics and respiratory characteristics of a target user, and analyzing to obtain feature confidence; predicting, based on the user characteristics and respiratory characteristics, the detection results after multiplex PCR detection according to multiple preset targets, to obtain a first predicted multiplex PCR detection spectrum and a second predicted multiplex PCR detection spectrum; collecting a test sample of the target user, and performing multiplex PCR parallel amplification detection according to the multiple preset targets to obtain an actual multiplex PCR detection spectrum; calculating the error between the actual multiplex PCR detection spectrum and the first predicted multiplex PCR detection spectrum and the second predicted multiplex PCR detection spectrum to obtain a detection error amplitude, combining the feature confidence, classifying to obtain a detection abnormality pattern, and marking the actual multiplex PCR detection spectrum as a data analysis result.

[0007] In a second aspect, the present invention provides a data analysis system for multiplex PCR multi-target parallel amplification, the system comprising: a feature acquisition module for collecting user features and respiratory features of a target user, and analyzing to obtain feature confidence; a detection spectrum acquisition module for predicting, based on the user features and respiratory features, the detection results after multiplex PCR detection according to multiple preset targets, to obtain a first predicted multiplex PCR detection spectrum and a second predicted multiplex PCR detection spectrum; an amplification detection module for collecting a detection sample of the target user, performing multiplex PCR parallel amplification detection according to the multiple preset targets, and obtaining an actual multiplex PCR detection spectrum; an error calculation module for calculating the error between the actual multiplex PCR detection spectrum and the first predicted multiplex PCR detection spectrum and the second predicted multiplex PCR detection spectrum, to obtain a detection error amplitude, and combining the feature confidence to classify and obtain a detection abnormality pattern, and mark the actual multiplex PCR detection spectrum as a data analysis result.

[0008] The beneficial effects of the present invention are: by collecting user characteristics and respiratory characteristics of the target user and analyzing them to obtain feature confidence, these features are used to predict the multiplex PCR test results, and then error calculation is performed with the actual test spectrum. In combination with the feature confidence, abnormal patterns are detected and the actual test spectrum is labeled, thereby realizing automated and precise analysis of multiplex PCR multi-target parallel amplification data, effectively improving the accuracy and efficiency of data analysis, helping to timely discover detection anomalies, and improving the reliability and accuracy of biological research. BRIEF DESCRIPTION OF THE DRAWINGS

[0009] Figure 1 A schematic flow chart of a data analysis method for multiplex PCR multi-target parallel amplification provided by the present invention.

[0010] Figure 2This is a structural schematic diagram of a data analysis system for multiplex PCR multi-target parallel amplification provided by the present invention.

[0011] Description of the accompanying drawings: feature acquisition module 11, detection spectrum acquisition module 12, amplification detection module 13, error calculation module 14. DETAILED DESCRIPTION

[0012] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. 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 those skilled in the art without making any creative efforts shall fall within the scope of protection of the present invention.

[0013] In the description of the present invention, the terms "first" and "second" are used for descriptive purposes only and should not be understood to indicate or imply relative importance or implicitly specify the number of the technical features indicated. Therefore, a feature specified as "first" or "second" may explicitly or implicitly include one or more of the specified features. In the description of the present invention, "plurality" means two or more, unless otherwise specifically defined.

[0014] In the description of the present invention, the term "for example" is used to mean "used as an example, illustration or illustration". Any embodiment of the present invention described as "for example" is not necessarily to be construed as being more preferred or advantageous than other embodiments. The following description is given to enable any person skilled in the art to implement and use the present invention. In the following description, details are listed for the purpose of explanation. It should be understood that a person of ordinary skill in the art can recognize that the present invention can be implemented without using these specific details. In other examples, well-known structures and processes are not elaborated in detail to avoid obscuring the description of the present invention with unnecessary details. Therefore, the present invention is not intended to be limited to the embodiments shown, but is consistent with the widest scope consistent with the principles and features disclosed herein.

[0015] Example 1, as Figure 1 As shown, an embodiment of the present invention provides a data analysis method for multiplex PCR multi-target parallel amplification, the method comprising: S10: collecting user characteristics and respiratory characteristics of a target user, and analyzing to obtain feature confidence.

[0016] For example, multiplex PCR is a variant of the polymerase chain reaction (PCR) that allows for the simultaneous amplification of multiple different DNA targets in a single reaction. Compared to traditional single-target PCR, multiplex PCR achieves simultaneous amplification of multiple targets in a single reaction by designing multiple pairs of specific primers that target different DNA sequences.

[0017] Among them, specific primers are specially designed primers that target different DNA sequences. Each pair of primers can accurately identify and bind to a specific target DNA fragment, guiding the DNA amplification reaction. This ensures that in a single multiplex PCR reaction, different primers each amplify the corresponding target DNA sequence, achieving simultaneous amplification of multiple targets without non-specific binding to other unrelated DNA sequences. For example, in a multiplex PCR experiment to detect multiple viruses, primers designed for the specific gene fragments of each virus are specific primers. They only work on the genes of the corresponding virus and will not mistakenly bind to the DNA of other viruses or normal cells.

[0018] Multiplex PCR technology is widely used in molecular biology research, particularly when simultaneous detection of multiple bacterial, viral, or genetic targets is required. For example, in respiratory infection research, multiplex PCR can simultaneously detect multiple respiratory viruses (such as influenza virus and respiratory syncytial virus) and bacteria (such as Streptococcus pneumoniae and Haemophilus influenzae).

[0019] Multi-target parallel amplification is achieved by adding multiple pairs of specific primers to the PCR reaction system. These primers target different DNA targets and are designed to prevent mutual interference, i.e., the formation of primer dimers or nonspecific amplification. During the PCR reaction, all primers bind simultaneously to the template DNA and are amplified by the DNA polymerase. Because each primer targets a specific DNA sequence, the amplified product is also specific, thus achieving multi-target parallel amplification.

[0020] The purpose of amplification is to increase the amount of a specific DNA sequence to a detectable level. In molecular biology research, the target DNA sequence content in many biological samples is extremely low, and direct detection often fails to obtain accurate results. Through PCR amplification, the amount of target DNA sequence can be amplified millions or even billions of times, greatly improving the sensitivity and accuracy of detection. In multiplex PCR, amplification requires processing multiple targets simultaneously, and the content of each target may be very low. Only through amplification can each target be accurately detected. The amplified DNA products can also be used for subsequent analysis, such as sequencing, cloning, expression, etc., providing rich materials for scientific research.

[0021] Preferably, in the data analysis method for multiplex PCR multi-target parallel amplification proposed in this scheme, the user characteristics and respiratory characteristics of the target user are first collected. The target user refers to the specific research object for multiplex PCR multi-target parallel amplification. After the target user is determined, the user characteristics are obtained. The user characteristics refer to the basic information of the target user, such as age, weight, etc. These characteristics are crucial for understanding the user's physiological state and possible influencing factors. At the same time, the respiratory characteristics of the target user are obtained. The respiratory characteristics refer to characteristics related to the user's respiratory health status, which may include symptom descriptions, such as cough, fever, etc.

[0022] After obtaining the target user's user and respiratory characteristics, the reliability of these features for the current detection, or confidence level, is calculated by analyzing their occurrence rates in the user database. A higher confidence level indicates a greater contribution of these features to the prediction result. The user database continuously updates and stores feature data for a large number of users, allowing for the calculation of feature occurrence rates.

[0023] Specifically, user characteristics, such as the frequency of occurrence of a specific age group or weight range, and respiratory characteristics, such as the frequency of occurrence of specific sample physical properties, are counted separately. The first occurrence rate and second occurrence rate of the two in the database are obtained in turn, and then the feature confidence is calculated based on these occurrence rates. Feature confidence, as an indicator to measure the reliability of a feature, reflects the stability and representativeness of the feature in the current analysis context, and provides an important reference for the subsequent prediction of multiplex PCR test results. For example, if users of a certain age group appear frequently in the database and their respiratory characteristics show a certain regularity, the feature confidence of this age group and related respiratory characteristics will be higher, which means that these features have a higher credibility when predicting the results of multiplex PCR tests.

[0024] S20: Based on the user characteristics and respiratory characteristics, the test results after the multiplex PCR test according to the multiple preset targets are predicted respectively to obtain a first predicted multiplex PCR test spectrum and a second predicted multiplex PCR test spectrum.

[0025] Optionally, after collecting the user characteristics and respiratory characteristics of the target user and analyzing them to obtain the feature confidence, these feature information are used to predict the test results after performing multiple PCR tests according to multiple preset targets.

[0026] Multiple preset targets refer to pre-defined targets for multiplex PCR testing. These targets are the objects of detection. The prediction network will predict the results of multiplex PCR testing based on these preset targets based on user characteristics and respiratory characteristics. For example, in actual testing, multiple preset targets may be set for different viral markers. The prediction network will predict the amplification of these targets, etc. If the physical properties of the sample indicate the presence of certain components at high concentrations, predictions may be made for targets related to these components.

[0027] Specifically, the multiplex PCR prediction channel configured on the cloud server is called. This channel has a built-in first and second multiplex PCR prediction networks, both of which are built and trained using deep learning technology. During the prediction process, collected user characteristics, such as age, weight, and other indicators reflecting individual physiological status, as well as respiratory characteristics, such as the physical properties of the sample and parameters related to environmental conditions, which may affect the PCR reaction, are used as input data and transmitted to the first and second multiplex PCR prediction networks respectively.

[0028] The first and second multiplex PCR prediction networks predict the results of multiplex PCR tests based on the input feature information and the complex patterns and associations they have learned internally. The prediction outputs two predictions: the first predicted multiplex PCR test profile and the second predicted multiplex PCR test profile. These profiles represent the predictions of the multiplex PCR test results from different prediction networks, respectively. They contain key information such as the amplification status and signal strength of each pre-set target.

[0029] For example, if user characteristics indicate that the user is in a certain age group, and the physical properties of the sample in the respiratory characteristics indicate that it may contain high concentrations of certain components, then the prediction network may predict based on these characteristics that targets associated with these components will show strong amplification signals in the multiplex PCR test, which will be reflected in the predicted multiplex PCR test profile. These two predicted profiles provide a benchmark for subsequent comparative analysis with actual test results, helping to evaluate the accuracy and reliability of the test.

[0030] S30: Collecting a test sample from the target user, performing multiplex PCR parallel amplification detection according to the multiple preset targets, and obtaining an actual multiplex PCR detection spectrum.

[0031] Furthermore, after completing the collection and predictive analysis of the target user characteristics and respiratory characteristics, the experimental operation stage is entered, that is, the test samples of the target user are collected, and multiple PCR parallel amplification detection is performed based on the preset multiple targets to obtain the actual multiple PCR detection spectrum.

[0032] Specifically, as the material basis for subsequent analysis, test samples must be collected according to strict operating procedures to ensure aseptic operation during the collection process and the integrity and representativeness of the sample. Sample types may include biological fluids and tissue fragments, and their selection depends on the experimental purpose and the characteristics of the intended target. For example, when collecting biological fluid samples, dedicated collection instruments and standard operating procedures must be followed to avoid contamination.

[0033] Furthermore, based on the preset multiple target information, which are usually pre-selected based on research needs and may cover specific gene fragments, markers and other DNA sequences with analytical value, multiplex PCR parallel amplification detection is performed.

[0034] Multiplex PCR technology is used to simultaneously add specific primers for multiple preset targets into the same reaction system. These primers are designed with precise base sequences to ensure that only the target sequence is specifically amplified in the PCR reaction. At the same time, the system also contains necessary components such as DNA polymerase and dNTPs to support the extension of the DNA chain. The purpose is to ensure that there is no interference during the amplification process and that the corresponding target sequences can be specifically identified and amplified.

[0035] Under PCR reaction conditions, the target DNA sequence is exponentially amplified through a thermal cycling process. As a key device in executing PCR reactions, a thermal cycler precisely controls temperature cycling, including the three key steps of denaturation, annealing, and extension, to achieve exponential amplification of the target DNA sequence.

[0036] After amplification is completed, high-precision detection equipment such as fluorescent quantitative PCR instrument or capillary electrophoresis instrument is used to detect and analyze the amplified products. These devices can accurately quantify the amplification of each target based on characteristic parameters such as fluorescence signal intensity or electrophoretic mobility, and integrate the detection data of each target to construct the actual multiplex PCR detection spectrum.

[0037] The actual multiplex PCR detection spectrum records the amplification status of each preset target in detail, including key parameters such as amplification efficiency and signal intensity, and is presented in graphical or digital form, intuitively demonstrating the amplification efficiency and relative content of each target, providing an experimental basis for subsequent comparative analysis with predicted detection spectra. For example, if the preset target contains a DNA sequence of a specific length, after parallel amplification through multiplex PCR, if the sequence is successfully amplified, a corresponding high-intensity fluorescent signal peak will appear in the actual multiplex PCR detection spectrum. Its position and intensity reflect the amplification efficiency and content of the target, providing intuitive and quantitative results for data analysis.

[0038] S40: Calculate the error between the actual multiplex PCR detection spectrum and the first predicted multiplex PCR detection spectrum and the second predicted multiplex PCR detection spectrum to obtain a detection error amplitude, combine the feature confidence, classify and obtain a detection abnormality pattern, and annotate the actual multiplex PCR detection spectrum as a data analysis result.

[0039] Specifically, after obtaining the actual multiplex PCR test profile, the error between it and the first and second predicted multiplex PCR test profiles must be calculated to quantify the degree of deviation from the test results. This error calculation compares the amplification signal intensity or relative content of each target in the actual test profile with the estimated value of the corresponding target in the predicted profile. Statistical metrics such as mean squared error (MSE) and absolute error are used for quantitative evaluation. The result is the test error margin, which reflects the difference between the actual test result and the predicted value.

[0040] Then, the calculated detection error amplitude is combined with the feature confidence obtained through analysis for analysis. Feature confidence is an indicator that measures the reliability of user features and respiratory features, and its level directly affects the accuracy of the prediction results. Through preset classification rules or algorithm models, such as decision trees and support vector machines, the detection abnormality pattern is classified and identified based on the comprehensive consideration of the detection error amplitude and feature confidence. For example, when the detection error amplitude exceeds a certain threshold and the feature confidence is low, it may be classified as a "high error and low confidence abnormal pattern", indicating that there is a significant difference between the detection result and the predicted value, and the reliability of the feature information is insufficient; on the contrary, if the error amplitude is small and the confidence is high, it is classified as a "normal detection pattern".

[0041] Ultimately, the actual multiplex PCR test profiles are annotated based on the detected anomaly patterns obtained through classification. This annotation may include key information such as the type and degree of anomaly, thereby forming a complete data analysis result. This result not only provides an important basis for subsequent data interpretation but also provides direction for optimizing PCR reaction conditions and improving detection accuracy. For example, if the annotation results indicate a "high error, low confidence anomaly" for a target, further review of the primer design, reaction conditions, or sample processing of that target is necessary to identify and correct potential sources of error.

[0042] In summary, the multiplex PCR multi-target parallel amplification data analysis process collects user characteristics and respiratory characteristics, predicts multiplex PCR test results, obtains actual test spectra, calculates errors, and classifies abnormal detection patterns, ultimately achieving accurate evaluation and annotation of multiplex PCR test results. This significantly improves the accuracy and reliability of the test and provides technical support for subsequent data interpretation and experimental optimization.

[0043] In a preferred embodiment, user characteristics and respiratory characteristics of the target user are collected and analyzed to obtain feature confidence, including: collecting user characteristics of the target user, wherein the user characteristics include age and weight; collecting respiratory characteristics of the target user, wherein the respiratory characteristics include at least sample physical properties, environmental parameters and respiratory symptom indicators; and analyzing to obtain feature confidence of the user characteristics and respiratory characteristics.

[0044] Specifically, in the preparation stage of multiplex PCR multi-target parallel amplification detection, it is necessary to systematically collect the user characteristics and respiratory characteristics of the target user, and obtain the feature confidence based on these feature analysis to ensure the accuracy and reliability of subsequent detection and analysis.

[0045] The collection of user characteristics focuses on two basic physiological indicators: age and weight. Age, as a continuous variable, reflects the user's life cycle stage and may affect the natural content or metabolic rate of certain components in their body. Weight, as an indicator of the user's body size and substance content, is indirectly related to the amount of sample collected or the distribution of substances in the body. For example, under the same conditions, a heavier user may produce more biological samples for testing.

[0046] Respiratory profile collection focuses on characteristics directly related to the respiratory system. These include environmental parameters at the time of sample collection, such as temperature and humidity. While these parameters are not directly considered as user characteristics, they may affect the physical state of the sample. These include the temperature, humidity, and air quality (dust concentration, microbial content) of the sampling environment, which can introduce external interference or affect sample preservation. Furthermore, the sample's physical and chemical properties, such as viscosity, transparency, color, pH, osmotic pressure, and protein concentration, reflect its appearance and chemical properties and may affect PCR reaction efficiency. Respiratory symptom indicators, such as cough frequency, sputum quality, wheezing intensity, temperature, respiratory rate, and abnormal lung auscultation, assist in determining infection type or inflammatory status. Furthermore, depending on the specific situation, respiratory profiles may also include microbial indicators, such as Gram stain results and direct microbial flora distribution, to guide target selection for multiplex PCR. Medical history data, such as previous respiratory infection types and records of previously positive targets, can assist in validating the validity of current test results.

[0047] After feature collection is completed, the frequency of occurrence of these features in the preset user database is counted to analyze and obtain the feature confidence of user features and respiratory features. Feature confidence is a quantitative indicator to measure the stability and representativeness of features, and its level directly reflects the reliability of the features in the current analysis context. For example, if users of a certain age group and weight range appear frequently in the database, and their respiratory features show a certain regularity, the feature confidence of this combined feature will be higher, which means that these features have higher credibility in the subsequent multiplex PCR multi-target parallel amplification detection analysis, providing strong support for the accurate interpretation of the test results.

[0048] In a preferred embodiment, the method further includes: retrieving the user features in a user database within a recent preset time range to obtain a first occurrence rate, wherein the user database updates and stores the user features and respiratory features within the recent preset time range; retrieving the respiratory features in the user database to obtain a second occurrence rate; and calculating the feature confidence based on the first occurrence rate and the second occurrence rate.

[0049] Specifically, in order to accurately assess the reliability of the target user's characteristics and respiratory characteristics, targeted retrieval and calculation are required within the user database within the most recent preset time range. The user database continuously updates and stores the characteristic information of all users within the most recent preset time period, covering user characteristics (such as age, weight, and other indicators reflecting the individual's physiological state) and respiratory characteristics (such as the sample's physicochemical properties, environmental parameters, and other factors that may affect the PCR reaction). The most recent preset time range refers to a specific pre-set time interval. This preset time range can be set to the last week, month, or three months, etc. The data within this range is used for subsequent targeted retrieval and calculation to assess the reliability of the target user's characteristics and respiratory characteristics.

[0050] The database is searched for the target user's user features. The first occurrence rate is calculated by counting the number of occurrences of these features in the database and dividing the result by the total number of records in the database. The first occurrence rate is a metric used to measure the prevalence and stability of user features, and its high or low level directly reflects the distribution of the feature within the target user population. For example, if users of a certain age group appear frequently in the database, the first occurrence rate of the feature for that age group will be high, indicating that the feature is highly prevalent within the current user population.

[0051] The target user's respiratory characteristics are retrieved in the same database. Similarly, the second occurrence rate is obtained by counting the number of occurrences and calculating the ratio. The second occurrence rate focuses on reflecting the distribution pattern of respiratory characteristics in the target user group.

[0052] Based on the first occurrence rate and the second occurrence rate, the feature confidence is calculated through preset algorithm models, such as weighted average, multiplication model, etc. Feature confidence is a quantitative indicator that comprehensively measures the reliability of user features and respiratory features. Its calculation process fully considers the frequency of occurrence of the two features and their potential impact in data analysis. For example, if a user's age characteristics and respiratory characteristics both have a high occurrence rate in the database, their feature confidence will be increased accordingly, which means that these features have higher credibility in subsequent multiplex PCR multi-target parallel amplification detection and analysis, providing a strong basis for the accurate interpretation of the test results.

[0053] In a preferred embodiment, based on the user characteristics and respiratory characteristics, the test results after multiple PCR detection according to multiple preset targets are predicted respectively to obtain a first predicted multiple PCR detection spectrum and a second predicted multiple PCR detection spectrum, including: calling a multiple PCR prediction channel configured on a cloud server, wherein the multiple PCR prediction channel includes a first multiple PCR prediction network and a second multiple PCR prediction network; transmitting the user characteristics and respiratory characteristics to the cloud server, respectively inputting the first multiple PCR prediction network and the second multiple PCR prediction network in the multiple PCR prediction channel, and predicting the output to obtain the first predicted multiple PCR detection spectrum and the second predicted multiple PCR detection spectrum for multiple PCR detection according to multiple preset targets.

[0054] Preferably, to accurately predict test results, a multiplex PCR prediction channel configured on a cloud server is invoked. This channel, as the core analysis module, integrates a first multiplex PCR prediction network and a second multiplex PCR prediction network. Both networks are built based on deep learning algorithms and trained on extensive data, enabling them to predict multiplex PCR test results.

[0055] Specifically, the collected target user characteristics (e.g., age, weight, and other indicators reflecting individual physiological status) and respiratory characteristics (e.g., physical and chemical properties of the sample, environmental parameters, and other factors that may affect the PCR reaction) are used as input data and transmitted to a cloud server via a secure data transmission protocol. These characteristic data are then fed into the first and second multiplex PCR prediction networks within the multiplex PCR prediction channel.

[0056] Within the prediction network, the input feature data is deeply analyzed and processed through a complex neural network structure and algorithm model to simulate various variables and influencing factors in the multiplex PCR detection process, and then predict and output the test results after multiplex PCR detection according to multiple preset targets.

[0057] Among them, the first multiplex PCR prediction network and the second multiplex PCR prediction network may output slightly different prediction results due to differences in network structure, training data or algorithm optimization, thereby obtaining the first predicted multiplex PCR detection spectrum and the second predicted multiplex PCR detection spectrum respectively. These two predicted detection spectra record in detail the predicted amplification of each preset target, including key parameters such as amplification efficiency and signal intensity, providing an important basis for subsequent comparative analysis with actual detection results. For example, if a target shows a higher amplification efficiency in the first predicted multiplex PCR detection spectrum, but a slightly lower amplification efficiency in the second predicted multiplex PCR detection spectrum, this may reflect the differentiated understanding of the amplification of the target by different prediction networks, providing multi-angle reference information for subsequent data interpretation.

[0058] In a preferred embodiment, the configuration step of the multiplex PCR prediction channel includes: based on sample data of multiplex PCR detection according to multiple preset targets, collecting a sample user feature set and a sample respiratory feature set of multiple sample users, and collecting a sample multiplex PCR detection spectrum set for multiplex PCR detection; using deep learning to construct a first multiplex PCR prediction network and a second multiplex PCR prediction network; using the sample user feature set and the sample respiratory feature set as input training data, and using the sample multiplex PCR detection spectrum set as output supervision data, respectively, to supervise the training, verification and testing of the first multiplex PCR prediction network and the second multiplex PCR prediction network until convergence; configuring the converged first multiplex PCR prediction network and the second multiplex PCR prediction network on a cloud server to obtain a multiplex PCR prediction channel.

[0059] Furthermore, the configuration of the multiplex PCR prediction channel aims to build a model that can accurately predict multiplex PCR test results through deep learning technology. The specific configuration process requires systematically collecting sample user feature sets and sample respiratory feature sets from multiple sample users based on sample data from multiplex PCR tests based on multiple preset targets.

[0060] The sample user feature set includes indicators such as age and weight that reflect individual physiological status, while the sample respiratory feature set includes factors such as the sample's physical and chemical properties and environmental parameters that may affect the PCR reaction. Furthermore, a set of multiplex PCR profiles is collected after multiplex PCR testing. This set details the actual amplification status of each pre-set target and serves as a supervisory basis for subsequent model training.

[0061] Subsequently, deep learning techniques were used to construct the first and second multiplex PCR prediction networks, respectively. These two networks may differ in structure, such as the number of layers, number of neurons, or choice of activation function, aiming to capture complex patterns in the data through different model architectures. These are not specified here and can be configured based on practical needs.

[0062] During the model training phase, supervised training of the first and second multiplex PCR prediction networks was performed using the sample user feature set and the sample respiratory feature set as input training data, respectively, and the sample multiplex PCR test profile set as output supervision data. During training, network parameters were continuously adjusted to gradually approximate the model output to the actual multiplex PCR test profile. Simultaneously, the model was validated using a validation set to evaluate its generalization ability to new data, and a final test was performed using a test set to ensure model stability and reliability.

[0063] When the model reaches a convergence state during the training, validation, and testing processes, it indicates that the model has learned the effective patterns in the data and can accurately predict the results of multiplex PCR tests. Finally, the converged first multiplex PCR prediction network and the second multiplex PCR prediction network are deployed to the cloud server to construct a multiplex PCR prediction channel. The multiplex PCR prediction channel can receive user characteristics and respiratory characteristics as input, and quickly output the predicted multiplex PCR detection spectrum, providing strong support for subsequent data analysis. For example, if the age and weight characteristics of a sample user are input into the multiplex PCR prediction channel, the channel will use the trained model to predict the amplification of each target in the multiplex PCR test of the user sample, and output the corresponding predicted detection spectrum.

[0064] In a preferred embodiment, calculating the error between the actual multiplex PCR detection spectrum and the first predicted multiplex PCR detection spectrum and the second predicted multiplex PCR detection spectrum to obtain the detection error amplitude includes: calculating the absolute error between the actual multiplex PCR detection spectrum and the first predicted multiplex PCR detection spectrum and the second predicted multiplex PCR detection spectrum to obtain the first detection spectrum error and the second detection spectrum error; and calculating the detection error amplitude based on the first detection spectrum error and the second detection spectrum error.

[0065] For example, to evaluate the accuracy of the prediction results, it is necessary to calculate the error between the actual multiplex PCR detection spectrum and the first predicted multiplex PCR detection spectrum and the second predicted multiplex PCR detection spectrum, thereby obtaining the detection error amplitude.

[0066] First, for each preset target, the absolute error between the actual multiplex PCR detection spectrum and the first predicted multiplex PCR detection spectrum and the second predicted multiplex PCR detection spectrum is calculated respectively. The absolute error is a quantitative indicator to measure the difference between the predicted value and the actual value. It is obtained by calculating the absolute value of the difference between the amplification signal intensity or relative content of each target in the predicted spectrum and the corresponding target value in the actual spectrum. For example, if the amplification signal intensity of a target in the actual multiplex PCR detection spectrum is 100 units, and the predicted value in the first predicted multiplex PCR detection spectrum is 90 units, then the error of the first detection spectrum of the target is 10 units. Similarly, the error of the second detection spectrum of the target in the second predicted multiplex PCR detection spectrum can be calculated.

[0067] Subsequently, based on the obtained first detection spectrum error and second detection spectrum error, the detection error amplitude is further calculated. The detection error amplitude, as an indicator that comprehensively reflects the accuracy of the prediction, can be obtained by calculating the average value, maximum value or other statistical quantities of the first detection spectrum error and the second detection spectrum error. The specific calculation method depends on the analysis requirements. For example, if the average value is used as the calculation method for the detection error amplitude, the first detection spectrum error and the second detection spectrum error of all targets are summed and averaged to obtain two average error values. These two average error values together constitute a quantitative representation of the detection error amplitude, which provides an important basis for subsequent data interpretation and model optimization. Through this process, the accuracy of the multiplex PCR prediction model can be systematically evaluated, providing strong support for improving the prediction algorithm and improving the detection accuracy.

[0068] In a preferred embodiment, in combination with the feature confidence, the detection anomaly pattern is classified and obtained, and the actual multiplex PCR detection spectrum is annotated as a data analysis result, including: when the detection error amplitude is greater than or equal to a preset error amplitude threshold, the detection error amplitude and the feature confidence are input into the detection anomaly pattern classification channel configured on the cloud server, and multiple classification detection anomaly patterns are obtained through classification output of multiple detection anomaly pattern classification branches; the classification detection anomaly pattern with the highest occurrence frequency among the multiple classification detection anomaly patterns is screened to obtain a detection anomaly pattern; and the detection anomaly pattern is used to annotate the actual multiplex PCR detection spectrum as a data analysis result.

[0069] Optionally, in order to accurately identify and mark abnormal situations during the detection process, a detection abnormality pattern classification channel is also configured on the cloud server, and then a comprehensive analysis of the detection error amplitude is performed in combination with the feature confidence.

[0070] Specifically, when the calculated detection error margin is greater than or equal to a preset error margin threshold, it indicates a significant difference between the actual multiplex PCR detection profile and the predicted profile. The detection error margin and feature confidence level are then used as input data and transmitted to the detection anomaly pattern classification channel configured on the cloud server for further analysis. The error margin threshold can be preset based on actual conditions.

[0071] The anomaly detection pattern classification channel integrates multiple anomaly detection pattern classification branches. Each branch is built based on a specific algorithm model or rule set, and is designed to classify and analyze input data from different perspectives. By processing multiple anomaly detection pattern classification branches in parallel, the classification channel can output multiple classified anomaly detection patterns, which reflect different anomaly types that may exist during the detection process.

[0072] The output of multiple classification anomaly detection patterns is then screened, and the frequency of each pattern in all classification results is counted. The classification anomaly detection pattern with the highest frequency is selected as the final anomaly detection pattern. This screening process, based on statistical principles, aims to identify the most likely anomaly pattern from multiple possible anomaly patterns, thereby improving the accuracy of anomaly identification.

[0073] Finally, the identified abnormality patterns are used to annotate the actual multiplex PCR profiles. This annotation may include key information such as the abnormality type. This annotation, the final output of data analysis, not only provides an important basis for subsequent data interpretation but also provides guidance for optimizing PCR reaction conditions and improving the testing process.

[0074] In a preferred embodiment, the configuration steps of the detection abnormal pattern classification channel include: based on the detection records of multiple PCR detection, collecting the sample feature confidence set and the sample detection error amplitude set, and collecting the abnormal types detected under different combinations of sample feature confidence and sample detection error amplitude, and marking to obtain the sample abnormal pattern set; performing N-fold division on the sample feature confidence set, sample detection error amplitude set and sample abnormal pattern set to obtain N sets of abnormal pattern classification training data, where N is a positive integer; using N sets of abnormal pattern classification training data respectively, constructing N detection abnormal pattern classification branches, obtaining the detection abnormal pattern classification channel, and configuring it on the cloud server.

[0075] Specifically, in the process of constructing a multiplex PCR detection abnormal pattern classification channel, it is necessary to systematically collect the sample feature confidence set, the sample detection error amplitude set, and the abnormality types detected under different combinations of sample feature confidence and sample detection error amplitude based on the detection records of the multiplex PCR test. Among them, the sample feature confidence set reflects the reliability of the feature in the prediction, the sample detection error amplitude set quantifies the deviation between the actual detection and the prediction, and the abnormality type, such as false negative and false positive, is the classification of erroneous results that may occur during the detection process. By collecting and labeling these data, a sample abnormality pattern set is formed. This set records in detail the abnormality types corresponding to various combinations of feature confidence and detection error amplitude, providing a rich data foundation for subsequent model training.

[0076] Subsequently, in order to improve the generalization ability and robustness of the model, the sample feature confidence set, the sample detection error amplitude set and the sample abnormal pattern set are divided into N folds, that is, they are divided into N non-overlapping subsets. Each time, N-1 parts are selected as training data, and the remaining 1 part is used as verification data. The cycle is repeated N times to ensure that each piece of data is used for both training and verification, thereby obtaining N pieces of abnormal pattern classification training data.

[0077] Next, N branches for detecting anomaly patterns are independently constructed using these N pieces of training data. Each branch is based on a specific algorithmic model (such as a decision tree, support vector machine, or neural network), aiming to learn the complex mapping relationship between feature confidence, detection error margin, and anomaly type from different perspectives. By training these N branches in parallel, the model can capture the diversity and complexity of the data, improving the accuracy of anomaly identification.

[0078] Finally, these N trained anomaly pattern classification branches are integrated into a complete anomaly pattern classification channel and deployed to a cloud server. This channel receives the feature confidence and detection error margin from actual testing as input. Through the collaborative work of internal branches, it quickly outputs the corresponding anomaly type, providing strong support for quality control and result interpretation of multiplex PCR testing. For example, if the feature confidence of a test is low and the detection error margin is large, the classification channel may output anomaly types such as "false negative" or "false positive", prompting the tester to further verify the test process or sample processing steps.

[0079] The data analysis method for multiplex PCR multi-target parallel amplification provided by the embodiment of the present invention has at least the following technical effects:

[0080] 1. By calling the multiplex PCR prediction channel configured on the cloud server and using the first and second multiplex PCR prediction networks constructed using deep learning technology, intelligent prediction of multiplex PCR test results is achieved. This prediction channel can comprehensively consider the user characteristics and respiratory characteristics of the target user and accurately output the first predicted multiplex PCR test spectrum and the second predicted multiplex PCR test spectrum. This intelligent prediction not only improves detection efficiency, but also enhances the reliability and accuracy of the prediction results through the setting of the dual prediction network, providing a basis for subsequent detection error assessment.

[0081] 2. After obtaining the actual multiplex PCR detection spectrum, the detection error amplitude was further obtained by calculating the absolute error between it and the predicted spectrum. Combined with the feature confidence, an innovative detection anomaly pattern classification channel was introduced. This channel can automatically classify and output multiple classified detection anomaly patterns based on a comprehensive analysis of the detection error amplitude and feature confidence. By screening the most frequently occurring classified detection anomaly patterns, it can accurately identify abnormal situations during the detection process, such as false negatives and false positives, providing strong support for the accurate interpretation of the test results.

[0082] 3. The computing power of the cloud server is fully utilized, and both the multiplex PCR prediction channel and the detection abnormal pattern classification channel are deployed on the cloud server. This deployment method not only improves the speed and efficiency of data processing, but also realizes centralized data management and sharing. At the same time, by partitioning the sample data into N-folds to construct multiple detection abnormal pattern classification branches, the generalization ability and robustness of the model are further enhanced, ensuring that the detection abnormal pattern can be accurately identified under different combinations of sample feature confidence and detection error amplitude. This efficient data processing and cloud server deployment strategy provides a strong guarantee for the large-scale application of multiplex PCR multi-target parallel amplification detection.

[0083] Example 2, as Figure 2 As shown, based on the same inventive concept as the data analysis method for multiplex PCR multi-target parallel amplification provided in Example 1, an embodiment of the present invention also provides a data analysis system for multiplex PCR multi-target parallel amplification, the system comprising: a feature acquisition module 11, for collecting user features and respiratory features of the target user, and analyzing to obtain feature confidence.

[0084] The detection spectrum acquisition module 12 is used to predict the detection results of multiplex PCR detection according to multiple preset targets based on the user characteristics and respiratory characteristics, and obtain a first predicted multiplex PCR detection spectrum and a second predicted multiplex PCR detection spectrum.

[0085] The amplification detection module 13 is used to collect the test sample of the target user, perform multiplex PCR parallel amplification detection according to the multiple preset targets, and obtain an actual multiplex PCR detection spectrum.

[0086] The error calculation module 14 is used to calculate the error between the actual multiplex PCR detection spectrum and the first predicted multiplex PCR detection spectrum and the second predicted multiplex PCR detection spectrum to obtain the detection error amplitude, combine the feature confidence, classify the detection abnormality pattern, and annotate the actual multiplex PCR detection spectrum as the data analysis result.

[0087] Furthermore, the feature collection module 11 is also used to perform the following steps: collecting user features of the target user, wherein the user features include age and weight; collecting respiratory features of the target user; and analyzing and obtaining feature confidence of the user features and respiratory features.

[0088] Furthermore, the feature acquisition module 11 is also used to perform the following steps: retrieving the user features in the user database within the most recent preset time range to obtain a first occurrence rate, wherein the user database updates and stores the user features and respiratory features within the most recent preset time range, and the respiratory features include at least sample physical properties, environmental parameters, and respiratory symptom indicators; retrieving the respiratory features in the user database to obtain a second occurrence rate; and calculating the feature confidence based on the first occurrence rate and the second occurrence rate.

[0089] Furthermore, the detection spectrum acquisition module 12 is also used to perform the following steps: calling the multiplex PCR prediction channel configured on the cloud server, wherein the multiplex PCR prediction channel includes a first multiplex PCR prediction network and a second multiplex PCR prediction network; transmitting the user characteristics and respiratory characteristics to the cloud server, and inputting the first multiplex PCR prediction network and the second multiplex PCR prediction network in the multiplex PCR prediction channel respectively, and predicting and outputting a first predicted multiplex PCR detection spectrum and a second predicted multiplex PCR detection spectrum for multiplex PCR detection according to multiple preset targets.

[0090] Furthermore, the detection spectrum acquisition module 12 is also used to perform the following steps: based on sample data of multiplex PCR detection according to multiple preset targets, a sample user feature set and a sample respiratory feature set of multiple sample users are collected, and a sample multiplex PCR detection spectrum set for multiplex PCR detection is collected; using deep learning, a first multiplex PCR prediction network and a second multiplex PCR prediction network are constructed; using the sample user feature set and the sample respiratory feature set as input training data, and using the sample multiplex PCR detection spectrum set as output supervision data, the first multiplex PCR prediction network and the second multiplex PCR prediction network are supervised trained, verified and tested until convergence; the converged first multiplex PCR prediction network and the second multiplex PCR prediction network are configured on a cloud server to obtain a multiplex PCR prediction channel.

[0091] Furthermore, the error calculation module 14 is also used to perform the following steps: calculating the absolute error between the actual multiplex PCR detection spectrum and the first predicted multiplex PCR detection spectrum and the second predicted multiplex PCR detection spectrum to obtain a first detection spectrum error and a second detection spectrum error; and calculating a detection error amplitude based on the first detection spectrum error and the second detection spectrum error.

[0092] Furthermore, the error calculation module 14 is also used to perform the following steps: when the detection error amplitude is greater than or equal to a preset error amplitude threshold, the detection error amplitude and feature confidence input are configured in the detection abnormality pattern classification channel of the cloud server, and multiple classification detection abnormality patterns are obtained through classification output of multiple detection abnormality pattern classification branches; the classification detection abnormality pattern with the highest occurrence frequency among the multiple classification detection abnormality patterns is screened to obtain a detection abnormality pattern; and the actual multiplex PCR detection spectrum is annotated using the detection abnormality pattern as a data analysis result.

[0093] Furthermore, the error calculation module 14 is also used to perform the following steps: based on the detection records of the multiple PCR detection, the sample feature confidence set and the sample detection error amplitude set are collected, and the abnormality types detected under different combinations of sample feature confidence and sample detection error amplitude are collected, and the sample abnormal pattern set is marked; the sample feature confidence set, the sample detection error amplitude set and the sample abnormal pattern set are divided into N parts to obtain N abnormal pattern classification training data, where N is a positive integer; N abnormal pattern classification training data are used respectively to construct N detection abnormal pattern classification branches, obtain detection abnormal pattern classification channels, and configure them on the cloud server.

[0094] Through the above detailed description of a data analysis method for multiplex PCR multi-target parallel amplification in this specification, those skilled in the art can clearly understand a data analysis system for multiplex PCR multi-target parallel amplification in this embodiment. For the system disclosed in the embodiment, since it corresponds to the method disclosed in the embodiment, the description is relatively simple, and the relevant details can be referred to the method section.

[0095] The above description of the disclosed embodiments is intended to enable one skilled in the art to implement or use the present application. Various modifications to these embodiments will be readily apparent to one skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present application. Therefore, the present application is not limited to the embodiments shown herein, but is intended to conform to the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A data analysis method for multiplex PCR multi-target parallel amplification, characterized in that: The method comprises: Collect the user characteristics and respiratory characteristics of the target user and analyze them to obtain the feature confidence; Predicting, based on the user characteristics and respiratory characteristics, the test results of multiplex PCR tests performed according to multiple preset targets, to obtain a first predicted multiplex PCR test profile and a second predicted multiplex PCR test profile; Collecting a test sample from the target user, performing multiplex PCR parallel amplification detection according to the multiple preset targets, and obtaining an actual multiplex PCR detection spectrum; The error between the actual multiplex PCR detection spectrum and the first predicted multiplex PCR detection spectrum and the second predicted multiplex PCR detection spectrum is calculated to obtain the detection error amplitude. Combined with the feature confidence, the detection abnormality pattern is classified and obtained, and the actual multiplex PCR detection spectrum is annotated as the data analysis result.

2. The data analysis method for multiplex PCR multi-target parallel amplification according to claim 1, characterized in that: Collect the target user's user characteristics and respiratory characteristics, and analyze them to obtain feature confidence, including: Collecting user characteristics of the target user, wherein the user characteristics include age and weight; Collecting respiratory characteristics of the target user, wherein the respiratory characteristics include at least sample physical properties, environmental parameters, and respiratory symptom indicators; Analyze and obtain feature confidences of the user features and respiratory features.

3. The data analysis method for multiplex PCR multi-target parallel amplification according to claim 2, characterized in that: include: Retrieving the user feature in a user database within a recent preset time range to obtain a first occurrence rate, wherein the user database updates and stores the user features and respiratory features within the recent preset time range; Retrieving the respiratory feature in the user database to obtain a second occurrence rate; A feature confidence is calculated based on the first occurrence rate and the second occurrence rate.

4. The data analysis method for multiplex PCR multi-target parallel amplification according to claim 1, characterized in that: Predicting the test results of multiplex PCR tests according to multiple preset targets based on the user characteristics and respiratory characteristics to obtain a first predicted multiplex PCR test profile and a second predicted multiplex PCR test profile, including: Invoking a multiplex PCR prediction channel configured on a cloud server, wherein the multiplex PCR prediction channel includes a first multiplex PCR prediction network and a second multiplex PCR prediction network; The user characteristics and respiratory characteristics are transmitted to the cloud server and respectively input into the first multiplex PCR prediction network and the second multiplex PCR prediction network in the multiplex PCR prediction channel, and the prediction output obtains the first predicted multiplex PCR detection spectrum and the second predicted multiplex PCR detection spectrum for multiplex PCR detection according to multiple preset targets.

5. The data analysis method for multiplex PCR multi-target parallel amplification according to claim 4, characterized in that: The configuration step of the multiplex PCR prediction channel includes: Collecting sample user feature sets and sample respiratory feature sets of multiple sample users based on sample data of multiplex PCR tests according to multiple preset targets, and collecting a multiplex PCR test profile set of samples for multiplex PCR tests; Using deep learning, we constructed the first multiplex PCR prediction network and the second multiplex PCR prediction network; Using the sample user feature set and the sample respiratory feature set as input training data and the sample multiplex PCR detection spectrum set as output supervision data, respectively, the first multiplex PCR prediction network and the second multiplex PCR prediction network are supervised trained, validated, and tested until convergence; The converged first multiplex PCR prediction network and the second multiplex PCR prediction network are configured on a cloud server to obtain a multiplex PCR prediction channel.

6. The data analysis method for multiplex PCR multi-target parallel amplification according to claim 1, characterized in that: Calculating the error between the actual multiplex PCR detection spectrum and the first predicted multiplex PCR detection spectrum and the second predicted multiplex PCR detection spectrum to obtain a detection error margin includes: calculating absolute errors between the actual multiplex PCR detection spectrum and the first predicted multiplex PCR detection spectrum and the second predicted multiplex PCR detection spectrum to obtain a first detection spectrum error and a second detection spectrum error; A detection error amplitude is obtained by calculation according to the first detection spectrum error and the second detection spectrum error.

7. The data analysis method for multiplex PCR multi-target parallel amplification according to claim 1, characterized in that: Combined with the feature confidence, the abnormal detection pattern is classified and obtained, and the actual multiplex PCR detection spectrum is annotated as a data analysis result, including: When the detection error amplitude is greater than or equal to a preset error amplitude threshold, the detection error amplitude and the feature confidence are input into a detection anomaly pattern classification channel configured in the cloud server, and multiple classification detection anomaly patterns are obtained through classification outputs of multiple detection anomaly pattern classification branches; Filtering the classification detection anomaly pattern with the highest frequency among the multiple classification detection anomaly patterns to obtain a detection anomaly pattern; The actual multiplex PCR detection spectrum is annotated using the detection anomaly pattern as a data analysis result.

8. The data analysis method for multiplex PCR multi-target parallel amplification according to claim 7, characterized in that: The configuration steps of the abnormal pattern detection classification channel include: Based on the test records of multiplex PCR tests, a set of sample feature confidences and a set of sample detection error margins are collected. The abnormality types detected under different combinations of sample feature confidences and sample detection error margins are collected and annotated to obtain a set of sample abnormality patterns. Performing N-fold division on the sample feature confidence set, the sample detection error amplitude set, and the sample abnormal pattern set to obtain N pieces of abnormal pattern classification training data, where N is a positive integer; N sets of abnormal pattern classification training data are used to construct N abnormal pattern detection classification branches, obtain abnormal pattern detection classification channels, and configure them on the cloud server.

9. A data analysis system for multiplex PCR multi-target parallel amplification, characterized in that: A data analysis method for implementing the multiplex PCR multi-target parallel amplification according to any one of claims 1 to 8, the system comprising: The feature collection module is used to collect the user features and respiratory features of the target user and analyze them to obtain feature confidence; a detection profile acquisition module, configured to predict, based on the user characteristics and respiratory characteristics, the detection results of a multiplex PCR test performed according to a plurality of preset targets, to obtain a first predicted multiplex PCR detection profile and a second predicted multiplex PCR detection profile; an amplification detection module, configured to collect a test sample from the target user, perform multiplex PCR parallel amplification detection according to the plurality of preset targets, and obtain an actual multiplex PCR detection profile; An error calculation module is configured to calculate the error between the actual multiplex PCR detection profile and the first and second predicted multiplex PCR detection profiles to obtain a detection error amplitude, classify the detection anomaly patterns based on the feature confidence, and annotate the actual multiplex PCR detection profile as a data analysis result.

Images