Metabolically unhealthy obesity identification model construction system, storage medium and kit

Abstract

A metabolically unhealthy obesity identification model construction system, a storage medium and a kit. The identification model construction system comprises a collection module, a feature selection module, a training module and an identification module. By means of performing machine learning on plasma metabolite biomarkers, early identification of metabolically unhealthy obese patients can be achieved before clinical symptoms manifest, thereby helping doctors and patients take intervention measures prior to an increase in cardiovascular risk. Compared with conventional physical examination procedures, the metabolically unhealthy obesity identification model construction system is simpler, and reduces the participation burden for patients.

Description

Metabolically unhealthy obesity identification model construction system, storage medium and kit TECHNICAL FIELD

[0001] The present application relates to the field of medical prediction, in particular to a metabolically unhealthy obesity identification model construction system, a storage medium and a kit and applications thereof. BACKGROUND

[0002] Studies have gradually shown that not all obese patients exhibit the same difference in cardiovascular disease risk. Based on these observations, two concepts of metabolically healthy obesity (MHO) and metabolically unhealthy obesity (MUO) have emerged. On the basis of obesity, if hypertension and glucose and lipid metabolism disorders occur, it is classified as MUO. Such patients usually have visceral fat and ectopic fat deposition, and under the physiological and pathological effects, adipose tissue can secrete a large amount of harmful adipokines and widely participate in the progression of various cardiovascular diseases. In contrast, MHO is a benign obesity. Although such patients are overweight, they do not have obvious metabolic disorders such as hypertension, hyperglycemia or dyslipidemia, and the risk of cardiovascular disease is relatively low. MUO patients are accompanied by hypertension, glucose and lipid metabolism disorders, and their cardiovascular event risk is significantly increased.

[0003] Currently, the differential diagnosis of MHO and MUO mainly relies on detecting whether the patient has developed symptoms such as elevated blood glucose, blood pressure or blood lipid levels. However, by the time these symptoms appear, the cardiovascular disease risk of MUO patients is already very high. Existing physical examinations require a variety of hematological measurements and a series of questionnaires involving past medical history, medication history, which is cumbersome and time-consuming and requires the full cooperation of the participants, and cannot effectively identify metabolically unhealthy status at an early stage. SUMMARY

[0004] The purpose of the present application is to provide a metabolically unhealthy obesity identification model construction system, a storage medium and a kit. The metabolically unhealthy obesity identification model construction system comprises a collection module, a feature selection module, a training module and an identification module. By using the relative concentration values of specific plasma small molecule metabolic biomarkers to construct the metabolically unhealthy obesity identification model, metabolically unhealthy obesity patients can be identified at an early stage before clinical symptoms appear, helping doctors and patients to take intervention measures before the cardiovascular risk increases.

[0005] To solve the above problems, the present application provides the following technical solutions.

[0006] In a first aspect, the present application provides a metabolic unhealthy obesity identification model construction system, which comprises a collection module, a feature selection module, a training module and an identification module. In this embodiment, the collection module is used to collect plasma samples of each subject of metabolic healthy obesity and metabolic unhealthy obesity, and obtain plasma small molecule metabolic biomarkers and their relative concentration values from each plasma sample. The feature selection module is used to perform feature selection based on the plasma small molecule metabolic biomarkers and their relative concentration values obtained from each plasma sample, to screen plasma small molecule metabolic biomarkers for early identification of metabolic unhealthy obesity as feature-selected plasma small molecule metabolic biomarkers. The training module is used to train a final prediction model for early identification of metabolic unhealthy obesity based on the feature-selected plasma small molecule metabolic biomarkers and their relative concentration values in each plasma sample. The identification module is used to obtain feature-selected plasma small molecule metabolic biomarkers and their relative concentration values of a subject to be detected, and input the feature-selected plasma small molecule metabolic biomarkers and their relative concentration values of the subject to be detected into the final prediction model to obtain a prediction result of whether the subject to be detected belongs to metabolic unhealthy obesity.

[0007] Further, in the metabolic unhealthy obesity identification model construction system described above, the collection module is used to obtain a first subject group and a second subject group which are not overlapped, wherein the first subject group and the second subject group each comprise subjects of metabolic healthy obesity and metabolic unhealthy obesity which are not overlapped. The collection module can also be used to collect first plasma samples of the first subject group, and obtain plasma small molecule metabolic biomarkers and their relative concentration values from each first plasma sample as a first set. The collection module can also be used to collect second plasma samples of the second subject group, and obtain plasma small molecule metabolic biomarkers and their relative concentration values from each second plasma sample as a test set.

[0008] Further, in the metabolic unhealthy obesity identification model construction system described above, the feature selection module is used to perform feature selection on all plasma small molecule metabolic biomarkers such as 303 plasma small molecule metabolic biomarkers on the first set using Lasso feature selection algorithm and five-fold cross-validation on R (4.3.1) software, and the implementation of Lasso feature selection algorithm depends on glmnet package. The glmnet package generates a sequence of 100 values corresponding to hyperparameter lambda by built-in program, which is used to control the strength of L1 regularization penalty on model coefficients, and determines lambda.min when the misclassification rate is the lowest in five-fold cross-validation. Finally, the calculated lambda.min is 0.02, and the metabolite biomarkers are screened based on lambda.min.

[0009] Further, in the above-mentioned identification model construction system for metabolically unhealthy obesity, the feature selection module is configured to divide the first set into five mutually non-overlapping data, each time taking one of the five mutually non-overlapping data that has not been selected as an internal validation set, and each time taking the remaining four data as a training set; wherein the internal validation set and the training set each include subjects of metabolically healthy obesity and metabolically unhealthy obesity that are mutually non-overlapping. The feature selection module is further configured to, based on the training set each time, use the Lasso feature selection algorithm on the R (4.3.1) software to train the feature selection model for 5 rounds to obtain 5 corresponding feature selection models.

[0010] The feature selection module is further configured to, for each value in the sequence of 100 values corresponding to the hyperparameter lambda, input a corresponding one of the 5 feature selection models to the internal validation set each round to obtain one misclassification rate corresponding to each value on each internal validation set; obtain one average misclassification rate corresponding to each value from the 5 misclassification rates corresponding to each value; select lambda.min corresponding to the lowest misclassification rate based on the smallest average misclassification rate among all values, wherein lambda is one of the hyperparameters; and obtain the corresponding metabolite markers based on the lambda.min corresponding to the lowest misclassification rate as the plasma small molecule metabolic biomarkers after feature selection.

[0011] Further, in the above-mentioned system for constructing a model for identifying metabolically unhealthy obesity, the feature-selected plasma small molecule metabolic biomarkers include: Cysteamine, 4-Methylpentanal, (R)-3-Hydroxybutyric acid, 4-Aminophenol, Pyruvic acid, L-Threonine, Succinic anhydride, D-Threitol, Glycerol, Hypotaurine, Creatinine, Adenine, L-Valine, L-Cysteine, Picolinic acid, Taurine, L-Glutamic acid, L-Methionine, Citrulline, D-Glucose, L-Arginine, Ascorbic acid, Lipoic acid, Propionylcarnitine, Glycerophosphocholine, Stearic acid and Sphingosine.

[0012] Further, in the above-mentioned system for constructing a model for identifying metabolically unhealthy obesity, the training module is configured to train a prediction model for early identification of metabolically unhealthy obesity based on the feature-selected plasma small molecule metabolic biomarkers and their relative concentration values in each plasma sample.

[0013] The training module is further configured to perform performance evaluation on the trained prediction model for early identification of metabolically unhealthy obesity, and select the prediction model for early identification of metabolically unhealthy obesity that meets preset requirements of performance evaluation as a final prediction model.

[0014] Further, in the above-mentioned system for constructing a model for identifying metabolically unhealthy obesity, the training module is configured to train a prediction model for early identification of metabolically unhealthy obesity based on the first set using an elastic network algorithm on R (4.3.1) software, and the elastic network algorithm is implemented by means of a glmnet package, wherein an ElasticNet algorithm hyperparameter alpha for controlling the ratio of L1 regularization and L2 regularization is set to 0.5, and s for controlling the punishment degree of parameters is set to 0.01.

[0015] Further, in the above-mentioned identification model construction system for early identification of metabolically unhealthy obesity, the training module is configured to obtain a trained prediction model for early identification of metabolically unhealthy obesity on the R (4.3.1) software, and to obtain the identification performance on the first set and the identification performance on the test set, respectively; if the trained prediction model for early identification of metabolically unhealthy obesity has an identification performance difference between the training set and the test set less than a preset difference threshold, the trained prediction model for early identification of metabolically unhealthy obesity is taken as the final prediction model.

[0016] In a second aspect, the present application provides a computer readable storage medium having computer executable instructions stored thereon, wherein the computer executable instructions, when executed by a processor, cause the processor to perform the following steps:

[0017] Step S1, a collection module collects blood plasma samples of each subject of metabolically healthy obesity and metabolically unhealthy obesity, and obtains blood plasma small molecule metabolic biomarkers and relative concentration values thereof from each blood plasma sample;

[0018] Step S2, a feature selection module performs feature selection based on the blood plasma small molecule metabolic biomarkers and relative concentration values thereof in each blood plasma sample, to screen blood plasma small molecule metabolic biomarkers for early identification of metabolically unhealthy obesity, as feature-selected blood plasma small molecule metabolic biomarkers;

[0019] Step S3, a training module trains a final prediction model for early identification of metabolically unhealthy obesity based on the feature-selected blood plasma small molecule metabolic biomarkers and relative concentration values thereof in each blood plasma sample;

[0020] Step S4, an identification module obtains feature-selected blood plasma small molecule metabolic biomarkers and relative concentration values thereof of a to-be-detected subject, and inputs the feature-selected blood plasma small molecule metabolic biomarkers and relative concentration values thereof of the to-be-detected subject into the final prediction model, to obtain a prediction result of whether the to-be-detected subject belongs to metabolically unhealthy obesity.

[0021] In this embodiment, the feature-selected blood plasma small molecule metabolic biomarkers include mercaptoethylamine, 4-methylpentanal, 3-hydroxybutyric acid, aminophenol, pyruvic acid, threonine, succinic anhydride, threitol, glycerol, hypotaurine, creatinine, adenine, valine, cysteine, isonicotinic acid, taurine, glutamic acid, methionine, citrulline, glucose, arginine, ascorbic acid, lipoic acid, acetylcarnitine, glycerophosphocholine, stearic acid, and sphingosine.

[0022] In a third aspect, the present application provides a detection kit for early identification of metabolically unhealthy obesity, comprising metabolite biomarkers, which at least include the following substances: mercaptoethylamine, 4-methylpentanal, 3-hydroxybutyric acid, aminophenol, pyruvic acid, threonine, succinic anhydride, threitol, glycerol, hypotaurine, creatinine, adenine, valine, cysteine, isonicotinic acid, taurine, glutamic acid, methionine, citrulline, glucose, arginine, ascorbic acid, lipoic acid, acetylcarnitine, glycerophosphocholine, stearic acid and sphingosine.

[0023] Specifically, the beneficial effects of the present application are as follows.

[0024] Firstly, strong early identification capability: through comprehensive analysis of metabolites in plasma by metabolomics, metabolic abnormalities in metabolically unhealthy obesity patients can be identified in time before the occurrence of clinical symptoms such as hyperglycemia and hypertension, thereby providing a scientific basis for early intervention.

[0025] Secondly, high throughput and high sensitivity: the metabolically unhealthy obesity identification model construction system is based on high-throughput mass spectrometry techniques such as nanoparticle-enhanced laser desorption / ionization time-of-flight mass spectrometry (MALDI-TOF-MS), which can simultaneously detect a large number of metabolites, and has extremely high sensitivity and specificity, ensuring the accuracy of metabolically unhealthy obesity identification.

[0026] In addition, convenience and generalizability: plasma sample collection is simple, the detection process is simplified, and it can be applied to large-scale population screening, suitable for various scenes such as clinical, scientific research and physical examination.

[0027] Finally, individualized health management support: based on the identification model and metabolite combination of the present application, data support can be provided for individualized obesity management, helping to develop personalized intervention programs and reduce the risk of obesity-related cardiovascular diseases.

[0028] Therefore, the present application not only provides an efficient and reliable technical solution for early identification of metabolically unhealthy obesity, but also provides important basic data and application prospects for cardiovascular disease prevention and obesity intervention. BRIEF DESCRIPTION OF DRAWINGS

[0029] Fig. 1 is a diagram of the relative concentration values of 303 metabolite biomarkers in the plasma of metabolically healthy obesity (MHO) and metabolically unhealthy obesity (MUO) subjects according to an embodiment of the present application.

[0030] Fig. 2 is a performance diagram of 27 metabolite biomarker combinations in the first set and independent test set for identifying MUO according to an embodiment of the present application.

[0031] Fig. 3a-3c are comparisons of 10-year cardiovascular event risks based on the independent test set in an embodiment of the present application, Fig. 3a is the number of MHO and MUO in the independent test set according to clinical judgment / metabolomics prediction; Fig. 3b is a comparison of 10-year cardiovascular event risks of MHO and MUO patients in the independent test set predicted by metabolomics; Fig. 3c is a comparison of 10-year cardiovascular event risks of MHO and MUO patients in the independent test set predicted by metabolomics in individuals judged as MHO according to clinical symptoms;

[0032] Fig. 4 shows a schematic diagram of the module structure of the system for constructing the identification model of the cardiovascular metabolic risk factor spectrum in an embodiment of the present application;

[0033] Fig. 5 shows a schematic diagram of the flow of the method for identifying the cardiovascular metabolic risk factor spectrum in an embodiment of the present application. DETAILED DESCRIPTION

[0034] The present application will be further described below in conjunction with the drawings.

[0035] In a typical configuration of the present application, the terminal, the device of the service network and the trusted party each comprises one or more processors (CPU), input / output interface, network interface and memory.

[0036] The memory can include non-permanent memory in the computer readable medium, random access memory (RAM) and / or non-volatile memory such as read-only memory (ROM) or flash memory (flash RAM). The memory is an example of the computer readable medium.

[0037] The computer readable medium includes permanent and non-permanent, removable and non-removable media, which can be implemented by any method or technology to store information. The information can be computer readable instructions, data structures, program modules or other data. Examples of computer storage media include, but are not limited to, phase change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technology, compact disc read-only memory (CD-ROM), digital versatile disc (DVD) or other optical storage, magnetic cassette, magnetic tape disk storage or other magnetic storage device, or any other non-transmission medium that can be used to store information accessible by a computing device. According to the definition herein, the computer readable medium does not include non-transitory computer readable media such as modulated data signals and carriers.

[0038] In order to make the purposes, technical solutions and advantages of the embodiments of the present application clearer, the various embodiments of the present application will be described in detail below with reference to the embodiments. Unless otherwise specified, the experimental methods in the embodiments are generally performed according to conventional conditions, for example, the conditions described in textbooks and experimental guidelines, or the conditions recommended by the manufacturers. The recommended conditions of the supporting software are run.

[0039] As shown in FIG. 5, the present application provides a recognition model construction system for metabolically unhealthy obesity, which comprises a collection module 10, a feature selection module 20, a training module 30 and a recognition module 40.

[0040] In one embodiment, as shown in FIG. 5, the present application provides an early recognition method for metabolically unhealthy obesity based on plasma metabolomics, which comprises:

[0041] Step S1, the collection module collects the plasma samples of each subject of metabolically healthy obesity and metabolically unhealthy obesity, and obtains the plasma small molecule metabolic biomarkers and their relative concentration values from each plasma sample;

[0042] Step S2, the feature selection module performs feature selection based on the plasma small molecule metabolic biomarkers and their relative concentration values obtained from each plasma sample, to screen the plasma small molecule metabolic biomarkers for early recognition of metabolically unhealthy obesity, as the plasma small molecule metabolic biomarkers after feature selection;

[0043] Step S3, the training module trains the final prediction model for early recognition of metabolically unhealthy obesity based on the plasma small molecule metabolic biomarkers after feature selection and their relative concentration values in each plasma sample;

[0044] Step S4, the recognition module obtains the plasma small molecule metabolic biomarkers after feature selection and their relative concentration values of the to-be-detected subject, and inputs the plasma small molecule metabolic biomarkers after feature selection and their relative concentration values of the to-be-detected subject into the final prediction model, to obtain the prediction result of whether the to-be-detected subject belongs to metabolically unhealthy obesity.

[0045] In one embodiment, the collection module can be used to collect the plasma samples of each subject of metabolically healthy obesity (MHO) and metabolically unhealthy obesity (MUO), and obtain the plasma small molecule metabolic biomarkers and their relative concentration values from each plasma sample.

[0046] Metabolomics, as a cutting-edge omics technology, can detect metabolic disorders in the body before the onset of disease symptoms by performing high-throughput, global analysis of metabolites in biological samples. Therefore, it has great potential in the early diagnosis of diseases. Plasma, as an easily accessible and information-rich biological fluid, is an important source of metabolites in the body. Plasma-based metabolomics analysis can provide valuable resources for the early identification of MUO patients and promote the prevention and intervention of obesity-related cardiovascular diseases.

[0047] First, the collection module can obtain a first subject group and a second subject group that do not overlap with each other, wherein the first subject group and the second subject group each include respective subjects of metabolic healthy obesity (MHO) and metabolic unhealthy obesity (MUO) that do not overlap with each other.

[0048] Specifically, the first subject group includes metabolic healthy obesity (MHO) and metabolic unhealthy obesity (MUO); the second subject group includes metabolic healthy obesity (MHO) and metabolic unhealthy obesity (MUO); and the first subject group and the second subject group do not overlap with each other.

[0049] Second, the collection module can collect first plasma samples of the first subject group, and obtain plasma small molecule metabolic biomarkers and their relative concentration values from each first plasma sample as a first set.

[0050] Finally, the collection module can collect second plasma samples of the second subject group, and obtain plasma small molecule metabolic biomarkers and their relative concentration values from each second plasma sample as a test set.

[0051] Specifically, with the approval of the ethics committee of a certain hospital, subjects from three different communities from January 2019 to March 2019 can be recruited to obtain the first set and the test set. The first set includes 167 cases of metabolic healthy obesity (MHO) and 167 cases of metabolic unhealthy obesity (MUO) from the first and second communities; the first set can be split into a training set and an internal validation set in subsequent feature screening. It should be noted that the feature screening model requires an internal validation set, and the obesity model does not require an internal validation set.

[0052] The test set includes 39 cases of metabolic healthy obesity (MHO) and 39 cases of metabolic unhealthy obesity (MUO) from the third community, which are used to independently verify the accuracy and clinical value of the metabolomics-based model.

[0053] The inclusion criteria for the subject group can be as follows:

[0054] Obesity criteria: According to the 2004 Expert Consensus on Obesity Prevention and Treatment for Chinese Residents published by the Chinese Nutrition Society, BMI ≥ 28 kg / m2 The subject is defined as obese.

[0055] Metabolic abnormality criteria: According to the metabolic syndrome diagnostic criteria of the Chinese Medical Association Diabetes Branch in 2004, the following three are included:

[0056] Hyperglycemia: fasting blood glucose ≥ 6.1 mmol / L (110 mg / dl) or two-hour postprandial blood glucose ≥ 7.8 mmol / L (140 mg / dl), or those who have been diagnosed as diabetes and treated;

[0057] Hypertension: systolic / diastolic blood pressure ≥ 140 / 90 mmHg, or those who have been diagnosed as hypertension and treated;

[0058] Dyslipidemia: fasting serum triglycerides ≥ 1.7 mmol / L (150 mg / dl), or high-density lipoprotein cholesterol (HDL-C) < 0.9 mmol / L (male) or < 1.0 mmol / L (female).

[0059] MHO is obese people without the above metabolic abnormalities, and MUO is obese people with at least one or more of the above metabolic abnormalities.

[0060] Exclusion criteria: Exclusion of subjects with acute or infectious clinical symptoms (such as fever, cough, sore throat, etc.) within three weeks before sampling to ensure that metabolic markers are not affected by acute diseases.

[0061] The blood plasma sample and the relative concentration value obtained from the subject can include the following steps S121 to S128.

[0062] Step S121 includes: the plasma sample of the peripheral venous blood of the subject in the morning fasting state is treated with EDTA anticoagulation and stored at -80℃. In this step, 1 ml of peripheral venous blood of the subject can be collected in the morning fasting state, and treated with EDTA anticoagulation tube. After plasma separation, the sample is stored at -80℃ to ensure the stability of metabolites.

[0063] Step S122 includes: preparation of instruments and reagents: preparation of deionized water, matrix (inorganic nanoparticles), and nanoparticle-enhanced laser desorption / ionization time-of-flight mass spectrometer (MALDI-TOF-MS, Bruker, Germany).

[0064] Step S123 includes: dilution treatment of the plasma sample: the plasma sample is diluted with deionized water at a standard ratio to obtain a diluted plasma sample, and the concentration of metabolites is adjusted to be suitable for mass spectrometry analysis. The plasma sample is diluted. Specifically, 100 nL of plasma sample can be diluted 10 times with deionized water to obtain a diluted plasma sample to adjust the concentration of metabolites suitable for mass spectrometry analysis.

[0065] Step S124 includes: preparing the inorganic nanoparticles into a nanoparticle matrix solution for enhancing the mass spectrum signal. Specifically, the matrix (inorganic nanoparticles) is prepared into a 1 mg / mL matrix solution with deionized water, ensuring that the metabolites are well combined with the matrix to obtain clear mass spectrum signals.

[0066] Step S125 includes: performing sample preparation on the mass spectrum target plate: the diluted plasma sample is spotted onto the mass spectrum target plate and dried at room temperature.

[0067] Specifically, sample preparation can be performed on the mass spectrum target plate, with each diluted plasma sample being spotted 500 nL, and dried at room temperature.

[0068] Step S126 includes: performing matrix preparation on the mass spectrum target plate: the nanoparticle matrix solution is spotted onto the mass spectrum target plate, ensuring that the nanoparticle matrix solution uniformly covers the plasma sample spot to obtain the sample. Specifically, matrix preparation can be performed on the mass spectrum target plate, with each matrix solution being spotted 500 nL on each plasma sample spot, and dried at room temperature.

[0069] Step S127 includes: using a nanoparticle-enhanced laser desorption / ionization time-of-flight mass spectrometer (MALDI-TOF-MS) to perform data acquisition in positive ion mode, with a 1000 Hz repetition rate, a 20 kV acceleration voltage, and a 250 ns delay time, and recording 2000 laser shots in each analysis to obtain the metabolite peak map in the sample.

[0070] Specifically, data acquisition is performed in a nanoparticle-enhanced laser desorption / ionization time-of-flight mass spectrometer, with data acquisition extracted in positive ion mode, using delayed extraction, a repetition rate of 1000 Hz, an acceleration voltage of 20 kV, a delay time of 250 ns, and 2000 laser shots per analysis.

[0071] Step S128 includes: based on the metabolite peak map in the sample, under the conventional standardization conditions of the nanoparticle-enhanced laser desorption / ionization time-of-flight mass spectrometer software, running the built-in data preprocessing pipeline, including spectral smoothing, baseline correction, spectral peak matching, and log transformation and Min-Max normalization in sequence, to obtain the relative intensity (%) of 303 metabolite biomarkers in plasma, i.e. the relative concentration value, as shown in the heat map of FIG. 1.

[0072] Specifically, the Savitzky-Golay (S-G) filter can be used for spectral line smoothing; then an adaptive iterative algorithm is used to identify baseline points, polynomial fitting is performed on the identified baseline points in the metabolite peak map, the fitted baseline is subtracted from the original spectrum, and the corrected spectrum is checked to ensure that the baseline is flat and has no negative values; then the peak detection parameters are set: signal-to-noise ratio threshold = 3, mass accuracy tolerance 50 ppm, the continuous wavelet transform (CWT) is used to identify the peak positions of possible metabolite markers, and further peak fitting is performed to obtain the corresponding peak position, peak height and peak area, and the metabolite marker level is relatively quantified according to the peak height size; finally, logarithmic transformation and Min-Max normalization are performed to obtain the relative concentration values of 303 markers.

[0073] Here, the relative concentration value obtained can be the peak intensity value after logarithmic transformation and min-max normalization, with a unit of %, used to measure the relative concentration of metabolites in plasma.

[0074] The 303 plasma small molecule metabolite biomarkers can be identified by comparing the HMDB (Human Metabolome Database) database.

[0075] In one embodiment, the feature selection module is used to perform feature selection based on the plasma small molecule metabolite biomarkers and their relative concentration values obtained in each plasma sample, to screen plasma small molecule metabolite biomarkers for early identification of metabolically unhealthy obesity, as the plasma small molecule metabolite biomarkers after feature selection.

[0076] First, the feature selection module Lasso feature selection: using the Lasso feature selection algorithm and five-fold cross-validation on the R (4.3.1) software, all plasma small molecule metabolite biomarkers such as 303 plasma small molecule metabolite biomarkers on the first set are subjected to feature selection, and the implementation of the Lasso feature selection algorithm depends on the glmnet package, wherein the glmnet package generates a sequence of 100 values corresponding to the hyperparameter lambda through the built-in program, which is used to control the strength of the L1 regularization penalty on the model coefficients, and determines lambda.min when the misclassification rate is the lowest in the five-fold cross-validation.

[0077] 100 value sequence, which is used to control the strength of the L1 regularization penalty on the model coefficients, and determines lambda.min when the misclassification rate is the lowest in the five-fold cross-validation.

[0078] The final calculated lambda.min = 0.02, and 27 metabolite biomarkers are screened out by taking the minimum misclassification rate as the standard, and the 27 metabolite biomarkers show good effect in identifying metabolically unhealthy obesity. Specifically, the misclassification rate is the smallest when lambda = 0.02, and the lambda at this time is called lambda.min.

[0079] Table 1: Information of 27 metabolite biomarkers

[0080] As shown in Table 1, the 27 metabolite biomarkers include mercaptoethanol, 4- methylpentanal, 3-hydroxybutyric acid, aminophenol, pyruvic acid, threonine, succinic anhydride, threitol, glycerol, hypotaurine, creatinine, adenine, valine, cysteine, isonicotinic acid, taurine, glutamic acid, methionine, citrulline, glucose, arginine, ascorbic acid, lipoic acid, acetylcarnitine, glycerophosphocholine, stearic acid and sphingosine.

[0081] In particular, R software is a language and environment for statistical computing and graphics. Cross-platform support for Windows Mac Linux, etc. It is a GNU project, similar to the S language and environment developed at Bell Labs (predecessor of AT&T, now Lucent Technologies) by John Chambers and his colleagues. R can be considered a different implementation of S. There are some important differences, but much of the code written for S remains unchanged when run under R. R provides a wide range of statistical (linear and nonlinear modeling, classical statistical tests, time series analysis, classification, clustering, etc.) and graphical techniques, and is highly extensible. The S language is usually the preferred tool for statistical method research, and R language provides an open source approach to participating in this activity.

[0082] In one embodiment, the feature selection module Lasso feature selection can include the following steps.

[0083] First, the feature selection module divides the first set into 5 non-overlapping data, each time taking 1 data not selected from the 5 non-overlapping data as an internal validation set, and each time taking the remaining 4 data as a training set; wherein the internal validation set and the training set each include non-overlapping metabolic healthy obese (MHO) and metabolic unhealthy obese (MUO) subjects. Here, the internal validation set includes metabolic healthy obese (MHO) and metabolic unhealthy obese (MUO) subjects; the training set includes metabolic healthy obese (MHO) and metabolic unhealthy obese (MUO) subjects; the internal validation set and the training set are non-overlapping. For example, the subjects are divided into a first set (a total of 234 cases, including 167 cases of MHO and 167 cases of UHO) and a test set (a total of 78 cases, including 39 cases of MHO and 39 cases of UHO).

[0084] Second, the feature selection module uses the Lasso feature selection algorithm based on the training set each time on the R (4.3.1) software, and trains the feature selection model for 5 rounds in a loop to obtain the corresponding 5 feature selection models.

[0085] Finally, the feature selection module inputs a corresponding one of the five feature selection models in each round for each value in the sequence of 100 values of the hyperparameter lambda, to obtain a corresponding one of the five misclassification rates of each value on each internal validation set; the feature selection module obtains an average misclassification rate corresponding to each value from the five misclassification rates corresponding to each value; the feature selection module selects a lambda.min corresponding to the lowest misclassification rate based on the smallest average misclassification rate among all values, where lambda is one of the hyperparameters; and the feature selection module obtains the 27 metabolic markers corresponding to the lambda.min with the lowest misclassification rate.

[0086] Here, based on the lambda.min being the best hyperparameter lambda, the corresponding features are extracted to screen out the 27 metabolic markers. The 27 metabolic biomarker-related information is shown in Table 1.

[0087] Specifically, the lambda hyperparameters of the five feature selection models are the same. For each lambda value, the used features are determined by the lambda value, and the corresponding metabolic markers can be obtained based on the features. The five feature selection models can be obtained by the training set. By observing the average performance of the lambda value on the internal validation set of the five models, it is determined that the average performance of which lambda value on the internal validation set is the best, i.e., the average misclassification rate is the lowest.

[0088] In an embodiment, the training module is configured to train a final prediction model for early identification of metabolically unhealthy obesity based on the selected plasma small molecule metabolic biomarkers and their relative concentration values in each plasma sample.

[0089] Specifically, the normalized relative concentrations (%) of the 27 metabolic biomarkers in the MHO group and the MHO group are shown in Table 2.

[0090] Table 2: Normalized relative concentrations (%) of 27 metabolic biomarkers in the MHO group and the MHO group

[0091] Training a final prediction model for early identification of metabolically unhealthy obesity based on the selected plasma small molecule metabolic biomarkers and their relative concentration values in each plasma sample includes: first, the training module trains a prediction model for early identification of metabolically unhealthy obesity based on the selected plasma small molecule metabolic biomarkers and their relative concentration values in each plasma sample; second, the training module performs performance evaluation on the trained prediction model for early identification of metabolically unhealthy obesity, and takes the prediction model for early identification of metabolically unhealthy obesity that meets the preset requirements of performance evaluation as the final prediction model.

[0092] The training module trains a prediction model for early identification of metabolically unhealthy obesity based on the selected plasma small molecule metabolic biomarkers and their relative concentration values in each plasma sample, including:

[0093] The training module trains the prediction model for early identification of metabolically unhealthy obesity based on the first set using an ElasticNet algorithm implemented in the glmnet package on R (4.3.1) software, wherein the ElasticNet algorithm hyperparameter alpha is set to 0.5 to control the ratio of L1 regularization and L2 regularization, and the s is set to 0.01 to control the penalty strength on parameters.

[0094] The training module evaluates the performance of the trained prediction model for early identification of metabolically unhealthy obesity, and selects the prediction model for early identification of metabolically unhealthy obesity that meets the preset requirements as the final prediction model, including: the training module obtains the trained prediction model for early identification of metabolically unhealthy obesity on R (4.3.1) software, and obtains the identification performance on the first set and the identification performance on the test set, respectively; if the identification performance of the trained prediction model for early identification of metabolically unhealthy obesity on the training set and the test set differs by less than a preset difference threshold, it means that the identification performances of the two are close, and the trained prediction model for early identification of metabolically unhealthy obesity is selected as the final prediction model.

[0095] Here, the training module uses the trained prediction model for early identification of metabolically unhealthy obesity to make predictions on the first set and the test set on R (4.3.1) software, respectively, to obtain the identification performance (identification performance) on the training set and the test set, respectively, including: the AUC values of the training set and the test set are 0.930 and 0.884, respectively, showing high accuracy and sensitivity, and the specific performance indicators are shown in Table 2; if the identification performance curves of the training set and the test set are close to each other (Delong test P = 0.26 > 0.05), as shown in FIG. 2, it means that the identification performances are close, and the trained prediction model for early identification of metabolically unhealthy obesity is a good model and can be used as the final prediction model.

[0096] The metabolic unhealthy obesity identification model provided by the application has the characteristics of less sample consumption and high reproducibility. By comparing the performance of the training set and the performance of the external test set, it is determined whether the model for identifying metabolic unhealthy obesity has the risk of overfitting. If the performance is close, it means that the model has universality, and if it is not close, it means that the generalization ability of the model is poor, and a better model needs to be trained. The performance of machine learning is shown in Table 3.

[0097] Table 3, performance of machine learning obtained by metabolic biomarker independent test set

[0098] A 2x2 confusion matrix is used to represent the model prediction results, where the row represents the actual label and the column represents the predicted label. The confusion matrix divides the samples into four categories, respectively:

[0099] TP (True Positive) True Positive: The number of samples that are actually positive and predicted to be positive.

[0100] TN (True Negative) True Negative: The number of samples that are actually negative and predicted to be negative.

[0101] FP (False Positive) False Positive: The number of samples that are actually negative and predicted to be positive.

[0102] FN (False Negative) False Negative: The number of samples that are actually positive and predicted to be negative.

[0103] Accuracy: The proportion of correctly predicted samples in the total samples, calculated as (TP+FP) / (FN+TN+TP+TN).

[0104] F1 value: The comprehensive index of the proportion of samples predicted to be positive in the actual positive samples and the proportion of correctly predicted samples in the positive samples, calculated as 2*P*R / (P+R), where P=TP / (TP+FP) and R=TP / (TP+FN).

[0105] Sensitivity: The proportion of samples predicted to be positive in the actual positive samples, calculated as TP / (TP+FN).

[0106] Specificity: The proportion of samples predicted to be negative in the actual negative samples, calculated as TN / (TN+FP).

[0107] PPV: Positive predictive value, i.e. the proportion of true positives in the population detected as positive, calculated as TP / (TP+FP).

[0108] NPV: negative predictive value, i.e. the proportion of people who are truly negative among those who are detected as negative, the calculation formula is TN / (TN+FN).

[0109] AUC: area under the ROC curve, the ROC curve is a curve composed of true positive rate and false positive rate at different thresholds, AUC is not affected by the proportion of positive and negative samples, which reflects the overall performance of the model at different thresholds,

[0110] The range is 0 to 1, the larger the better the overall performance of the model, which can be calculated by the following formula:

[0111] Plasma small molecule metabolic biomarkers and their relative concentration values, the selected plasma small molecule metabolic biomarkers and their relative concentration values of the to-be-detected person are input into the final prediction model to obtain the prediction result of whether the to-be-detected person belongs to metabolic unhealthy obesity.

[0112] Here, the acquisition method of the selected plasma small molecule metabolic biomarkers and their relative concentration values of the to-be-detected person can refer to the acquisition method in step 1.

[0113] In addition, the ability of the model to predict future cardiovascular risk can be evaluated by performing experiments: In the test set, the metabolic unhealthy obese (MUO) individuals identified by the final prediction model based on the aforementioned 27 metabolite biomarkers and the metabolic healthy obese (MHO) individuals identified by the model, China-PAR estimated the 10-year cardiovascular event rate according to risk factors, here, China-PAR is a tool recommended by the Joint Committee of the Guidelines for Cardiovascular Disease Risk Assessment and Management in China for estimating the 10-year cardiovascular event rate in the Guidelines for Cardiovascular Disease Risk Assessment and Management in China (2019), wherein the cardiovascular events include fatal or non-fatal atherosclerotic cardiovascular disease and stroke, and the calculation is performed under the instructions of the China-PAR online tool (www.cvdrisk.com.cn). In the test set, the average 10-year cardiovascular event rate of the MHO estimated by the metabolomics prediction of the final prediction model of the present application was 5.91, with a standard deviation of 1.66; the average 10-year cardiovascular event rate of the MUO estimated by the metabolomics prediction of the final prediction model of the present application was 8.11, with a standard deviation of 1.57, and the difference between the two groups was significant (Mann-Whitney test P<0.001), as shown in Figures 3a and 3b. In the test set, the MUO patients identified by the final prediction model of the present application were estimated for 10-year cardiovascular event risk according to China-PAR, and the MUO identified by the model was compared with the MHO identified by the model. Analysis showed that the final prediction model of the present application identified MUO patients with significantly increased 10-year cardiovascular event risk.

[0114] In addition, the ability of the model to identify potential metabolically unhealthy obese individuals was evaluated experimentally: in the individuals of the test set clinically evaluated as MHO, the individuals identified as metabolically unhealthy obese (MUO) and metabolically healthy obese (MHO) by the final prediction model based on the aforementioned 27 metabolite biomarkers were evaluated for 10-year cardiovascular event rates according to China-PAR, respectively. The mean of the estimated 10-year cardiovascular event rates for MHO predicted by metabolomics in the individuals of the test set clinically evaluated as MHO was 5.60, with a standard deviation of 1.56; the mean of the estimated 10-year cardiovascular event rates for MUO predicted by metabolomics through the final prediction model of the application was 7.58, with a standard deviation of 1.48, and the difference between the two groups was significant (Mann-Whitney test P<0.001). Clinically asymptomatic MHO patients can express similar metabolomic patterns to MUO patients, with a higher 10-year cardiovascular event risk, and thus be predicted as potential MUO individuals by the final prediction model of the application. As shown in Figures 3a and 3c. Through cardiovascular risk increase analysis, data analysis showed that in the MHO population without obvious clinical symptoms, the China-PAR estimated 10-year cardiovascular event risk of the population predicted as potential MUO by the final model of the application was significantly increased, suggesting the early identification value based on the metabolomics model.

[0115] Figures 3a to 3c are comparisons of 10-year cardiovascular event risks based on an independent test set in an embodiment of the application, Figure 3a shows the number of MHO and MUO individuals in the independent test set according to clinical symptoms / metabolomics prediction; Figure 3b shows a comparison of 10-year cardiovascular event risks of MHO and MUO patients predicted by metabolomics in the independent test set; and Figure 3c shows individuals clinically evaluated as MHO in the independent test set, some of which are predicted as MUO by the metabolomics model and have a higher 10-year cardiovascular risk, and thus should be considered as potential MUO patients.

[0116] The application further provides a computer readable storage medium, which stores computer executable instructions, wherein the computer executable instructions are executed by a processor to make the processor execute the following steps: step S1, a collection module collects blood plasma samples of each subject of metabolic healthy obesity and metabolic unhealthy obesity, and obtains plasma small molecule metabolic biomarkers and relative concentration values thereof from each blood plasma sample; step S2, a feature selection module performs feature selection based on the plasma small molecule metabolic biomarkers and relative concentration values thereof obtained from each blood plasma sample, to screen plasma small molecule metabolic biomarkers for early identification of metabolic unhealthy obesity as the plasma small molecule metabolic biomarkers after feature selection; step S3, a training module trains a final prediction model for early identification of metabolic unhealthy obesity based on the plasma small molecule metabolic biomarkers after feature selection and relative concentration values thereof in each blood plasma sample; step S4, an identification module obtains the plasma small molecule metabolic biomarkers after feature selection and relative concentration values thereof of a to-be-detected subject, and inputs the plasma small molecule metabolic biomarkers after feature selection and relative concentration values thereof of the to-be-detected subject into the final prediction model, to obtain a prediction result of whether the to-be-detected subject belongs to metabolic unhealthy obesity. The plasma small molecule metabolic biomarkers after feature selection include mercaptoethylamine, 4-methylpentanal, 3-hydroxybutyric acid, aminophenol, pyruvic acid, threonine, succinic anhydride, threitol, glycerol, hypotaurine, creatinine, adenine, valine, cysteine, isonicotinic acid, taurine, glutamic acid, methionine, citrulline, glucose, arginine, ascorbic acid, lipoic acid, acetylcarnitine, glycerophosphocholine, stearic acid and sphingosine.

[0117] The application further provides a detection kit for early identification of metabolic unhealthy obesity, which protects metabolite biomarkers, and the metabolite biomarkers at least include the following substances: mercaptoethylamine, 4-methylpentanal, 3-hydroxybutyric acid, aminophenol, pyruvic acid, threonine, succinic anhydride, threitol, glycerol, hypotaurine, creatinine, adenine, valine, cysteine, isonicotinic acid, taurine, glutamic acid, methionine, citrulline, glucose, arginine, ascorbic acid, lipoic acid, acetylcarnitine, glycerophosphocholine, stearic acid and sphingosine.

[0118] In summary, the identification model construction system for metabolically unhealthy obesity disclosed by the present application utilizes high-throughput metabolite detection technology and machine learning algorithms to provide an innovative and effective solution for the early diagnosis of metabolically unhealthy obesity (MUO). Metabolomics is an advanced omics technology that can perform global analysis of small metabolites in biological samples with high throughput. Metabolomics can systematically detect metabolic disorders in the body before the appearance of disease phenotypes, thus showing broad prospects in disease diagnosis. Plasma, as an easily accessible biological fluid, contains rich metabolite information that can reflect the dynamic changes of various metabolic pathways in the body. Plasma metabolites are considered early markers of diseases such as cardiovascular disease and glucose-lipid metabolic disorders. The present application selects specific metabolic markers related to metabolically unhealthy obesity by systematically analyzing plasma metabolites, thereby establishing an effective tool for early identification of MUO.

[0119] The present application exhibits excellent risk assessment ability. Through machine learning of plasma metabolite biomarkers, metabolically unhealthy obesity patients can be identified early when clinical symptoms have not yet appeared, helping doctors and patients to take intervention measures before cardiovascular risk increases. The identification model construction system for metabolically unhealthy obesity provided by the present application is based on small sampling of plasma metabolites, which is more convenient than traditional medical examination procedures, reduces the participation burden of patients, and is suitable for wide clinical application. Through the combination of machine learning models and high-throughput mass spectrometry technology, high sensitivity (87.18%) and specificity (71.79%) identification of MUO can be achieved, ensuring its reliability and accuracy in clinical application. The identification model construction system for metabolically unhealthy obesity provided by the present application provides a scientific basis for individualized health management, which helps to customize personalized intervention measures, especially in the prevention of cardiovascular disease.

[0120] The identification model construction system for metabolically unhealthy obesity of the present application is used for the detection and prevention of metabolically unhealthy obesity (MUO). By identifying MUO patients, even when no obvious clinical symptoms such as hypertension, elevated blood glucose, or dyslipidemia appear, the potential cardiovascular disease risk can be accurately judged. This system has excellent sensitivity, specificity, and accuracy, and the detection process is simple, fast, and suitable for large-scale clinical screening and intervention, bringing great benefits to the prevention and early intervention of cardiovascular disease. The present application can efficiently distinguish between metabolically healthy obesity (MHO) and metabolically unhealthy obesity (MUO) through metabolite data in plasma, and provides a scientific basis for individualized health management, especially in the early prevention of cardiovascular events.

[0121] In addition, the present application discloses a metabolite composition for early identification of metabolically unhealthy obesity (MUO), which can be applied to develop a kit to provide a convenient and rapid detection tool to support clinical diagnosis and preventive measures.

[0122] The detailed content of each device embodiment of the present application can be referred to the corresponding part of each system embodiment, which will not be described here.

[0123] Obviously, those skilled in the art can make various modifications and variations to the present application without departing from the spirit and scope of the present application. Thus, if these modifications and variations of the present application fall within the scope of the claims of the present application and their equivalents, the present application also intends to include these modifications and variations.

[0124] It should be noted that the present application can be implemented in software and / or a combination of software and hardware, for example, can be implemented by using an application specific integrated circuit (ASIC), a general purpose computer or any other similar hardware device. In one embodiment, the software program of the present application can be executed by a processor to implement the steps or functions described above. Similarly, the software program of the present application (including related data structures) can be stored in a computer readable recording medium, such as a RAM memory, a magnetic or optical drive or a soft disk and the like. In addition, some steps or functions of the present application can be implemented by hardware, for example, as a circuit cooperating with the processor to perform the steps or functions.

[0125] In addition, part of the present application can be applied as a computer program product, for example, computer program instructions, when executed by a computer, through the operation of the computer, the method and / or technical solutions according to the present application can be invoked or provided. The program instructions invoking the method of the present application can be stored in a fixed or removable recording medium, and / or transmitted through a data stream in a broadcast or other signal bearing medium, and / or stored in the working memory of the computer device running according to the program instructions. Here, according to one embodiment of the present application includes a device including a memory for storing computer program instructions and a processor for executing program instructions, wherein when the computer program instructions are executed by the processor, the device triggers the operation of the method and / or technical solutions based on the above-mentioned multiple embodiments according to the present application.

[0126] It will be obvious to a person skilled in the art that the application is not limited to the details of the above-described exemplary embodiments but can be implemented in other embodiments without departing from the scope of the application. The embodiments are to be considered in all respects as being illustrative and not restrictive, the scope of the application being indicated by the appended claims rather than by the above description, and all changes which come within the meaning and range of equivalents of the claims are therefore intended to be embraced therein. Any reference signs in the claims should not be construed as limiting the scope of the claims. The word "comprising" does not exclude other elements or steps not mentioned. The singular is not excluded from the plural. Multiple units or devices also can be presented by one unit or device, either by software or hardware. The words "first", "second", etc. are used to indicate names and not a specific order.

Claims

1. A metabolic unhealthy obesity identification model construction system, characterized in that, The method comprises the following steps: A collection module is configured to collect blood plasma samples of each subject with metabolically healthy obesity and metabolically unhealthy obesity, and obtain plasma small molecule metabolic biomarkers and their relative concentration values from each blood plasma sample; A feature selection module is configured to perform feature selection based on the plasma small molecule metabolic biomarkers and their relative concentration values obtained from each blood plasma sample, so as to screen out plasma small molecule metabolic biomarkers for early identification of metabolically unhealthy obesity as feature-selected plasma small molecule metabolic biomarkers; A training module is configured to train a final prediction model for early identification of metabolically unhealthy obesity based on the feature-selected plasma small molecule metabolic biomarkers and their relative concentration values in each blood plasma sample; An identification module is configured to obtain feature-selected plasma small molecule metabolic biomarkers and their relative concentration values of a to-be-detected subject, and input the feature-selected plasma small molecule metabolic biomarkers and their relative concentration values of the to-be-detected subject into the final prediction model to obtain a prediction result of whether the to-be-detected subject belongs to metabolically unhealthy obesity. The feature-selected plasma small molecule metabolic biomarkers comprise: Mercaptoethylamine, 4-methylpentanal, 3-hydroxybutyric acid, aminophenol, pyruvic acid, threonine, succinic anhydride, threitol, glycerol, hypotaurine, creatinine, adenine, valine, cysteine, isonicotinic acid, taurine, glutamic acid, methionine, citrulline, glucose, arginine, ascorbic acid, lipoic acid, acetylcarnitine, glycerophosphocholine, stearic acid and sphingosine.

2. The metabolic unhealthy obesity identification model construction system of claim 1, wherein, The collection module is configured to obtain a first subject group and a second subject group which are mutually exclusive, wherein the first subject group and the second subject group each comprise subjects with metabolically healthy obesity and metabolically unhealthy obesity which are mutually exclusive; The collection module is configured to collect first blood plasma samples of the first subject group, and obtain plasma small molecule metabolic biomarkers and their relative concentration values from each first blood plasma sample as a first set; The collection module is configured to collect second blood plasma samples of the second subject group, and obtain plasma small molecule metabolic biomarkers and their relative concentration values from each second blood plasma sample as a test set.

3. The metabolic unhealthy obesity identification model construction system of claim 2, wherein, The feature selection module is configured to perform feature selection on all plasma small molecule metabolic biomarkers in the first set by using a Lasso feature selection algorithm and five-fold cross-validation on R software, and the implementation of the Lasso feature selection algorithm depends on a glmnet package, wherein the glmnet package generates a sequence of 100 values corresponding to a hyperparameter lambda by an internal program, which is used to control the strength of L1 regularization on model coefficient penalty, and determines lambda.min when the misclassification rate is the lowest in five-fold cross-validation, and finally calculates lambda.min, and the metabolite biomarkers are screened out based on lambda.min.

4. The metabolic unhealthy obesity identification model construction system of claim 3, wherein, The feature selection module is configured to divide the first set into five non-overlapping data, each time taking one of the five non-overlapping data that has not been selected as an internal validation set, and taking the remaining four data as a training set each time; wherein the internal validation set and the training set each include subjects of metabolic healthy obesity and metabolic unhealthy obesity that are non-overlapping; The feature selection module is configured to, based on the training set each time, use a Lasso feature selection algorithm on R software to cyclically train the feature selection model for five rounds to obtain five corresponding feature selection models; The feature selection module is configured to, for each value in a sequence of 100 values corresponding to the hyperparameter lambda, input a corresponding one of the five feature selection models to the internal validation set of each round to obtain one misclassification rate corresponding to each value on each internal validation set; obtain one average misclassification rate corresponding to each value from the five misclassification rates corresponding to each value; select lambda.min corresponding to the lowest misclassification rate based on the smallest average misclassification rate among all values, wherein lambda is one of the hyperparameters; and obtain the corresponding metabolite marker based on lambda.min corresponding to the lowest misclassification rate as the plasma small molecule metabolic biomarker after feature selection.

5. The metabolic unhealthy obese identification model construction system of claim 2, wherein, The training module is configured to train a prediction model for early identification of metabolic unhealthy obesity based on the plasma small molecule metabolic biomarker after feature selection and the relative concentration value thereof in each plasma sample; and perform performance evaluation on the trained prediction model for early identification of metabolic unhealthy obesity, and take the prediction model for early identification of metabolic unhealthy obesity that meets the preset requirements of the performance evaluation as the final prediction model.

6. The metabolic unhealthy obese identification model construction system of claim 5, wherein, The training module is configured to train the prediction model for early identification of metabolic unhealthy obesity based on the first set using an elastic network algorithm on R software, and the implementation of the elastic network algorithm depends on the glmnet package, wherein the ElasticNet algorithm hyperparameter alpha used to control the ratio of L1 regularization and L2 regularization is set to 0.5; and s used to control the penalty degree of the parameters is set to 0.

01.

7. The metabolic unhealthy obesity identification model construction system of claim 6, wherein, The training module is configured to obtain the identification performance of the trained prediction model for early identification of metabolic unhealthy obesity on the first set and the identification performance on a test set on R software; and take the trained prediction model for early identification of metabolic unhealthy obesity as the final prediction model if the identification performances of the trained prediction model for early identification of metabolic unhealthy obesity on the training set and the test set differ by less than a preset difference threshold.

8. A computer-readable storage medium having stored thereon computer- executable instructions, wherein, The computer executable instructions, when executed by the processor, cause the processor to perform the following steps: The collection module collects plasma samples of subjects of metabolic healthy obesity and metabolic unhealthy obesity, and obtains plasma small molecule metabolic biomarkers and relative concentration values thereof from each plasma sample; The collection module collects plasma samples of subjects of metabolic healthy obesity and metabolic unhealthy obesity, and obtains plasma small molecule metabolic biomarkers and relative concentration values thereof from each plasma sample; The feature selection module performs feature selection based on the plasma small molecule metabolic biomarkers and their relative concentration values in each plasma sample, to screen the plasma small molecule metabolic biomarkers for early identification of metabolically unhealthy obesity, as the feature-selected plasma small molecule metabolic biomarkers; The training module trains the final prediction model for early identification of metabolically unhealthy obesity based on the feature-selected plasma small molecule metabolic biomarkers and their relative concentration values in each plasma sample; The identification module obtains the feature-selected plasma small molecule metabolic biomarkers and their relative concentration values of the to-be-detected subject, and inputs the feature-selected plasma small molecule metabolic biomarkers and their relative concentration values of the to-be-detected subject into the final prediction model, to obtain the prediction result of whether the to-be-detected subject belongs to metabolically unhealthy obesity. The feature-selected plasma small molecule metabolic biomarkers include: mercaptoethylamine, 4-methylpentanal, 3-hydroxybutyric acid, aminophenol, pyruvic acid, threonine, succinic anhydride, threitol, glycerol, hypotaurine, creatinine, adenine, valine, cysteine, isonicotinic acid, taurine, glutamic acid, methionine, citrulline, glucose, arginine, ascorbic acid, lipoic acid, acetylcarnitine, glycerophosphocholine, stearic acid and sphingosine.

9. A test kit for early identification of metabolically unhealthy obesity, characterized by, The detection kit comprises metabolite biomarkers, which at least include: mercaptoethylamine, 4-methylpentanal, 3-hydroxybutyric acid, aminophenol, pyruvic acid, threonine, succinic anhydride, threitol, glycerol, hypotaurine, creatinine, adenine, valine, cysteine, isonicotinic acid, taurine, glutamic acid, methionine, citrulline, glucose, arginine, ascorbic acid, lipoic acid, acetylcarnitine, glycerophosphocholine, stearic acid and sphingosine; The metabolite biomarkers are subjected to feature selection and identification by the following modules: The collection module is used to collect plasma samples of each subject with metabolically healthy obesity and metabolically unhealthy obesity, and obtain plasma small molecule metabolic biomarkers and their relative concentration values from each plasma sample; The feature selection module is used to perform feature selection based on the plasma small molecule metabolic biomarkers and their relative concentration values in each plasma sample, to screen the plasma small molecule metabolic biomarkers for early identification of metabolically unhealthy obesity, as the feature-selected plasma small molecule metabolic biomarkers; The training module is used to train the final prediction model for early identification of metabolically unhealthy obesity based on the feature-selected plasma small molecule metabolic biomarkers and their relative concentration values in each plasma sample; The identification module is used to obtain the feature-selected plasma small molecule metabolic biomarkers and their relative concentration values of the to-be-detected subject, and input the feature-selected plasma small molecule metabolic biomarkers and their relative concentration values of the to-be-detected subject into the final prediction model, to obtain the prediction result of whether the to-be-detected subject belongs to metabolically unhealthy obesity.

Images