Metabolic marker combination for assisting in staging of CKD with diabetes and applications, methods

Abstract

The application discloses a metabolic marker combination and application and a method for assisting in CKD staging accompanied by diabetes, and belongs to the field of medical treatment. The metabolic marker combination comprises eicosapentaenoic acid ethyl ester, tetradecenoyl carnitine, chlorocruorin, guanidyl succinate, lycoramine and 3-palmitoyl-sn-glycerol for assisting in the staging of CKD 1-2 stages and CKD 3-4 stages accompanied by diabetes; indolelactic acid, acetylcarnitine, L-tryptophan, L-leucine, guanidyl succinate, lycoramine and L-octanoylcarnitine for assisting in the staging of CKD 1-2 stages and CKD 5 stage accompanied by diabetes; and acetylcarnitine, n-butylbenzenesulfonamide, L-leucine, 5-hydroxyindole-3-acetic acid and hesperetin for assisting in the staging of CKD 3-4 stages and CKD 5 stage accompanied by diabetes. The metabolic marker combination and the constructed model improve the accuracy of CKD staging accompanied by diabetes, and sample acquisition is simple.

Description

Technical Field

[0001] This invention belongs to the field of disease staging and identification technology, specifically relating to a combination and application of metabolic biomarkers for assisting in the staging of CKD with diabetes, and a method thereof, namely, a combination of metabolic biomarkers for the identification of CKD with diabetes and its application, and a method for model construction. Background Technology

[0002] Chronic kidney disease (CKD) is a chronic progressive disease characterized by the gradual loss of kidney function, affecting approximately 10% of the global population. In recent years, its incidence has been steadily increasing due to the rising prevalence of metabolic diseases such as diabetes and hypertension. Early CKD is typically insidious, leading to low early diagnosis rates and a risk of progressing to end-stage renal disease. Currently, clinical diagnosis and staging of CKD mainly rely on serum creatinine (SCr), urine albumin / creatinine ratio (ACR), estimated glomerular filtration rate (eGFR), and renal biopsy. However, these indicators have relatively low sensitivity in early diagnosis, CKD subtyping, and staging. Renal biopsy, while the gold standard for pathological diagnosis, is an invasive procedure and cannot be used as a universally applicable method for routine screening.

[0003] Metabolic disorders are a typical pathophysiological feature of CKD development, reflecting systemic changes in waste excretion, energy homeostasis, and endocrine function. Amino acids, sugars, and fatty acids are common metabolic products in the human body, widely present in bodily fluids such as blood and urine. Regarding metabolic analysis platforms, traditional liquid chromatography-mass spectrometry (LC-MS) and nuclear magnetic resonance (NMR) techniques require cumbersome chromatographic separation, long pretreatment times, and complex spectral analysis. Matrix-assisted laser desorption / ionization time-of-flight mass spectrometry (MALDI-TOF MS), as a soft ionization biomolecular mass spectrometry technique, offers advantages such as simple sample preparation, fast analysis speed, and high detection sensitivity. In recent years, nano-enhanced MALDI-TOF MS technology has selectively captured small-molecule metabolites from complex biological samples through structural advantages, obtaining high-throughput metabolic fingerprints and providing a powerful alternative to traditional metabolic analysis. Therefore, using enhanced MALDI-TOF MS technology to analyze serum and urine samples of CKD and establish a highly efficient combination of metabolic biomarkers for typing and staging has significant potential.

[0004] However, there are currently no reports on two-fluid metabolic fingerprinting of serum and urine for CKD. Therefore, this study aims to construct a CKD typing and staging model based on nano-enhanced MALDI-TOF MS technology, screen combinations of biological metabolites in two-fluid samples, promote accurate typing and staging of clinical CKD, and ultimately improve the early diagnosis rate of CKD.

[0005] It should be noted that the information disclosed in the background section above is only used to enhance the understanding of the background of the present invention, and therefore may include information that does not constitute prior art known to those skilled in the art. Summary of the Invention

[0006] To address the aforementioned problems in the existing technology, this invention provides a combination and application of metabolic biomarkers for assisting in the staging of CKD associated with diabetes, as well as a method thereof. The technical problem to be solved by this invention is achieved through the following technical solution: In a first aspect, the present invention provides a combination of metabolic markers for assisting in the staging of CKD with diabetes, the combination of metabolic markers comprising a first metabolite combination, a second metabolite combination, and a third metabolite combination; The first metabolite combination comprises the serum metabolites eicosapentaenoic acid ethyl ester, tetradecenecarnitine, biliverdin, and the urinary metabolites guanidinosuccinate, lycopodium, and 3-palmitoyl-sn-glycerol, for auxiliary staging of CKD stages 1-2 and CKD stages 3-4 with diabetes. The second metabolite combination includes serum metabolites indolelactate, acetylcarnitine, and L-tryptophan, and urinary metabolites L-leucine, guanidinosuccinate, lycopodium, and L-octanoylcarnitine, for auxiliary staging of patients with CKD stage 1-2 and CKD stage 5 diabetes. The third metabolite combination includes the serum metabolites acetylcarnitine and n-butylbenzenesulfonamide, and the urinary metabolites L-leucine, 5-hydroxyindole-3-acetic acid, and hesperidin, and is used as an adjunct to staging of CKD stages 3-4 and CKD stage 5 in patients with diabetes.

[0007] Secondly, the present invention provides the application of a reagent for detecting the above-mentioned combination of metabolic markers in the preparation of a kit to assist in the staging of CKD with diabetes.

[0008] Thirdly, the present invention provides a method for constructing a model to assist in the staging of CKD with diabetes, comprising the following steps: S1. Obtain in vitro serum and urine samples from patients with CKD stages 1-2, 3-4, and 5 of diabetes. S2. Multidimensional carbon-based composite material reinforced matrix-assisted laser desorption / ionization time-of-flight mass spectrometry was used to detect each of the serum and urine samples to obtain serum metabolic fingerprint peaks and urine metabolic fingerprint peaks. S3. The MALDIquant package was used to perform baseline correction, peak extraction, spectral smoothing, and normalization preprocessing on the serum metabolic fingerprint peaks and the urine metabolic fingerprint peaks to obtain the corresponding peaks for each serum sample. m / zFeatured serum metabolic profiles and corresponding data for each urine sample m / z A characteristic set of urinary metabolic profiles; S4. Based on the serum metabolic profile set and the urine metabolic profile set, screen the above-mentioned combinations of metabolic biomarkers; S5. Based on the combination of metabolic biomarkers, a machine learning algorithm is used to construct a first staging model for CKD stages 1-2 and CKD stages 3-4 with diabetes, a second staging model for CKD stages 1-2 and CKD stages 5 with diabetes, and a third staging model for CKD stages 3-4 and CKD stages 5 with diabetes.

[0009] In one embodiment of the present invention, step S2 includes: A multidimensional carbon-based composite material was obtained, and the multidimensional carbon-based composite material was dissolved in an ethanol solution to obtain a matrix solution; The matrix solution is dropped onto the target plate, dried, and then a serum or urine sample is added. After drying again, a co-crystallized layer is formed. In positive ion mode, matrix-assisted laser desorption / ionization time-of-flight mass spectrometry was used to obtain serum metabolic fingerprint peaks and urine metabolic fingerprint peaks, respectively.

[0010] In one embodiment of the present invention, the multidimensional carbon-based composite material is one of one-dimensional carbon nanotubes, two-dimensional graphene, and three-dimensional carbon nanospheres.

[0011] In one embodiment of the present invention, step S4 includes: processing the serum metabolic profile set and the urine metabolic profile set... m / z Perform a t-test on the features, according to p The combination of metabolic biomarkers was selected if the value was <0.05 and the score was >0.7. The screening of the metabolic biomarker combination includes: screening a first metabolite combination for auxiliary staging of patients with CKD stages 1-2 and CKD stages 3-4, a second metabolite combination for auxiliary staging of patients with CKD stages 1-2 and CKD stages 5, and a third metabolite combination for auxiliary staging of patients with CKD stages 3-4 and CKD stages 5. The first metabolite group includes a first serum metabolite group and a first urine metabolite group; the first serum metabolite group includes ethyl eicosapentaenoate, tetradecylcarnitine, and biliverdin, and the first urine metabolite group includes guanidinosuccinate, lycopodium, and 3-palmitoyl-sn-glycerol. The second metabolite group comprises a second serum metabolite group and a second urine metabolite group; the second serum metabolite group comprises indolelacic acid, acetylcarnitine, and L-tryptophan, and the second urine metabolite group comprises L-leucine, guanidinosuccinate, lycophylline, and L-octanoylcarnitine; The third metabolite group includes a third serum metabolite group and a third urine metabolite group; the third serum metabolite group includes acetylcarnitine and n-butylbenzenesulfonamide, and the third urine metabolite group includes L-leucine, 5-hydroxyindole-3-acetic acid, and hesperidin.

[0012] In one embodiment of the present invention, step S5, the construction process of the first staging identification model includes: Obtain the first dataset, which consists of samples with CKD stages 1-2 and CKD stages 3-4, and analyze the metabolite markers in the first serum metabolite group and the first urine metabolite group. m / z The mass spectrum response signal intensity at the location; 75% to 85% of the first dataset is used as the first training set, and the remainder is used as the first validation set; Based on the first training set, a first stage identification model is constructed using Lasso regression machine learning, and the first validation set is used to validate the first stage identification model.

[0013] In one embodiment of the present invention, step S5, the construction process of the second staging identification model includes: Obtain a second dataset consisting of samples with stage 1-2 diabetes mellitus and stage 5 diabetes mellitus. Metabolic markers were analyzed in the second serum metabolite group and the second urine metabolite group. m / z The mass spectrometry response signal intensity at the location; 75% to 85% of the second dataset is used as the second training set, and the remainder is used as the second validation set; Based on the second training set, a second staging identification model is constructed using Lasso regression machine learning, and the second validation set is used to validate the second staging identification model.

[0014] In one embodiment of the present invention, step S5, the construction process of the third staging identification model includes: A third dataset was obtained, consisting of samples with CKD stages 3-4 and CKD stage 5 diabetes. Metabolic markers were analyzed in the third serum metabolite group and the third urinary metabolite group. m / z The mass spectrometry response signal intensity at the location; 75% to 85% of the third dataset is used as the third training set, and the remainder is used as the third validation set; Based on the third training set, a third staging identification model is constructed using Lasso regression machine learning, and the third validation set is used to validate the third staging identification model.

[0015] Compared with the prior art, the beneficial effects of the present invention are as follows: 1. This invention uses MALDI-TOF MS technology enhanced with multidimensional carbon-based composite materials to obtain serum and urinary metabolic fingerprints of CKD patients with diabetes mellitus. It screens combinations of metabolic biomarkers for CKD stages 1-2, 3-4, 5, 3-4, and 5 with diabetes mellitus. The Lasso regression machine learning method is used to train the model on the training set and validate it on the test set to obtain a staging identification model.

[0016] 2. The CKD subtyping and diagnostic model constructed using the Lasso regression algorithm in this invention achieves an AUC value of 1.00. Furthermore, this invention further constructs a combined serum and urine two-fluid metabolic marker model for differentiating CKD combined with diabetes staging (lowest AUC = 0.736, highest AUC = 1.00), which has higher AUC values ​​than serum alone (lowest AUC = 0.43, highest AUC = 0.97) or urine alone (lowest AUC = 0.63, highest AUC = 0.97), providing a reliable technical solution for achieving accurate subtyping and staging differentiation of clinical CKD.

[0017] 3. This invention uses a multidimensional carbon-based composite material as the inorganic nanomatrix for a novel MALDI-TOF MS. This material has a curvature structure with different dimensions and abundant C, O, and N heteroatoms, and has a strong built-in electric field and photothermal conversion efficiency. In positive ion mode, it can efficiently ionize metabolic standards such as amino acids and glucose, and realize high-throughput (384 samples per chip) and rapid (20 s / sample) analysis of metabolites in serum and urine.

[0018] The present invention will be further described in detail below with reference to the accompanying drawings and embodiments. Attached Figure Description

[0019] Figure 1 This is a serum metabolic fingerprint characterization diagram of CKD with and without diabetes in Example 1 of the present invention. Figure 2 This is an example of training multiple machine learning models for CKD with and without diabetes using embodiments of the present invention. Figure 3This is a box plot comparing the distribution of six serum and urine metabolic markers in CKD stages 1-2 and CKD stages 3-4 with diabetes in Example 2 of the present invention. Figure 4 This is a box plot comparing the distribution of seven serum and urine metabolic markers in CKD stages 1-2 and CKD stage 5 with diabetes in Example 2 of the present invention. Figure 5 This is a box plot comparing the distribution of five serum and urine metabolic markers in CKD stages 3-4 and CKD stage 5 with diabetes in Example 2 of the present invention. Figure 6 The receiver operating characteristic (ROC) curves of the combination of individual serum metabolic markers in Example 3 of the present invention in the training set and the test set; Figure 7 The figure shows the receiver operating characteristic (ROC) curves of the combination of individual urinary metabolic markers in the training and test sets of Embodiment 4 of the present invention. Detailed Implementation

[0020] To further illustrate the technical means and effects adopted by the present invention to achieve its intended purpose, the following detailed description, in conjunction with the accompanying drawings and specific embodiments, provides a detailed explanation of a combination and application of metabolic markers for assisting in the staging of CKD associated with diabetes, based on the present invention. The foregoing and other technical contents, features, and effects of the present invention are clearly presented in the following detailed description of the specific embodiments in conjunction with the accompanying drawings. Through the description of the specific embodiments, a more in-depth and specific understanding of the technical means and effects adopted by the present invention to achieve its intended purpose can be obtained. However, the accompanying drawings are for reference and illustration only and are not intended to limit the technical solution of the present invention.

[0021] It should be noted that, in this document, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of indicated technical features. Thus, a feature defined as "first" or "second" may explicitly or implicitly include one or more of that feature. In the description of this invention, "a plurality of" means two or more, unless otherwise explicitly specified. Furthermore, the terms "comprising," "including," or any other variations are intended to cover non-exclusive inclusion, such that an article or apparatus comprising a list of elements includes not only those elements but also other elements not explicitly listed.

[0022] In this article, m / z "Feature" refers to a signal spike with identifiable intensity and shape that appears at a specific mass-to-charge ratio position, representing the presence of a specific molecule (usually an ion) in the sample being tested. Example 1 This embodiment uses matrix-assisted laser desorption / ionization time-of-flight mass spectrometry (MALDI-TOF MS) technology to collect serum metabolic fingerprints of CKD patients without diabetes (105 cases) and CKD with diabetes (92 cases). The fingerprints are then trained using various machine learning models to construct a classification model for CKD with and without diabetes.

[0023] The instruments and reagents used included: matrix-assisted laser desorption / ionization time-of-flight mass spectrometer, serum samples, matrices (one-dimensional carbon nanotubes, two-dimensional graphene, three-dimensional carbon nanospheres), water, and ethanol solvent.

[0024] Specifically, the following steps are included: Step 11: Dilute the original serum sample 5 times to obtain the serum sample to be tested.

[0025] Step 12: Prepare a matrix solution of 1 mg / mL using deionized water / ethanol (V / V=1 / 1) for the multidimensional carbon-based composite material.

[0026] Step 13: Take 1 μL of the above matrix solution and drop it onto a 384 stainless steel target plate. After the matrix solution dries at room temperature (15-30 °C), continue to drop 1 μL of the serum sample to be tested and dry it to form a uniform co-crystallized layer.

[0027] Step 14: Acquire serum metabolic fingerprints using a Bruker Tims TOF-flex matrix-assisted laser desorption / ionization time-of-flight mass spectrometer. For example... Figure 1 As shown, an exemplary illustration is provided of the characterization of serum metabolic fingerprint profiles for patients with and without diabetes-related CKD.

[0028] Step 15: Use the MALDIquant package to perform baseline correction, peak extraction, spectral smoothing, and normalization on the above serum metabolic fingerprint peaks. m / z Data processing.

[0029] Step 16: On RStudio (version 4.5.2), process the 119 extracted files mentioned above. m / z The features were trained using six machine learning algorithms: Lasso regression, random forest, Naive Bayes, decision tree, K-nearest neighbor, and logistic regression. Step 17: Evaluate each model using the ten-fold cross-validation method to screen diagnostic models for CKD without diabetes and CKD with diabetes.

[0030] like Figure 2 As shown, the results indicate that among the six machine learning algorithms, the Lasso regression model and the random forest model both achieved an AUC value of 1.00, demonstrating the highest diagnostic performance.

[0031] Example 2 Metabolic biomarker combinations for auxiliary staging of CKD stages 1-2 and 3-4 with diabetes were screened separately. These combinations were used for auxiliary staging of CKD stages 1-2 and 5 with diabetes, and for auxiliary staging of CKD stages 3-4 and 5 with diabetes. The specific steps included: Step 21: In the group with diabetes mellitus and CKD, 1 μL of in vitro serum and urine samples were obtained from 72 patients with known diabetes mellitus and CKD stage 1-2, 75 patients with diabetes mellitus and CKD stage 3-4, and 75 patients with diabetes mellitus and CKD stage 5. Each serum and urine sample was diluted 5 times and processed according to the procedure in step 13 of Example 1.

[0032] Step 22: Serum and urine metabolic fingerprint peaks for patients with CKD stages 1-2, 3-4, and 5 of diabetes were detected and acquired using a MALDI-TOF MS instrument. MALDIquant data preprocessing was then used to identify the corresponding peaks in the serum and urine samples. m / z Metabolic profiles were ultimately obtained from serum and urine two-fluid metabolic profiles, containing 365 features. m / z A metabolic atlas.

[0033] Step 23: Analyze the 365 features of each sample obtained in Step 22. m / z As a dataset, t-tests were performed on each pair of groups with CKD stage 1-2 diabetes and CKD stage 3-4 diabetes, CKD stage 1-2 diabetes and CKD stage 5 diabetes, and CKD stage 3-4 diabetes and CKD stage 5 diabetes to screen for corresponding combinations of metabolic biomarkers.

[0034] Specifically, it includes: Step 231, press p Six differential statistical markers were identified between patients with CKD stages 1-2 and CKD stages 3-4 with diabetes, based on scores <0.05 and >0.76. m / z(Differential characteristic peaks) were used to compare and identify specific combinations of metabolic biomarkers using the HMDB library and ultra-high performance liquid chromatography-mass spectrometry. The results are shown in Table 1, including three serum metabolites: ethyl eicosapentaenoic acid ester, tetradecenylcarnitine, and biliverdin; and three urinary metabolites: guanidinosuccinate, lycoperodine, and 3-palmitoyl-sn-glycerol.

[0035] Table 1. Metabolic markers of CKD stages 1-2 and CKD stages 3-4 with diabetes.

[0036] Note: In the table, "S" represents serum sample and "U" represents urine sample.

[0037] like Figure 3 The image shows a box plot comparing the distribution of six serum and urinary metabolic markers in CKD stages 1-2 and CKD stages 3-4 with diabetes. Combined with... Figure 3 As shown in Table 1, serum eicosapentaenoic acid ethyl ester levels were significantly lower in CKD stages 3-4 with diabetes than in CKD stages 1-2 with diabetes, while urinary 3-palmitoyl-sn-glycerol levels were significantly higher in CKD stages 3-4 with diabetes.

[0038] Step 232, according to p Screening for CKD stage 1-2 with diabetes and CKD stage 5 with diabetes using values ​​<0.05 and score>0.76 yielded 7 differences. m / z (Differential characteristic peaks) were used to compare and identify specific combinations of metabolic biomarkers using the HMDB library and ultra-high performance liquid chromatography-mass spectrometry. The results are shown in Table 2, including three serum metabolites: indolelactic acid, acetylcarnitine, and L-tryptophan; and four urinary metabolites: L-leucine, guanidinosuccinate, lycoperodine, and L-octanoylcarnitine.

[0039] Table 2. Metabolic markers of CKD stages 1-2 and CKD stage 5 with diabetes.

[0040] Figure 4 Box plots comparing the distribution of seven serum and urinary metabolic markers in patients with stage 1-2 and stage 5 diabetes mellitus. Figure 4 As shown in Table 2, serum acetylcarnitine and urinary lycoperodine were downregulated in CKD stage 5 with diabetes mellitus, while serum indolelactic acid was upregulated in CKD stage 5 with diabetes mellitus.

[0041] Step 233, according to p Five differences were identified between patients with stage 3-4 CKD and stage 5 CKD with diabetes, based on scores <0.05 and >0.76. m / z (Differential characteristic peaks) were used to compare and identify specific combinations of metabolic biomarkers using the HMDB library and ultra-high performance liquid chromatography-mass spectrometry. The results are shown in Table 3, including two serum metabolites: acetylcarnitine and N-butylbenzenesulphonamide; and three urinary metabolites: L-leucine, 5-hydroxyyindole-3-acetic acid, and hesperetin (Table 3).

[0042] Table 3. Metabolic markers of CKD stages 3-4 and CKD stage 5 with diabetes.

[0043] like Figure 5 The image shows a box plot comparing the distribution of five serum and urinary metabolic markers in patients with CKD stages 3-4 and CKD stage 5. Figure 5 As shown in Table 3, serum acetylcarnitine and urine hesperetin were significantly upregulated in CKD stages 3-4 with diabetes.

[0044] It should be noted that the order of steps 231-233 is not fixed.

[0045] Step 24: Construct a staged identification model, which specifically includes: Step 241: Based on the selected combination of 6 characteristic metabolic biomarkers (shown in Table 1), 80% of the 57 samples with CKD stage 1-2 diabetes were randomly selected as the training set, and the remaining 20% ​​as the test set; 80% of the 60 samples with CKD stage 3-4 diabetes were randomly selected as the training set, and the remaining 20% ​​as the test set. Lasso regression machine learning was used to construct staging identification models for CKD stage 1-2 and CKD stage 3-4 diabetes; this staging identification model was defined as the first staging identification model, and its diagnostic efficacy was validated using the test set.

[0046] Step 242: Based on the selected combination of 7 characteristic metabolic biomarkers (shown in Table 2), 80% of the 57 samples with CKD stage 1-2 diabetes were randomly selected as the training set, and the remaining 20% ​​as the test set; 80% of the 60 samples with CKD stage 5 diabetes were randomly selected as the training set, and the remaining 20% ​​as the test set. Lasso regression machine learning was used to construct staging identification models for CKD stage 1-2 and CKD stage 5 diabetes; this staging identification model was defined as the second staging identification model, and its diagnostic efficacy was validated using the test set.

[0047] Step 243: Based on the selected combination of six characteristic metabolic biomarkers (shown in Table 3), 80% of the 60 samples with CKD stage 3-4 diabetes were randomly selected as the training set, and the remaining 20% ​​as the test set; 80% of the 60 samples with CKD stage 5 diabetes were randomly selected as the training set, and the remaining 20% ​​as the test set. Lasso regression machine learning was used to construct staging identification models for CKD stages 3-4 and CKD stage 5 diabetes; this staging identification model was defined as the third staging identification model, and its diagnostic efficacy was validated using the test set.

[0048] As shown in Table 4, the detailed diagnostic performance of the first, second, and third stage identification models based on Lasso regression machine learning was evaluated using the training and test sets, respectively.

[0049] Table 4. Detailed performance of LASSO regression learning in different staging identification.

[0050] Note: AUC is the area under the curve, CA is the accuracy rate, and F1 is the F1 score.

[0051] Example 3 A staging model was constructed based on the Lasso regression machine learning algorithm using a combination of three serum metabolic markers (eicosapentaenoic acid ethyl ester, tetradecylcarnitine, and biliverdin) for patients with CKD stages 1-2 and 3-4 with diabetes. (See [link to relevant documentation]). Figure 6 The staging identification model is defined as the fourth staging identification model. The diagnostic efficacy of the constructed fourth staging identification model is verified by the test set. The AUC value obtained in the test set is 0.947, which shows high identification and prediction performance.

[0052] Example 4 A staging model was constructed based on the Lasso regression machine learning algorithm using a combination of three individual urinary metabolic markers (guanidinosuccinate, lycopodium, and 3-palmitoyl-sn-glycerol) for patients with CKD stages 1-2 and 3-4 diabetes. (See [link to relevant documentation]). Figure 7 The staging identification model is defined as the fifth staging identification model. The diagnostic efficacy of the constructed fifth staging identification model is verified by the test set. The AUC value obtained in the test set is 0.864, which also shows good identification and prediction performance.

[0053] The above description, in conjunction with specific preferred embodiments, provides a further detailed explanation of the present invention. It should not be construed that the specific implementation of the present invention is limited to these descriptions. For those skilled in the art, various simple deductions or substitutions can be made without departing from the concept of the present invention, and all such modifications and substitutions should be considered within the scope of protection of the present invention.

Claims

1. A combination of metabolic biomarkers for assisting in the staging of CKD associated with diabetes, characterized in that, The combination of metabolic markers includes a first metabolite combination, a second metabolite combination, and a third metabolite combination; The first metabolite combination comprises the serum metabolites eicosapentaenoic acid ethyl ester, tetradecenecarnitine, biliverdin, and the urinary metabolites guanidinosuccinate, lycopodium, and 3-palmitoyl-sn-glycerol, for auxiliary staging of CKD stages 1-2 and CKD stages 3-4 with diabetes. The second metabolite combination includes serum metabolites indolelactate, acetylcarnitine, and L-tryptophan, and urinary metabolites L-leucine, guanidinosuccinate, lycopodium, and L-octanoylcarnitine, for auxiliary staging of patients with CKD stage 1-2 and CKD stage 5 diabetes. The third metabolite combination includes the serum metabolites acetylcarnitine and n-butylbenzenesulfonamide, and the urinary metabolites L-leucine, 5-hydroxyindole-3-acetic acid, and hesperidin, and is used as an adjunct to staging of CKD stages 3-4 and CKD stage 5 in patients with diabetes.

2. The use of the reagent for detecting the combination of metabolic markers as described in claim 1 in the preparation of a kit to assist in the staging of CKD with diabetes.

3. A method for constructing a model to assist in the staging of CKD with diabetes, characterized in that, Includes the following steps: S1. Obtain in vitro serum and urine samples from patients with CKD stages 1-2, 3-4, and 5 of diabetes. S2. Multidimensional carbon-based composite material reinforced matrix-assisted laser desorption / ionization time-of-flight mass spectrometry was used to detect each of the serum and urine samples to obtain serum metabolic fingerprint peaks and urine metabolic fingerprint peaks. S3. The MALDIquant package was used to perform baseline correction, peak extraction, spectral smoothing, and normalization preprocessing on the serum metabolic fingerprint peaks and the urine metabolic fingerprint peaks to obtain the corresponding peaks for each serum sample. m / z Featured serum metabolic profiles and corresponding data for each urine sample m / z A characteristic set of urinary metabolic profiles; S4. Based on the serum metabolic profile set and the urine metabolic profile set, screen to obtain the combination of metabolic biomarkers described in claim 1; S5. Based on the combination of metabolic biomarkers, a machine learning algorithm is used to construct a first staging model for CKD stages 1-2 and CKD stages 3-4 with diabetes, a second staging model for CKD stages 1-2 and CKD stages 5 with diabetes, and a third staging model for CKD stages 3-4 and CKD stages 5 with diabetes.

4. The method for constructing a model to assist in the staging of CKD with diabetes as described in claim 3, characterized in that, Step S2 includes: A multidimensional carbon-based composite material was obtained, and the multidimensional carbon-based composite material was dissolved in an ethanol solution to obtain a matrix solution; The matrix solution is dropped onto the target plate, dried, and then a serum or urine sample is added. After drying again, a co-crystallized layer is formed. In positive ion mode, matrix-assisted laser desorption / ionization time-of-flight mass spectrometry was used to obtain serum metabolic fingerprint peaks and urine metabolic fingerprint peaks, respectively.

5. The method for constructing a model to assist in staging CKD with diabetes as described in claim 4, characterized in that, The multidimensional carbon-based composite material is one of one-dimensional carbon nanotubes, two-dimensional graphene, and three-dimensional carbon nanospheres.

6. The method for constructing a model to assist in staging CKD with diabetes as described in claim 3, characterized in that, Step S4 includes: processing the serum metabolic profile set and the urine metabolic profile set... m / z Perform a t-test on the features, according to p If the value is < 0.05 and the score is > 0.7, the combination of metabolic biomarkers is selected. The screening of the metabolic biomarker combination includes: screening a first metabolite combination for auxiliary staging of patients with CKD stages 1-2 and CKD stages 3-4, a second metabolite combination for auxiliary staging of patients with CKD stages 1-2 and CKD stages 5, and a third metabolite combination for auxiliary staging of patients with CKD stages 3-4 and CKD stages 5. The first metabolite group includes a first serum metabolite group and a first urine metabolite group; the first serum metabolite group includes ethyl eicosapentaenoate, tetradecylcarnitine, and biliverdin, and the first urine metabolite group includes guanidinosuccinate, lycopodium, and 3-palmitoyl-sn-glycerol. The second metabolite group comprises a second serum metabolite group and a second urine metabolite group; the second serum metabolite group comprises indolelacic acid, acetylcarnitine, and L-tryptophan, and the second urine metabolite group comprises L-leucine, guanidinosuccinate, lycophylline, and L-octanoylcarnitine; The third metabolite group includes a third serum metabolite group and a third urine metabolite group; the third serum metabolite group includes acetylcarnitine and n-butylbenzenesulfonamide, and the third urine metabolite group includes L-leucine, 5-hydroxyindole-3-acetic acid, and hesperidin.

7. The method for constructing a model to assist in staging CKD with diabetes as described in claim 6, characterized in that, In step S5, the construction process of the first staged identification model includes: Obtain the first dataset, which consists of samples with CKD stages 1-2 and CKD stages 3-4, and analyze the metabolite markers in the first serum metabolite group and the first urine metabolite group. m / z The mass spectrum response signal intensity at the location; 75% to 85% of the first dataset is used as the first training set, and the remainder is used as the first validation set; Based on the first training set, a first stage identification model is constructed using Lasso regression machine learning, and the first validation set is used to validate the first stage identification model.

8. The method for constructing a model to assist in staging CKD with diabetes as described in claim 6, characterized in that, In step S5, the construction process of the second staging identification model includes: Obtain a second dataset consisting of samples with stage 1-2 and stage 5 CKD diabetes. Metabolic markers were analyzed in the second serum metabolite group and the second urine metabolite group. m / z The mass spectrometry response signal intensity at the location; 75% to 85% of the second dataset is used as the second training set, and the remainder is used as the second validation set; Based on the second training set, a second staging identification model is constructed using Lasso regression machine learning, and the second validation set is used to validate the second staging identification model.

9. The method for constructing a model to assist in staging CKD with diabetes as described in claim 6, characterized in that, In step S5, the construction process of the third stage identification model includes: A third dataset was obtained, consisting of samples with CKD stages 3-4 and CKD stage 5 diabetes. Metabolic markers were analyzed in the third serum metabolite group and the third urinary metabolite group. m / z The mass spectrometry response signal intensity at the location; 75% to 85% of the third dataset is used as the third training set, and the remainder is used as the third validation set; Based on the third training set, a third staging identification model is constructed using Lasso regression machine learning, and the third validation set is used to validate the third staging identification model.

Images