Glycosylated peptide fragment marker for early prediction of cardiovascular complications of peritoneal dialysis patient and application thereof
By screening and applying glycated peptide biomarkers and the glmnet machine learning model, the problem of insufficient sensitivity and specificity in predicting cardiovascular complications in peritoneal dialysis patients in existing technologies has been solved, achieving high stability and accuracy in early prediction and screening out biomarkers with significant diagnostic potential.
Patent Information
- Application Number
- CN202510873642.0
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-06-27
- Publication Date
- 2025-10-17
AI Technical Summary
Existing biomarkers lack sufficient sensitivity and specificity in predicting cardiovascular complications in peritoneal dialysis patients and are easily affected by renal function and dialysis status. Traditional screening methods have a high false positive rate in high-dimensional proteomics data processing, and lack stability and reliability, making early prediction impossible.
By combining glycated peptide biomarkers with the glmnet machine learning model, serum samples from peritoneal dialysis patients were processed and quantitatively analyzed to screen for glycated peptide biomarkers with early predictive capabilities, including specific amino acid sequences and related reagents and consumables. Machine learning was used to handle nonlinear relationships and multidimensional features.
Early prediction of cardiovascular complications in peritoneal dialysis patients was achieved. The screening of glycated peptide biomarkers showed high stability and accuracy. The combined AUC value of 54 biomarkers was 0.964, with a sensitivity of 85.7%, a specificity of 92.9%, and an accuracy of 92.3%. The three preferred biomarkers showed significant diagnostic effects.
Smart Images
Figure CN120795076A_ABST
Abstract
Description
TECHNICAL FIELD
[0001] The present application relates to the field of biomedical technology, in particular to a glycated peptide marker for early prediction of cardiovascular complications in peritoneal dialysis patients and use thereof. BACKGROUND
[0002] For a long time, cardiovascular disease (CVD) has always been the most common complication and the most common cause of death in dialysis patients. The 2024 annual report of the US Renal Data System shows that more than half of the deaths of dialysis patients are related to CVD, and arrhythmia, cardiac arrest (or cardiac sudden death) is the main cause of death. Although traditional CVD risk factors are prevalent in dialysis patients, the pathological and physiological changes caused by dialysis itself and loss of kidney function further exacerbate the risk of CVD, and the high incidence of cardiovascular complications has become a major problem that needs to be solved in the field of peritoneal dialysis treatment. At present, a variety of biomarkers have been found to be closely related to CVD complications in peritoneal dialysis patients, such as cardiac troponin, amino-terminal B-type natriuretic peptide precursor, soluble human matrix metalloproteinase 2, and angiopoietin 2, and many markers have been widely used in clinical practice. However, the existing biomarkers still have the problems of insufficient sensitivity and specificity, and are easily affected by kidney function and dialysis status.
[0003] Proteomics opens up a new way for biomarker screening of various diseases by systematically analyzing the composition and change rule of proteins in the body. The structure and function of proteins may change significantly after modification. Therefore, modified proteomics has become an important branch of proteomics research. Among them, the glycation modification of proteins is the spontaneous covalent binding reaction between reducing monosaccharides and proteins. The reaction process can be divided into two stages: the formation of early glycation products and the generation of advanced glycation end products. Studies have shown that glycosylated proteins are closely related to CVD complications in dialysis patients, but there are still many limitations in using them as biomarkers for CVD complications in dialysis patients, such as AGE, which is an advanced product of protein glycation, cannot predict the occurrence of the disease in a timely and early manner, and the glycosylation modification sequence and site information provided by a single protein is limited. Although proteomics technology has made significant progress in disease marker screening, due to the high dimensionality and complexity of proteomic data, traditional screening methods based on statistical analysis still have great limitations. In contrast, machine learning can significantly improve the stability and accuracy of marker screening by self-learning to handle non-linear relationships and multi-dimensional features.
[0004] Therefore, it is necessary to use glycosylated proteomics technology and machine learning methods to conduct comprehensive and systematic analysis, screen markers that can early predict CVD complications in peritoneal dialysis patients, and further improve their clinical application value. SUMMARY
[0005] The purpose of the present application is to provide a plurality of glycated peptide markers capable of early prediction of cardiovascular complications in peritoneal dialysis patients and applied to early prediction of cardiovascular complications in peritoneal dialysis patients in clinical practice.
[0006] To achieve the above-mentioned purpose, the specific technical solutions adopted by the present application are as follows:
[0007] In the first aspect, the present application provides a glycated peptide marker capable of early prediction of cardiovascular complications in peritoneal dialysis patients, and the amino acid sequence (N-terminal to C-terminal) is SEQ ID NO: 1-SEQ ID NO: 54 shown in Table 1 below, and g in each sequence is a glycosylation modification site.
[0008] Table 1: Amino acid sequence information of 54 glycated peptide markers
[0009]
[0010]
[0011] The concentration level of the above-mentioned glycated peptide marker in the serum of peritoneal dialysis patients with cardiovascular complications is higher than that of peritoneal dialysis patients without cardiovascular complications. The glycosylation modification process can generally be divided into two stages: in the first stage, glycosylation modification mainly occurs on the side chain amino group of lysine, and the product formed is early glycation product; in the second stage, the early glycation product further reacts to form late glycation product with more complex structure. The glycated peptide marker of the present application is early glycation product in which glycosylation reaction occurs on lysine, so detecting it can realize early prediction of cardiovascular complications in peritoneal dialysis patients.
[0012] Further, the application method is to quantify the concentration level of the glycated peptide marker in the serum of the patient, and the steps are as follows:
[0013] S1. Serum sample pretreatment: adding sodium cyanoborohydride solution to the serum sample to reduce the protein; adding dithiothreitol and iodoacetamide solution to the reduced protein solution to perform reduction alkylation reaction; then adding peptide N-glycosidase F to remove N-glycan chain; collecting the protein filtrate and adding trypsin for enzymolysis;
[0014] S2. Enrichment and desalination of glycated peptides: adding the collected enzymolysis filtrate to boric acid material for glycated peptide enrichment; after enrichment is completed, taking the supernatant and adding it to a Sep-Pak C18 solid-phase extraction column for desalination, and collecting the eluate for freeze-drying;
[0015] S3. Quantification of glycated peptides: after the freeze-dried powder is redissolved, the glycated peptides in the sample are quantified by LC-MS / MS analysis.
[0016] In a second aspect, the present invention provides the use of a glycated peptide marker in the preparation of a product for early prediction of cardiovascular complications in peritoneal dialysis patients, wherein the amino acid sequence of the α1-antitrypsin glycated peptide marker is selected from any one or more of SEQ ID NO: 1-SEQ ID NO: 54, preferably SGLSTGWTQLSK(g)LLELTGPK, GEAFTLK(g)ATVLNYLPK or K(g)QHLFVK, i.e., SEQ ID NO: 1, SEQ ID NO: 8 or SEQ ID NO: 10.
[0017] In a third aspect, the present invention provides a product for predicting cardiovascular complications in peritoneal dialysis patients, comprising reagents and / or consumables for detecting the concentration level of glycosylated peptide markers in the patient's serum, wherein the sequence of the glycosylated peptide is selected from one or more of SEQ ID NO: 1-SEQ ID NO: 54, preferably one or more of SGLSTGWTQLSK(g)LLELTGPK, GEAFTLK(g)ATVLNYLPK, and K(g)QHLFVK.
[0018] Furthermore, the product includes the following reagents / consumables: sodium cyanoborohydride, dithiothreitol, iodoacetamide, peptide N-glycosidase F, trypsin, boric acid material, Sep-Pak C18 solid phase extraction column.
[0019] The present invention has the following beneficial effects:
[0020] 1. Traditional screening methods typically rely on statistical analysis. These methods often struggle to effectively control false positive rates when processing high-throughput proteomic data, and the stability and reliability of screening results are insufficient. The present invention utilizes the GLMNet machine learning model for screening glycosylated peptide markers. This model, through self-learning, can process nonlinear relationships and multidimensional features in data, uncovering underlying patterns from high-dimensional data and significantly improving the stability and accuracy of marker screening.
[0021] Currently, there is a lack of early predictive biomarkers for CVD complications in peritoneal dialysis patients. Although AGEs have been shown to be closely associated with the development of CVD in dialysis patients, they are late products of protein glycation and therefore cannot provide timely and early prediction of the disease's onset. The glycated peptide biomarkers screened in this invention enable early diagnosis of CVD complications in peritoneal dialysis patients.
[0022] 3. The present invention screened 54 glycosylated peptide markers that significantly contributed to the differentiation of CVD patients from those without CVD. The combined AUC value of the 54 differentially glycosylated peptides was 0.964, with a sensitivity of 85.7%, a specificity of 92.9%, and an accuracy of 92.3%.
[0023] 4、The application further screens 3 preferred glycated peptide markers from the 54 markers according to the target peptide segment selection criteria, the mean of the glmnet model importance coefficient, and the chemical properties of the glycated peptide segments, which exhibit better diagnostic efficiency. BRIEF DESCRIPTION OF DRAWINGS
[0024] Figure 1 : Comparison results of different machine learning models in Example 1. In the figure: ROC AUC is the area under the ROC curve, used to evaluate the ability of the classification model to distinguish between positive and negative samples; Accuracy is the accuracy rate, which refers to the proportion of samples predicted correctly by the model in the total samples; F1 score is the harmonic mean of precision and recall, used to comprehensively evaluate the performance of the model.
[0025] Figure 2 : Specific information of the 54 glycated peptide markers screened in Example 1.
[0026] Figure 3 : Performance of the 54 glycated peptide markers screened in Example 1 in the test set. In the figure: (a) ROC curve of the test set classification results; (b) confusion matrix diagram of the test set classification results, with the horizontal axis Truth representing the true class of the sample and the vertical axis Response representing the predicted class of the model.
[0027] Figure 4 : Prediction effect diagram of the 3 glycated peptide markers in Example 2. DETAILED DESCRIPTION
[0028] Glycation modification can change the structure and function of proteins and plays an important role in the occurrence and development of diseases. Previous studies have shown that glycosylated proteins are closely related to the occurrence of cardiovascular diseases. Glycated proteomics provides a powerful research tool for revealing the mechanism of disease occurrence by systematically analyzing the composition and change rules of glycosylated proteins in the body, and also opens up a new way for early diagnosis of diseases. In this application, the technology is used to compare the composition of serum glycated proteomes of patients with and without CVD complications in peritoneal dialysis, and machine learning methods are used to screen glycated peptide markers for CVD complications, which are applied to the diagnosis of CVD complications in peritoneal dialysis patients.
[0029] The application will be further described below in conjunction with the drawings and specific examples.
[0030] Example 1
[0031] This example performs screening of glycated peptide markers, as follows:
[0032] 1. Inclusion of research subjects:
[0033] Forty-eight pairs of peritoneal dialysis patients were enrolled, and serum samples before dialysis and related clinical data were collected. According to whether CVD occurred in the past 6 months, the patients were divided into CVD group and non-CVD group. The gender, age and dialysis age of the patients in the two groups were matched.
[0034] 2. Pretreatment of serum samples:
[0035] ① 1.11 μL of 75 mmol / L sodium cyanoborohydride solution was added to 10 μL of serum sample, and the protein was reduced at 37°C and 1100 rpm for 4 hours;
[0036] ② 500 μg of reduced protein solution was taken into a 10KDa ultrafiltration tube, and ammonium bicarbonate solution was used to remove excess sodium cyanoborohydride solution. Then 4.04 μL of 1 mol / L dithiothreitol was added, and the reaction was carried out at 56°C for 1 hour. Subsequently, 8.08 μL of 1 mol / L iodoacetamide solution was added, and the reaction was carried out at room temperature and in the dark for 45 minutes. After the above reduction and alkylation reaction was completed, the unreacted reagents were removed by ultrafiltration;
[0037] ③ 0.5 μL of peptide N-glycosidase F was added, and the reaction was carried out at 37°C and 150 rpm for 16 hours to remove N-glycan chains. After the reaction was completed, the protein filtrate was collected;
[0038] ④ Trpsin was added to the protein filtrate at a mass ratio of 50:1 (serum protein: trypsin), and the enzyme digestion was carried out at 37°C and 1100 rpm for 16 hours. After the enzyme digestion was completed, the filtrate was collected.
[0039] 3. Enrichment and desalination of glycopeptides:
[0040] ① The collected filtrate was added to 20 mg of equilibrated and activated boronic acid material (Bio-rad, Affi-Gel Boronate Media, 1536103), and the reaction was carried out at 37°C and 1100 rpm for 16 hours;
[0041] ② After the glycosylation enrichment was completed, the supernatant was added to a Sep-Pak C18 solid-phase extraction column for desalination. After 3 times of loading and 2 times of washing, the desalted eluate was collected and freeze-dried for storage.
[0042] 4. Identification and screening of glycopeptide markers:
[0043] The lyophilized sample was reconstituted with 0.1% formic acid solution to a concentration of 0.5 μg / μL, and then analyzed by high-performance liquid chromatography-mass spectrometry using an Orbitrap Exploris 480 high-resolution mass spectrometer to obtain signal intensity data. The raw data file generated by LC-MS / MS testing was processed using PEAKS Online software, and the database was the Swiss-Prot annotated human database downloaded from the Uniport official website (release time 2023-04). The de novo sequencing algorithm was used to obtain the amino acid sequence of the peptide segment, the enzyme cutting mode was set to full enzyme cutting, the false discovery rate was set to less than 1%, and the maximum allowed missed cutting site was set to 4. The database retrieval fixed modification was set to carbamidomethylation (+57.021 Da); the variable modification was set to protein N-term acetylation (+42.010 Da), methionine oxidation (+15.995 Da), asparagine deamidation (+0.984 Da), and lysine glycation (+164.068 Da). The precursor mass error was set to 10 ppm, and the fragment mass error was set to 0.05 Da. The relative intensity obtained based on the peptide mass spectrum peak was extracted using the LFQ algorithm in the software.
[0044] Machine learning analysis was performed using the R package mlr3verse in R software (version 4.3.0). First, 16 mainstream machine learning models were used for analysis, including glmnet, Neural network, glmboost, MARS, GBM, XGBoost, GAMboost, LightGBM, random forest, Naive Bayes, LDA, SVM, FNN, logistic regression, and kNN. The performance of the model in distinguishing between the CVD group and the non-CVD group was evaluated, and the evaluation indicators mainly included the area under the receiver operating characteristic curve, accuracy, and F1 score. The parameters of all models were optimized using auto_tuner in mlr3tuning. After comprehensive comparison, the glmnet model showing the best discrimination ability was selected as the marker screening model. Figure 1 Based on the results of four-fold cross-validation, the coefficient values of each feature variable were obtained and used as an important basis for screening markers. The screening group samples were randomly divided into training and test sets in a 7:3 ratio. After training the optimal model in the training set, the model performance was further tested in the test set. The evaluation methods included AUC and confusion matrix test, where the judgment threshold of the confusion matrix was 0.5, and higher than 0.5 was determined as the CVD group, and less than or equal to 0.5 was the non-CVD group.
[0045] After excluding one ELDRDTVFALVNYIFFK(g)GK peptide segment which has been studied, 54 significant differential glycation peptides were obtained from the glmnet model. The 54 significant differential glycation peptides were ranked according to the mean of the importance coefficients, and the specific information is shown in Table 2. Figure 2 As shown in Table 2, the 54 significant differential glycation peptides were ranked according to the mean of the importance coefficients, and the specific information is shown in Table 2. Figure 3 As shown in Table 2, the 54 significant differential glycation peptides were ranked according to the mean of the importance coefficients, and the specific information is shown in Table 2.
[0046] According to the selection criteria of the target peptide segment, the 54 differential glycation peptides were further screened, and the screening criteria were as follows: 1. unique peptide segment of a specific protein; 2. length of 7-25 amino acids; 3. mass <6000 Da and detectability ≥0.5; 4. no methionine, cysteine or other post-translational modification sites. Finally, three candidate glycation peptide segment markers SGLSTGWTQLSK(g)LLELTGPK, GEAFTLK(g)ATVLNYLPK and K(g)QHLFVK were obtained.
[0047] The diagnostic performance of the three glycation peptide segment markers was tested: the FC values (i.e. the ratio of the mass spectrometric signal intensity of the CVD group to the CVD-free group) of the CVD group and the CVD-free group were calculated using the mass spectrometric relative signal intensity of 48 pairs of peritoneal dialysis patients, and the paired t test and the receiver operating characteristic curve were drawn using the relative quantitative data of the glycation peptide segment measured by mass spectrometry. The results are shown in Table 3.
[0048] Table 2. Diagnostic performance of the three candidate glycation peptide segment markers
[0049]
[0050] As can be seen, the areas under the receiver operating characteristic curves of the three glycation peptide segment markers tested were 0.848, 0.820 and 0.760, respectively, indicating that the three glycation peptide segment markers had strong ability to distinguish between CVD and non-CVD patients, and had good diagnostic effect on CVD complications in peritoneal dialysis patients.
[0051] This specific embodiment is only an explanation of the present application, and is not a limitation of the present application. Any changes made by those skilled in the art after reading the specification of the present application will be protected by the patent law as long as they are within the scope of the claims of the present application.
Claims
1. Application of glycosylated peptide markers in the preparation of products for early prediction of cardiovascular complications in peritoneal dialysis patients, characterized in that: The amino acid sequence of the glycosylated peptide marker is selected from one or more of SEQ ID NO: 1 to SEQ ID NO: 54, and g in each sequence represents a glycosylation modification site.
2. The use according to claim 1, characterized in that The amino acid sequence of the glycosylated peptide marker is selected from SGLSTGWTQLSK(g)LLELTGPK, GEAFTLK(g)ATVLNYLPK or K(g)QHLFVK.
3. The use according to claim 1, characterized in that The serum concentrations of the glycated peptide markers in peritoneal dialysis patients with cardiovascular complications were higher than those in peritoneal dialysis patients without cardiovascular complications.
4. The use according to claim 3, characterized in that The application method is to quantify the concentration level of glycosylated peptide markers in the patient's serum. The steps are as follows: S1. Serum sample pretreatment: Sodium cyanoborohydride solution is added to the serum sample to reduce the protein; dithiothreitol and iodoacetamide solution are added to the reduced protein solution for reductive alkylation; peptide N-glycosidase F is then added to remove N-glycan chains; Collect the protein filtrate and add trypsin for enzymatic digestion; S2. Enrichment and desalting of glycated peptides: The collected enzymatic filtrate is added to a boric acid material to enrich the glycated peptides; After the enrichment was completed, the supernatant was added to a Sep-Pak C18 solid phase extraction column for desalting, and the eluate was collected and freeze-dried; S3. Quantification of Glycosylated Peptides: After the lyophilized powder is reconstituted, the glycated peptides in the sample are quantified by LC-MS / MS analysis.
5. A product for early prediction of cardiovascular complications in peritoneal dialysis patients, characterized in that: The invention comprises reagents and / or consumables for detecting the concentration level of glycosylated peptide markers in patient serum, wherein the sequence of the enzymatically glycosylated peptide is selected from one or more of SEQ ID NO: 1-SEQ ID NO:
54.
6. The product according to claim 5, characterized in that The sequence of the glycosylated peptide segment is selected from one or more of SGLSTGWTQLSK(g)LLELTGPK, GEAFTLK(g)ATVLNYLPK, and K(g)QHLFVK.
7. The product according to claim 5 or 6, characterized in that include: Sodium cyanoborohydride, dithiothreitol, iodoacetamide, peptide N-glycosidase F, trypsin, boric acid material, Sep-Pak C18 solid phase extraction column.