Biological age evaluation method based on N-glycosylation spectrum pattern of plasma immunoglobulin G

Abstract

The invention discloses a method for constructing an N-glycosylation spectrum type biological age evaluation model based on plasma immunoglobulin G and application of the biological age evaluation model. Based on a glycosyl quantitative result, a clinical biochemical detection index, a lifestyle measurement index, a multi-dimensional health score and the like, a glycosyl metabolic index, a glycosyl inflammation index, a glycosyl cardiovascular index, a glycosyl stroke index, a glycosyl liver index and a glycosyl kidney index are fitted. And inputting the obtained different glycosyl data into the model, and speculating the aging process and health risk of each organ system of the detected person according to the risk threshold in the early-stage crowd. The method can predict a'glycosyl biological age system 'which can reflect health level and disease risk better than'calendar age', has individual specificity, is very sensitive to change of inflammatory response in vivo, is relatively stable in a short time, can be used as an index for early detection and screening of senescence, and can be used as an index for early detection and screening of senescence. And thus, effective reference is provided for individualized health suggestions of the detected personnel.

Description

Technical Field

[0001] The present invention relates to the field of life and health, and in particular to a biological age assessment method based on the N-glycosylation spectrum of plasma immunoglobulin G. Background Art

[0002] Aging refers to the process of physiological integrity, including the gradual damage of molecules, cells and organs, the intensification of chronic inflammatory responses, the increase in disease susceptibility, and ultimately death. Studies have shown that although the increase in "calendar age" during the aging process is uncontrollable, a more active and healthy lifestyle, such as continuous physical exercise and smoking cessation, can delay the aging process. Therefore, the research on screening molecular markers of the aging process and calculating "biological age" to predict the aging process and the risk of age-related diseases has made unprecedented progress in recent years.

[0003] Disturbances in protein homeostasis, such as the secretion of unfolded, misfolded or aggregated proteins and misdegradation of proteins, have been shown to be associated with aging and age-related diseases. The polypeptide sequence of a protein is determined by the nucleotide sequence in the gene and does not change with age. However, more than half of the proteins and almost all membrane proteins and secretory proteins undergo a complex post-translational modification process, namely glycosylation, during their formation. These proteins are called glycoproteins. The activity and function of these glycoproteins are significantly affected by N-glycosylation, including immunoglobulins (Ig), the main molecules of humoral immunity.

[0004] Immunoglobulin (Ig) is a class of proteins with antibody activity, including IgG, IgM, IgA, IgE, and IgD, accounting for about 20% of plasma proteins, of which IgG has the largest content (>80%). IgG is considered a perfect model glycoprotein because it has a well-defined glycan structure and function. It consists of two identical light chains and heavy chains, which are covalently linked to form a "y"-shaped structure. The two domains of IgG in vivo are the antigen-binding fragment and the crystallizable fragment (Fc). In the heavy chain of all IgG subclasses, there is a single N-linked glycosylation site at amino acid 297. Slight changes in the composition of attached glycans can significantly change the conformation of the Fc domain and have a significant impact on the effector function of IgG. For example, the addition of terminal sialic acid to this glycan can reduce the binding of Fcg receptors (FcgRs) and convert IgG antibodies from pro-inflammatory mediators to anti-inflammatory mediators through the lectin receptor SIGN-R1 or the intercellular adhesion molecule DCSIGN, thereby upregulating the expression of the inhibitory Fc receptor FcgRIIb on inflammatory cells, thereby attenuating autoantibody-induced inflammation. Another important role played by glycosylation in IgG is the regulation of antibody-dependent cellular cytotoxicity (ADCC): the addition of fucose residues to the glycan core can change the conformation of the Fc domain, significantly reducing its affinity for FcgRIIIa; therefore, core-fucosylated IgG is less efficient in activating ADCC. Most IgG (>95%) are core-fucosylated, which prevents them from erroneously initiating ADCC. Studies have found that the level of galactosylation in Japanese women decreases with age.

[0005] Currently, the methods for measuring biological age mainly focus on epigenetic clocks (such as DNA methylation levels), telomere length, and proteomics and transcriptomics data, but they have the following defects:

[0006] 1. Inaccurate assessment: Traditional calendar age cannot accurately reflect the biological aging process of an individual, resulting in low accuracy in health risk prediction. Some existing biomarkers, such as telomere length and epigenetic clocks, also have certain limitations and cannot fully reflect an individual's overall health status and aging rate.

[0007] 2. Lack of comprehensiveness: Existing biomarkers, such as epigenetic clocks and telomere length, can only reflect biological changes in a certain aspect and lack a comprehensive assessment of the overall physiological state. Different biomarkers may have complex relationships that are independent or interactive with each other, and a single marker cannot fully reflect complex physiological and pathological processes.

[0008] 3. Complexity and high cost: The existing methods for detecting multiple biomarkers are complex and expensive, which limits their application in clinical and large-scale populations. In particular, some high-throughput omics technologies, such as epigenomics and transcriptomics, can provide rich information, but their high cost and technical barriers hinder their widespread application.

[0009] In the prior art, there has been an analysis of N-glycoproteins in plasma proteins of healthy people to predict the development of type 2 diabetes. This invention involves predicting the development of type 2 diabetes by quantitatively analyzing N-glycans attached to plasma proteins of healthy people, which makes it possible to determine whether the person under investigation belongs to the risk group for the development of type 2 diabetes. Its shortcomings are that only the risk of developing type 2 diabetes is predicted. The parameters used for prediction are only the N-glycan values ​​on plasma proteins, without the assistance of individual epidemiological and clinical indicators.

[0010] The prior art already has a method for establishing a serum glycoprotein N-glycome profile model for chronic hepatitis liver damage. This invention discloses a method for establishing a serum glycoprotein N-glycome profile model for chronic hepatitis liver damage. The serum glycoprotein N-glycome profile was detected using the G-Test detection method, and an N-glycome profile model with significant differences between liver damage patients and normal controls was established, and NG1A2F with significant expression differences between the liver damage group and the normal control group was screened out. Insufficient: Only liver damage was predicted.

[0011] The prior art includes a method for analyzing N-glycans linked to immunoglobulin G in human plasma and its application

[0012] The invention discloses a method for analyzing N-glycans or glycopeptides linked to immunoglobulin G in human plasma; wherein, two or more glycans are selected from a total of six glycans that are clearly strongly correlated with age, and their relative abundance is determined. Glycans have the following structure: GlcNAc = N-acetylglucosamine, Fuc = fucose, Man = mannose, NeuAc = N-acetylneuraminic acid, Gal = galactose. Based on the analysis results, the glycan age index (GAI) is calculated, which can be used to: predict the biological and / or chronological age of the subject; monitor the effectiveness of methods for delaying the aging process. Disadvantages: Only six glycan information is used, and all glycan information is not fully used.

[0013] Therefore, there is an urgent need for a new technical means to more comprehensively and accurately assess an individual's biological age.

[0014] Unlike the "biological age" markers found in previous studies that are only related to genes (such as certain gene loci) or can only reflect the process of cell apoptosis (such as telomere length), the human plasma IgG N-glycosylation profile is affected by both genes and the environment. Its changes can dynamically reflect the overall health status of the human body and are highly reproducible under certain physiological conditions. Previous studies have found that IgG N-glycosylation profiles are not only related to aging and inflammation, but also significantly associated with cardiovascular metabolic diseases such as hypertension, diabetes, and dyslipidemia. However, most of these studies use a single death, aging-related health outcome, or cardiovascular disease outcome as a prediction target, and the aging rates of various organ systems are different. If the prediction of individual health risks distinguishes different organ systems, it will be more targeted.

[0015] The present invention is based on the IgG N-glycosylation spectrum and personal health examination database, applies the polynomial regression method to establish a prediction model, and constructs a glycosyl biological age system based on multiple organ systems, which includes glycosyl metabolic index, glycosyl liver index, glycosyl cardiovascular index, glycosyl inflammatory index, glycosyl kidney index and glycosyl stroke index, which comprehensively reflects the individual aging process. Summary of the invention

[0016] The complex carbohydrate molecules attached to IgG are called IgG glycosyl. A typical IgG N-glycosyl has two branches. Each branch can be composed of different monosaccharides, including galactose, sialic acid, mannose, fucose, and N-acetylglucosamine. When a branch ends with galactose, this carbohydrate residue is called terminal galactose. Similarly, when a branch ends with sialic acid, it is called terminal sialic acid. The composition of IgG glycosyl in the human body is very complex. At present, more than 190 different glycosyl structures have been discovered. The distribution of these structures varies with gender, age, health status, physiological state and environmental factors. Therefore, everyone has a different IgG N-glycosyl spectrum.

[0017] The present invention discloses a method for constructing a biological age assessment model based on the N-glycosylation spectrum of plasma immunoglobulin G, comprising the following steps:

[0018] 1) Collect blood samples and extract plasma IgG;

[0019] 2) Detect the N-glycosylation of IgG to obtain the N-glycosylation spectrum;

[0020] 3) The area under the chromatographic peak of each sugar group is divided by the area under the total chromatographic peak as the horizontal data of the sugar group, and the obtained sugar group data are pre-processed to obtain sugar group data with normal distribution;

[0021] 4) The glycosylated metabolic index was calculated based on the metabolic syndrome severity index and plasma IgG N-glycosylation data; the glycosylated liver index was calculated based on the liver fibrosis FIB-4 index and plasma IgG N-glycosylation data; the glycosylated cardiovascular index was calculated based on the Framingham cardiovascular disease risk score and plasma IgG N-glycosylation data; the glycosylated inflammatory index was calculated based on the systemic immune inflammation index and plasma IgG N-glycosylation data; the glycosylated renal index was calculated based on the urea nitrogen to creatinine ratio and plasma IgG N-glycosylation data; and the glycosylated stroke index was calculated based on the combined cohort stroke risk equation and plasma IgG N-glycosylation data;

[0022] 5) The obtained scores of glycosylation metabolism index, glycosylation liver index, glycosylation cardiovascular index, glycosylation inflammation index, glycosylation kidney index and glycosylation stroke index are integrated as the glycosylation biological age score;

[0023] 6) Use Gaussian Process Regression (GPR) or polynomial regression to construct a biological age assessment model;

[0024] 7) Evaluate the predictive ability of the constructed biological age assessment model.

[0025] Preferably, in step 2), high performance liquid chromatography or ultra high performance liquid chromatography is used to obtain the N-glycosylation profile.

[0026] Preferably, in step 3), Z-score variation, logarithmic variation and polynomial variation are used to obtain normally distributed glycosyl data.

[0027] Preferably, in step 4), glycosyl data dimension reduction is performed based on LASSO regression to obtain glycosyl structures associated with each index.

[0028] Preferably, the glycosyl structures associated with the glycosyl metabolism index are: GP2, GP3, GP4, GP5, GP6, GP9, GP10, GP11, GP12, GP14, GP15, GP16, GP20, GP22, GP24;

[0029] The glycosylated structures associated with the glycosylated inflammatory index are: GP1, GP4, GP6, GP8, GP9, GP12, GP15, GP19, GP20, GP22, GP24;

[0030] The glycosylated structures associated with the glycosyl cardiovascular index are: GP1, GP4, GP5, GP6, GP11, GP12, GP15, GP19, GP20, GP21, GP23, GP24;

[0031] The glycosylated structures associated with the glycosylated stroke index are: GP1, GP4, GP5, GP6, GP8, GP9, GP11, GP12, GP15, GP17, GP19, GP20, GP21, GP23, GP24;

[0032] The glycosylated structures associated with the glycosyl liver index are: GP2, GP4, GP5, GP6, GP8, GP9, GP10, GP11, GP12, GP13, GP14, GP19, GP20;

[0033] The glycosylated structures associated with the glycosylated renal index are: GP10, GP11.

[0034] Preferably, the biological age assessment model in step 6) is: Male: GlycanAge = 51.48 + 34.97*GP3 + 26.18*GP5 - 40.15*GP6 2 +52.74*GP8 2 +38.50*GP14 2 -34.71*GP16 2 -51.14*GP17 2 -58.50*GP22;

[0035] Female: GlycanAge=49.18–29.55*GP2 2 +28.32*GP3-37.72*GP4 2 –53.04*GP8 2 +53.70*GP14 2 +33.68*GP18-48.11*GP20 2 +38.04*GP21.

[0036] Preferably, the prediction ability evaluation in step 7) is evaluated using root mean square error (RMSE) and R-square.

[0037] Preferably, the GP value is the percentage of the area under the chromatographic peak of each structural sugar to the area under the total chromatographic peak; the sugar peak structure corresponding to the GP2 is FA1, the sugar peak structure corresponding to the GP3 is A2B, the sugar peak structure corresponding to the GP4 is FA2, the sugar peak structure corresponding to the GP5 is M5, the sugar peak structure corresponding to the GP6 is FA2B, the sugar peak structure corresponding to the GP8 is FA2[6]G1, the sugar peak structure corresponding to the GP14 is FA2G2, the sugar peak structure corresponding to the GP16 is FA2G1S1, the sugar peak structure corresponding to the GP17 is A2G2S1, the sugar peak structure corresponding to the GP18 is FA2G2S1, the sugar peak structure corresponding to the GP20 is FA2FG2S1, the sugar peak structure corresponding to the GP21 is A2G2S2, and the sugar peak structure corresponding to the GP22 is A2BG2S2.

[0038] The present invention discloses a method for evaluating biological age of a sample, comprising the following steps:

[0039] 1) Collect blood samples and extract plasma IgG;

[0040] 2) Detect the N-glycosylation of IgG to obtain the N-glycosylation spectrum;

[0041] 3) The area under the chromatographic peak of each sugar group divided by the area under the total chromatographic peak is the horizontal data of the sugar group;

[0042] 4) Evaluate the biological age of the sample using a biological age evaluation model, the evaluation model is: Male: GlycanAge = 51.48 + 34.97*GP3 + 26.18*GP5 - 40.15*GP6 2 +52.74*GP8 2 +38.50*GP14 2 -34.71*GP16 2 -51.14*GP17 2 -58.50*GP22;

[0043] Female: GlycanAge=49.18–29.55*GP2 2 +28.32*GP3-37.72*GP4 2 –53.04*GP8 2 +53.70*GP14 2 +33.68*GP18-48.11*GP20 2 +38.04*GP21.

[0044] Preferably, in step 2), high performance liquid chromatography or ultra high performance liquid chromatography is used to obtain the N-glycosylation profile.

[0045] The present invention discloses the use of a biological age evaluation model in predicting the degree of individual aging, wherein the evaluation model is: Male: GlycanAge = 51.48 + 34.97*GP3 + 26.18*GP5 - 40.15*GP6 2 +52.74*GP8 2 +38.50*GP14 2 -34.71*GP16 2 -51.14*GP17 2 -58.50*GP22;

[0046] Female: GlycanAge=49.18–29.55*GP2 2 +28.32*GP3-37.72*GP4 2 –53.04*GP8 2 +53.70*GP14 2 +33.68*GP18-48.11*GP20 2 +38.04*GP21.

[0047] The "glycosylation metabolism index" disclosed in the present invention should be an index calculated by an equation including glycosylation data and metabolism-related indicators.

[0048] GP, or Glycan Peak, is the percentage of the area under the chromatographic peak of each structural sugar to the area under the total chromatographic peak.

[0049] The present invention discloses a male biological age evaluation model, wherein the evaluation model is: GlycanAge=51.48+34.97*GP3+26.18*GP5-40.15*GP6 2 +52.74*GP8 2 +38.50*GP14 2 -34.71*GP16 2 -51.14*GP17 2 -58.50*GP22.

[0050] The present invention discloses a female biological age evaluation model, wherein the evaluation model is: GlycanAge = 49.18-29.55*GP2 2 +28.32*GP3-37.72*GP4 2 –53.04*GP8 2 +53.70*GP14 2 +33.68*GP18-48.11*GP20 2 +38.04*GP21.

[0051] The present invention is based on the immunoglobulin G (IgG) N-glycosylation spectrum and personal health examination database, and applies machine learning algorithms and other methods to establish a "glycosylated biological age" prediction model, which solves the technical problem of multi-dimensional prediction of individual health risks.

[0052] The sugar group data involved in the present invention are all obtained by high performance liquid chromatography and ultra-high performance liquid chromatography. This method is currently the most mature and stable detection method with high throughput, but the sample requirement is large. With the development of sugar group detection technology, capillary electrophoresis and other technologies can also be used in the future to quantitatively and qualitatively analyze plasma IgG N-sugar groups.

[0053] The present invention combines plasma IgG N-glycosylation spectra with personal health databases, not only taking into account changes in N-glycosylation, but also integrating individual epidemiological data and clinical indicators, such as lifestyle, metabolic indicators, etc., to construct a multi-dimensional "glycosylated biological age system". This method combines the biological age indexes of multiple organ systems, including glycosylated metabolic index, glycosylated inflammatory index, glycosylated cardiovascular index, glycosylated liver index, glycosylated stroke index, glycosylated kidney index, etc., thereby providing a comprehensive individual health risk assessment. This method can not only assess an individual's metabolic status, inflammation level, etc., but also assess the individual's overall health status.

[0054] By analyzing the N-glycosylation spectrum of plasma IgG, the present invention takes into account more glycosylation structures and constructs a multi-dimensional "glycosylated biological age system", which can more comprehensively reflect the health status of an individual, including metabolism, inflammation, cardiovascular health and other aspects, and obtain a comprehensive assessment of the individual dimension.

[0055] The present invention can predict a "glycogen-based biological age system" that can better reflect health levels and disease risks than "calendar age". Changes in the "glycogen-based biological age system" are affected by both the environment and genetics, are individual-specific, and are very sensitive to changes in inflammatory responses in the body. It can be used as an indicator for early detection and screening of aging, and thus provide an effective reference for individualized health recommendations for the tested person. BRIEF DESCRIPTION OF THE DRAWINGS

[0056] Figure 1 Results of plasma IgG N-glycosylation profiling using ultra-performance liquid chromatography.

[0057] Figure 2 Scatter plot and fitting curve of glycosylbiological age based on Gaussian process regression (male).

[0058] Figure 3 Scatter plot and fitting curve of glycosylbiological age based on Gaussian process regression (female).

[0059] Figure 4 Scatter plot and fitted curve of glycosylbiological age based on polynomial regression (male).

[0060] Figure 5 Scatter plot and fitted curve of glycosylbiological age based on polynomial regression (female). DETAILED DESCRIPTION

[0061] Example 1 Quantitative and qualitative analysis of plasma IgG N-glycosyl groups by high performance and ultra-high performance liquid chromatography

[0062] (1) Blood sample collection

[0063] The person being tested is required to have a light diet, avoid drinking a lot of alcohol, stop smoking the day before blood collection, and cannot take hypoglycemic drugs on the morning of blood collection. On the day of blood collection, a professional nurse will use ethylenediaminetetraacetic acid (EDTA) anticoagulation vacuum blood collection tubes to draw 5 ml of fasting arm venous blood from the person being tested in the morning, and the blood sample will be immediately refrigerated and transported to the sugar base testing laboratory.

[0064] (2) Blood sample processing

[0065] Take out the blood collection tube, paste the sample number (try to keep it stable during the process and keep the blood collection tube in a vertical state), and place it on a test tube rack for 1 hour at room temperature. Then, put the blood collection tube into a centrifuge at 2000rpm and centrifuge at 4℃ for 15 minutes. After taking it out, use a pipette to absorb the upper plasma sample and divide it into 3 1.5ml centrifuge tubes. Paste the same sample number as the blood collection tube on the wall and cap of the centrifuge tube. The plasma sample is immediately subjected to subsequent IgG N-glycosylation testing, or stored in a -80℃ ultra-low temperature refrigerator for subsequent testing.

[0066] (3) Plasma IgG extraction

[0067] IgG was extracted using a hydrophilic liquid chromatography extraction plate containing Protein G. First, the extraction plate was washed with ultrapure water, binding buffer, and neutralization buffer in sequence. 60 μl of plasma sample was diluted with binding buffer at a ratio of 1:7, and then the diluted plasma sample was filtered using a filter plate with a pore size of 0.45 μm. The filtered sample was transferred to the IgG extraction plate, and the extraction plate was washed three times with 2 ml of binding buffer. Finally, IgG was eluted from the extraction plate with 1 ml of 0.1 M formic acid solution, and the eluate was collected. The extracted IgG sample was stored in a -20 °C refrigerator for the next step of analysis.

[0068] (4) N-glycosylase cleavage

[0069] 30 μl 1.33% sodium dodecyl sulfate (SDS) was added to each IgG sample and incubated in a water bath at 65°C for 10 min. After cooling at room temperature for 20 min, 10 μl 4% Igepal-CA630 was added to each sample. 100 units of glycosidase (N-Glycosidase F) were dissolved in 100 μl ultrapure water. 10 μl glycosidase solution and 10 μl 5×PBS (pH=8.0) were added to each IgG sample, and then 30 μl 0.1M sodium hydroxide solution was added, and incubated in a water bath at 37°C for 18 h.

[0070] (5) Glycosyl 2-AB fluorescence labeling

[0071] Each N-glycan sample after enzyme digestion requires 35 μl of 2-AB labeling reagent, shaken on a shaker for 10 minutes, and then incubated at 65°C for 3 hours. Add 700 μl of 100% acetonitrile to each labeled sample, wash the labeled N-glycans, and use a vacuum extraction device to filter the samples and remove the waste liquid.

[0072] (6) N-Glycan UPLC Detection

[0073] The 2-AB labeled IgG N-glycosylation was detected by ultra-high pressure liquid chromatography system, and the N-glycosylation with different structures was separated by hydrophilic interaction chromatography principle, and the level of N-glycosylation was detected by fluorescence detector. The system includes a quaternary solvent management system, a sample manager and a fluorescence detector. The excitation wavelength of the fluorescence detector is 330nm and the emission wavelength is 420nm. The chromatographic column is a BEH C18 column with a volume of 2.1×50mm and internally filled with BEH particles with a particle size of 1.7μm. The mobile phase A is 50mmol / L ammonium formate solution with pH=4.4, and the mobile phase B is acetonitrile solution. The temperature of the sample before loading is 5℃, the column temperature is set to 60℃, and the elution gradient and flow rate settings are shown in Table 1 below:

[0074] Table 1 Ultra-high pressure liquid chromatography elution program settings

[0075]

[0076] The final glycosylation profile contains 24 glycan peaks (GP) (such as Figure 1 ).

[0077] The IgG N-glycosylation structure is represented by the following letter and number codes commonly used in the field of glycosylome research: F represents a core fucose attached to the core N-acetylglucosamine; Mx represents "x" mannoses attached to the core N-acetylglucosamine; Ax represents "x" branches (N-acetylglucosamine) attached to the trimannose core; A2 represents that both branches of the core mannose are attached to N-acetylglucosamine; B represents bisecting N-acetylglucosamine attached to the core mannose; Gx represents "x" galactoses attached to the branches; [3] G1 and [6] G1 represent galactoses attached to α1-3 and α1-6 mannoses, respectively; Sx represents "x" sialic acids attached to galactose. Table 2 shows the meanings of the 22 glycosylation peaks measured experimentally and the meanings of the glycosylation peak structures. See Table 3 for the structural diagrams.

[0078] Table 2 IgG N-glycosyl structure correspondence table under HPLC peak

[0079]

[0080] IgG N-glycosylated structure under HPLC peak

[0081]

[0082]

[0083]

[0084] Among them, graphic symbols are used to represent different oligosaccharide residues, and different glycosyl structures are represented according to the oligosaccharide residues and linkages, such as the blue square represents N-acetylglucosamine, the green circle represents mannose, the red triangle represents fucose, the yellow circle represents galactose, and the purple diamond represents sialic acid.

[0085] Example 2: Correlation between glycosylation level and inflammatory response

[0086] 107 individuals who were about to undergo surgery were collected. Since surgery can cause acute inflammatory response in the body, this method can be used to observe the changes in the IgG N-glycosylation of individuals after acute inflammation. Individual plasma samples were collected before surgery and 24 and 48 hours after surgery, and ultra-high pressure liquid chromatography was used to analyze the IgG N-glycosylation of individual plasma samples at different time points. 1) Data preprocessing, batch effect correction of glycosylation data, and the percentage of different glycosylation structures in total glycosylation as quantitative data, and logarithmic transformation of glycosylation data. 2) The study observed that individuals had changes in glycosylation levels 24 hours after surgery, and the changes continued from 24 to 48 hours, proving that glycosylation is highly sensitive to inflammation in the body. 3) By studying the correlation between glycosylation and the inflammatory indicator C-reactive protein, it was found that the correlation coefficient was 0.4, and the correlation was good.

[0087] Example 3 Construction of a model for sugar-based biological age

[0088] Clinical information and plasma samples of 1157 Chinese people were collected, and high-throughput quantitative detection of plasma IgG N-glycosylation was performed using ultra-high pressure liquid chromatography to obtain individual N-glycosylation data. 1. Data preprocessing: The percentage of different structural glycosylation in the total glycosylation was used as the glycosylation quantitative data, and the data was standardized by Z-Score change, and the data was converted into a normal distribution using logarithmic change. 2. Calculate the individual metabolic syndrome severity index (Metabolic SyndromSeverity Score), systemic immune inflammation index (Systemic immune inflammation index), Framingham cardiovascular disease risk score (Framingham CVD Risk Score), liver fibrosis FIB-4 index (Fibrosis4Score), cerebrovascular disease risk score (Pooled Cohort Stroke Risk Equations), urinary system score (BUN / Creatine Ratio). 3. Use LASSO regression to reduce the dimension of N-glycosylation data, and use N-glycosylation data to fit the prediction model for 6 scores. The fitting model methods include Gaussian process regression and polynomial regression. The predicted values ​​of the 6 scores fitted by the model are used as the patient's glycosylation metabolism index, glycosylation inflammation index, glycosylation cardiovascular index, glycosylation liver index, glycosylation cerebrovascular index, and glycosylation kidney index. 4. The final 6 indexes fitted are added to the individual's calendar age as the individual's biological age. If the individual's glycosylation biological age is less than the calendar age, it proves that the individual is in good health, and the more the biological age is less than the calendar age, the better the individual's health. Conversely, if the patient's biological age is greater than the calendar age, it proves that the individual's aging process is accelerated and the health is poor.

[0089] The specific operations are as follows:

[0090] 2.1 Glycosyl Quantification Data Preprocessing

[0091] The area under the chromatographic peak of each glycan divided by the area under the total chromatographic peak is the level of the glycan, that is, the content of each glycan index is expressed in percentage. After obtaining the glycan data, the batch effect of the glycan data was corrected using the ComBat method of R software. The GlycanR program was used for quantitative calculation of glycan-derived structures, and the glycan data was preprocessed using Z-score changes, logarithmic changes, polynomial changes, etc., so that the data conformed to the normal distribution.

[0092] 2.2 Construction of the “Glyco-based Biological Age System” Scoring System

[0093] Based on the IgG N-glycosylation database of 1157 Chinese Han people accumulated by the research team in the early stage, the glycosyl biological age system score was constructed. The database contains plasma IgG N-glycosylation data, clinical measurement indicators, and lifestyle measurement indicators. The glycosyl metabolic index is calculated based on the metabolic syndrome severity index and plasma IgG N-glycosylation data, the glycosyl liver index is calculated based on the liver fibrosis FIB-4 index and plasma IgG N-glycosylation data, the glycosyl cardiovascular index is calculated based on the Framingham cardiovascular disease risk score and plasma IgG N-glycosylation data, the glycosyl inflammatory index is calculated based on the systemic immune inflammation index and plasma IgG N-glycosylation data, the glycosyl renal index is calculated based on the urea nitrogen creatinine ratio and plasma IgG N-glycosylation data, and the glycosyl stroke index is calculated based on the combined cohort stroke risk equation and plasma IgG N-glycosylation data. Finally, the glycosyl metabolic index, glycosyl liver index, glycosyl cardiovascular index, glycosyl inflammatory index, glycosyl renal index, and glycosyl stroke index were obtained to evaluate health risks from these six dimensions. The final fitted 6 indexes and the individual's calendar age are added together as the individual's glycosyl biological age. If the individual's glycosyl biological age is less than the calendar age, it proves that the individual is in good health, and the more the biological age is less than the calendar age, the better the individual's health. Conversely, if the patient's biological age is greater than the calendar age, it proves that the individual's aging process is accelerated and the health is poor.

[0094] 2.3 Data Dimensionality Reduction Based on LASSO Regression

[0095] Before model building, LASSO regression (Least Absolute Shrinkage and Selection Operator) was first used to reduce the dimensionality of the glycosylation data. LASSO regression is a linear regression method used in statistics and machine learning, mainly used for variable selection and regularization. LASSO will reduce the coefficients of unimportant features to zero, thereby playing a role in variable selection. Moreover, LASSO can prevent the model from overfitting and improve the interpretability of the model. After the 6 system scores were reduced by LASSO regression, the selected glycosylation structures are as follows:

[0096] Glycosylation metabolism index: GP2, GP3, GP4, GP5, GP6, GP9, GP10, GP11, GP12, GP14, GP15, GP16, GP20, GP22, GP24.

[0097] Glycolic inflammatory index: GP1, GP4, GP6, GP8, GP9, GP12, GP15, GP19, GP20, GP22, GP24.

[0098] Glycosyl cardiovascular index: GP1, GP4, GP5, GP6, GP11, GP12, GP15, GP19, GP20, GP21, GP23, GP24.

[0099] Glycosylated stroke index: GP1, GP4, GP5, GP6, GP8, GP9, GP11, GP12, GP15, GP17, GP19, GP20, GP21, GP23, GP24.

[0100] Glycosyl liver index: GP2, GP4, GP5, GP6, GP8, GP9, GP10, GP11, GP12, GP13, GP14, GP19, GP20.

[0101] Glycosylated renal index: GP10, GP11.

[0102] 2.4 Model construction based on Gaussian process regression

[0103] Gaussian Process Regression (GPR) is used for model building. GPR is a probability-based nonparametric regression method used to solve regression problems. It predicts the value of new data points and gives prediction uncertainty by assuming that the data obeys a Gaussian process. Gaussian process regression is particularly suitable for modeling noisy data, nonlinear functions, and small data sets.

[0104] The kernel function used in the GPR model of this project is the radial basis kernel (RBF kernel), and the kernel function is:

[0105]

[0106] The analysis process was based on R language version 4.4.1, and the model was constructed using the gausspr function of the kernlab package. The model parameters were as follows: type = "regression", variance.model = TRUE, cross = 10.

[0107] 2.5 Model construction based on polynomial regression

[0108] Use polynomial regression to build the model. Polynomial regression is a regression analysis method that extends linear regression and is used to fit nonlinear data. Although it is still a linear model, it can capture nonlinear relationships by introducing polynomial features to the independent variables. The polynomial regression model of this project is as follows:

[0109] GlycanAge(male)=51.48+34.97*GP3+26.18*GP5-40.15*GP6 2 +52.74*GP8 2

[0110] +38.50*GP14 2 -34.71*GP16 2 -51.14*GP17 2 -58.50*GP22.

[0111] GlycanAge(female)=49.18–29.55*GP2 2 +28.32*GP3-37.72*GP42–53.04

[0112] *GP8 2 +53.70*GP14 2 +33.68*GP18-48.11*GP20 2 +38.04*GP21.

[0113] Scatter plot and fitting curve of glycosyl biological age based on Gaussian process regression (male) Figure 2 As shown. The scatter plot and fitting curve of glycosyl biological age based on Gaussian process regression (female) are shown in Figure 3 shown.

[0114] Scatter plot and fitting curve of glycosyl biological age based on polynomial regression (male) Figure 4 As shown in Figure 2, the scatter plot of glycosyl biological age based on polynomial regression and the fitting curve (female) are as follows Figure 5 shown.

[0115] Example 4: Evaluation of male biological age by N-glycosylation profile

[0116] GlycanAge(male)=51.48+34.97*GP3+26.18*GP5-40.15*GP6 2 +52.74*GP8 2

[0117] +38.50*GP14 2 -34.71*GP16 2 -51.14*GP17 2 -58.50*GP22.

[0118] Individual A, male, had the following IgG N-glycosylation test results: GP1: 0.08, GP2: 0.6, GP3: 0.27, GP4: 13.32, GP5: 0.26, GP6: 4.47, GP7: 0.87, GP8: 18.54, GP9: 7.69, GP10: 5.94, GP11: 0.59, GP12: 1.9, GP13: 0.61, GP14: 20.49, GP15: 2.3, GP16: 2.15, GP17: 1.5, GP18: 12.06, GP19: 1.79, GP20: 0.5, GP21: 0.82, GP22: 0.14, GP23: 1.73, GP24: 1.38.

[0119] After standardization and logarithmic transformation, the sugar test results were substituted into the above formula to calculate the sugar biological age: 44.3 years old. The individual's calendar age was 56.4 years old, and his biological age was much younger than his calendar age, proving that he was in good health.

[0120] Example 5: Evaluation of female biological age by N-glycosylation profile

[0121] GlycanAge(female)=49.18–29.55*GP2 2 +28.32*GP3-37.72*GP42–53.04

[0122] *GP8 2 +53.70*GP14 2 +33.68*GP18-48.11*GP20 2 +38.04*GP21.

[0123] Individual B, female, had the following IgG N-glycosylation test results: GP1: 0.1, GP2: 0.11, GP3: 0.16, GP4: 11.37, GP5: 2.77, GP6: 4.18, GP7: 0.89, GP8: 10.52, GP9: 9.09, GP10: 3.35, GP11: 1.96, GP12: 0.87, GP13: 0.32, GP14: 14.41, GP15: 1.21, GP16: 3.49, GP17: 5.32, GP18: 12.04, GP19: 2.76, GP20: 0.5, GP21: 1.47, GP22: 1.38, GP23: 2.08, GP24: 9.67.

[0124] After standardization and logarithmic transformation, the sugar test results were substituted into the above formula to calculate the sugar biological age: 63.2 years old. The calendar age of this individual was 48.5 years old. His biological age was much greater than his calendar age, proving that his health condition was poor.

[0125] Example 6 Evaluation Method for Model Prediction Ability

[0126] The root mean square error (RMSE) and R-square were used to evaluate the predictive ability of each model of the glycobiological age system. Variable importance analysis was performed using RMSE as the loss function. The root mean square error is a measure of the difference between the actual observed value and the model predicted value. It is the square root of the mean of the sum of squared residuals. The smaller the RMSE, the smaller the difference between the model's predicted value and the actual value, and the better the model's predictive performance. R-square, also known as the coefficient of determination, is a measure of the proportion of variability explained by the model to the total variability. Its value ranges between 0 and 1, and the closer the value is to 1, the better the model fits the data.

[0127] The RMSE of the GlycanAge (male) model is 9.73, indicating that the average error between the biological age predicted by the model and the actual biological age is about 9.73 years. 2 It is 0.3304, indicating that the model can explain about 33.04% of the variability.

[0128] The RMSE of the GlycanAge (female) model is 6.92, indicating that the average error between the biological age predicted by the model and the actual biological age is about 6.92 years. 2 It is 0.6111, indicating that the model can explain about 61.11% of the variability.

Claims

1. A method for constructing a biological age assessment model based on the N-glycosylation profile of plasma immunoglobulin G, characterized in that: The steps include: 1) Collect blood samples and extract plasma IgG; 2) Detect the N-glycosylation of IgG to obtain the N-glycosylation spectrum; 3) The area under the chromatographic peak of each sugar group is divided by the area under the total chromatographic peak as the horizontal data of the sugar group, and the obtained sugar group data are pre-processed to obtain sugar group data with normal distribution; 4) The glycosylated metabolic index was calculated based on the metabolic syndrome severity index and plasma IgG N-glycosylation data; the glycosylated liver index was calculated based on the liver fibrosis FIB-4 index and plasma IgG N-glycosylation data; the glycosylated cardiovascular index was calculated based on the Framingham cardiovascular disease risk score and plasma IgG N-glycosylation data; the glycosylated inflammatory index was calculated based on the systemic immune inflammation index and plasma IgG N-glycosylation data; the glycosylated renal index was calculated based on the urea nitrogen to creatinine ratio and plasma IgG N-glycosylation data; and the glycosylated stroke index was calculated based on the combined cohort stroke risk equation and plasma IgG N-glycosylation data; 5) The obtained scores of glycosylation metabolism index, glycosylation liver index, glycosylation cardiovascular index, glycosylation inflammation index, glycosylation kidney index and glycosylation stroke index are integrated as the glycosylation biological age score; 6) Use Gaussian process regression (GPR) or polynomial regression to construct a biological age assessment model; Optionally, 7) evaluating the predictive ability of the constructed biological age assessment model.

2. The method for constructing a biological age assessment model according to claim 1, characterized in that: In step 2), high performance liquid chromatography or ultra high performance liquid chromatography is used to obtain the N-glycosylation profile.

3. The method for constructing a biological age assessment model according to claim 1, characterized in that: In step 3), Z-score variation, logarithmic variation and polynomial variation are used to obtain normally distributed sugar data.

4. The method for constructing a biological age assessment model according to claim 1, characterized in that: In step 4), glycosylation data dimensionality reduction is performed based on LASSO regression to obtain glycosylation structures associated with each index.

5. The method for constructing a biological age assessment model according to claim 3, characterized in that: The glycosylation structures associated with the glycosylation metabolism index are: GP2, GP3, GP4, GP5, GP6, GP9, GP10, GP11, GP12, GP14, GP15, GP16, GP20, GP22, GP24; The glycosylated structures associated with the glycosylated inflammatory index are: GP1, GP4, GP6, GP8, GP9, GP12, GP15, GP19, GP20, GP22, GP24; The glycosylated structures associated with the glycosyl cardiovascular index are: GP1, GP4, GP5, GP6, GP11, GP12, GP15, GP19, GP20, GP21, GP23, GP24; The glycosylated structures associated with the glycosylated stroke index are: GP1, GP4, GP5, GP6, GP8, GP9, GP11, GP12, GP15, GP17, GP19, GP20, GP21, GP23, GP24; The glycosylated structures associated with the glycosyl liver index are: GP2, GP4, GP5, GP6, GP8, GP9, GP10, GP11, GP12, GP13, GP14, GP19, GP20; The glycosylated structures associated with the glycosylated renal index are: GP10, GP11, The glycosyl structure is shown in the following table:

6. The method for constructing a biological age assessment model according to claim 1, characterized in that: The biological age evaluation model in step 6) is: Male: GlycanAge = 51.48 + 34.97*GP3 + 26.18*GP5 - 40.15*GP6 2 +52.74*GP8 2 +38.50*GP14 2 -34.71*GP16 2 -51.14*GP17 2 -58.50*GP22; Female: GlycanAge=49.18–29.55*GP2 2 +28.32*GP3-37.72*GP4 2 –53.04*GP8 2 +53.70*GP14 2 +33.68*GP18-48.11*GP20 2 +38.04*GP21; the GP value is the percentage of the area under the chromatographic peak of each structural sugar to the area under the total chromatographic peak; The sugar peak structure corresponding to the GP2 is FA1, the sugar peak structure corresponding to the GP3 is A2B, the sugar peak structure corresponding to the GP4 is FA2, the sugar peak structure corresponding to the GP5 is M5, the sugar peak structure corresponding to the GP6 is FA2B, the sugar peak structure corresponding to the GP8 is FA2[6]G1, the sugar peak structure corresponding to the GP14 is FA2G2, the sugar peak structure corresponding to the GP16 is FA2G1S1, the sugar peak structure corresponding to the GP17 is A2G2S1, the sugar peak structure corresponding to the GP18 is FA2G2S1, the sugar peak structure corresponding to the GP20 is FA2FG2S1, the sugar peak structure corresponding to the GP21 is A2G2S2, and the sugar peak structure corresponding to the GP22 is A2BG2S2.

7. The method for constructing a biological age assessment model according to claim 1, characterized in that: The prediction ability evaluation in step 7) is evaluated using the root mean square error RMSE and R square.

8. A method for evaluating biological age of a sample, characterized in that: The steps include: 1) Collect blood samples and extract plasma IgG; 2) Detect the N-glycosylation of IgG to obtain the N-glycosylation spectrum; 3) The area under the chromatographic peak of each sugar group divided by the area under the total chromatographic peak is the horizontal data of the sugar group; 4) Evaluate the biological age of the sample using a biological age evaluation model, wherein the evaluation model is: Male: GlycanAge = 51.48 + 34.97 * GP3 + 26.18 * GP5 - 40.15 * GP6 2 +52.74*GP8 2 +38.50*GP14 2 -34.71*GP16 2 -51.14*GP17 2 -58.50*GP22; Female: GlycanAge=49.18–29.55*GP2 2 +28.32*GP3-37.72*GP4 2 –53.04*GP8 2 +53.70*GP14 2 +33.68*GP18-48.11*GP20 2 +38.04*GP21; the GP value is the percentage of the area under the chromatographic peak of each structural sugar to the area under the total chromatographic peak; the sugar peak structure corresponding to the GP2 is FA1, the sugar peak structure corresponding to the GP3 is A2B, the sugar peak structure corresponding to the GP4 is FA2, the sugar peak structure corresponding to the GP5 is M5, the sugar peak structure corresponding to the GP6 is FA2B, the sugar peak structure corresponding to the GP8 is FA2[6]G1, the sugar peak structure corresponding to the GP14 is FA2G2, the sugar peak structure corresponding to the GP16 is FA2G1S1, the sugar peak structure corresponding to the GP17 is A2G2S1, the sugar peak structure corresponding to the GP18 is FA2G2S1, the sugar peak structure corresponding to the GP20 is FA2FG2S1, the sugar peak structure corresponding to the GP21 is A2G2S2, and the sugar peak structure corresponding to the GP22 is A2BG2S2.

9. The biological age assessment method according to claim 8, characterized in that: In step 2), high performance liquid chromatography or ultra high performance liquid chromatography is used to obtain the N-glycosylation profile.

10. The use of a biological age assessment model in predicting the degree of individual aging, or assessing the health status of an individual, or assessing the health risk of an individual, characterized in that: The evaluation model is: Male: GlycanAge = 51.48 + 34.97 * GP3 + 26.18 * GP5 - 40.15 * GP6 2 +52.74*GP8 2 +38.50*GP14 2 -34.71*GP16 2 -51.14*GP17 2 -58.50*GP22; Female: GlycanAge=49.18–29.55*GP2 2 +28.32*GP3-37.72*GP4 2 –53.04*GP8 2 +53.70*GP14 2 +33.68*GP18-48.11*GP20 2 +38.04*GP21; the GP value is the percentage of the area under the chromatographic peak of each structural sugar to the area under the total chromatographic peak; the sugar peak structure corresponding to the GP2 is FA1, the sugar peak structure corresponding to the GP3 is A2B, the sugar peak structure corresponding to the GP4 is FA2, the sugar peak structure corresponding to the GP5 is M5, the sugar peak structure corresponding to the GP6 is FA2B, the sugar peak structure corresponding to the GP8 is FA2[6]G1, the sugar peak structure corresponding to the GP14 is FA2G2, the sugar peak structure corresponding to the GP16 is FA2G1S1, the sugar peak structure corresponding to the GP17 is A2G2S1, the sugar peak structure corresponding to the GP18 is FA2G2S1, the sugar peak structure corresponding to the GP20 is FA2FG2S1, the sugar peak structure corresponding to the GP21 is A2G2S2, and the sugar peak structure corresponding to the GP22 is A2BG2S2.

Images