Immunity assessment methods, systems, and applications based on pre-trained models

Abstract

This invention discloses a method, system, and application of immunity assessment based on a pre-trained model. The method of this invention can be used to assess immunity based on biological functions, and can also be used to predict disease outcomes, such as post-infection outcomes.

Description

Technical Field

[0001] This invention belongs to the field of biotechnology, specifically relating to an immune assessment method, system, and application based on a pre-trained model. Background Technology

[0002] Immunity and individual differences in immunity are important variables affecting the severity of symptoms after infection, the presence of sequelae, and the severity of sequelae. However, there is currently no systematic technical framework for assessing immunity and its application in outcomes.

[0003] Therefore, there is an urgent need in this field for an immunity assessment method. Summary of the Invention

[0004] As mentioned above, there is an urgent need in this field for an immunity assessment method.

[0005] Based on a biological pathway database and combined with existing DNA methylation data, the inventors scanned all biological functions and established a pathway aging assessment model for each biological function. Furthermore, they used the aging index of biological functions to predict infection outcomes. The inventors unexpectedly discovered that pathways with strong predictive power primarily involve a series of immunological functions, and these immunological functions are significantly correlated with the proportions of immune cells and lymphocytes. This led to the realization of this invention.

[0006] In a first aspect, the present invention provides a method for assessing immunity, the method comprising the following steps:

[0007] (1) Data acquisition: Acquire aging gene set data, pathway gene set data, and DNA methylation data of subjects containing age information. Pathway refers to a set of genes with biological functions.

[0008] (2) Establish an age prediction model: Using the DNA methylation data, the methylation feature value of the intersection of each pathway gene set and the aging gene set is used as the independent variable, and age is used as the dependent variable. An algorithm is used to construct an age prediction model, and the pathway aging index is calculated using the model. The range of the pathway aging index is 0-1.

[0009] (3) Immunity assessment: Immunity is assessed using the pathway aging index, which includes two aspects:

[0010] Immune aging index = pathway aging index

[0011] Immune vitality index = 1 - pathway aging index.

[0012] In a second aspect, the present invention provides a system for assessing immunity, the system implementing the method of the first aspect, the system comprising:

[0013] The data acquisition module is used to acquire aging gene sets, pathway gene sets, and DNA methylation data of subjects containing age information. A pathway represents a set of genes with biological functions.

[0014] The model building module is used to construct an age prediction model using the DNA methylation data, with the methylation feature value of the intersection of each pathway gene set and the aging gene set as the independent variable and age as the dependent variable, and to calculate the pathway aging index using the model. The pathway aging index is in the range of 0-1.

[0015] The immunity assessment module is used to assess immunity using the pathway aging index. This immunity assessment includes two aspects:

[0016] Immune aging index = pathway aging index

[0017] Immune vitality index = 1 - pathway aging index.

[0018] In a third aspect, the present invention provides a medium for assessing immunity, the medium comprising a procedure for implementing the method of the first aspect, the medium comprising:

[0019] (1) Obtain aging gene set, pathway gene set data, and subject DNA methylation data containing age information. Pathway represents a set of genes with biological functions. (2) Using the DNA methylation data, with the methylation feature value of the intersection of each pathway gene set and the aging gene set as the independent variable and age as the dependent variable, construct an age prediction model using an algorithm. Use the model to calculate the pathway aging index, which is in the range of 0-1.

[0020] (3) Immunity assessment: Immunity is assessed using the pathway aging index, which includes two aspects:

[0021] Immune aging index = pathway aging index

[0022] Immune vitality index = 1 - pathway aging index.

[0023] In a fourth aspect, the present invention provides a method for obtaining phenotype-related pathway aging characteristics, the method comprising:

[0024] (1) Obtain phenotypic data, wherein the phenotypic data is DNA methylation feature value × sample matrix;

[0025] (2) Calculate the pathway aging index of the phenotypic data using the method of any one of claims 1-3, and obtain the pathway aging index × sample matrix;

[0026] (3) Based on the pathway aging index × sample matrix, obtain pathway aging features that are significantly correlated with the phenotype, for example, through machine learning algorithms.

[0027] In a fifth aspect, the present invention provides a system for obtaining phenotype-related pathway aging characteristics, the system comprising:

[0028] A data acquisition device is used to acquire phenotypic data, which is DNA methylation feature values ​​× sample matrix;

[0029] The pathway aging index calculator is used to calculate the pathway aging index of the phenotypic data using the method described in any one of claims 1-3, and obtain the pathway aging index × sample matrix.

[0030] A phenotypic-pathway analyzer is used to obtain pathway aging features that are significantly correlated with the phenotype based on the pathway aging index × sample matrix, for example, through machine learning algorithms.

[0031] In a sixth aspect, the present invention provides a medium for obtaining phenotype-related pathway aging characteristics, the medium comprising a program for implementing the method of the fourth aspect, the medium comprising:

[0032] (1) Obtain phenotypic data, wherein the phenotypic data is DNA methylation feature value × sample matrix;

[0033] (2) Calculate the pathway aging index of the phenotypic data using the method of any one of claims 1-3, and obtain the pathway aging index × sample matrix;

[0034] (3) Based on the pathway aging index × sample matrix, obtain pathway aging features that are significantly correlated with the phenotype, for example, through machine learning algorithms.

[0035] In a seventh aspect, the present invention provides a method for assessing the immunity of an infected subject, the method using 10 pathways to assess immunity using the method described in the first aspect, comprising:

[0036] GO_ALPHA_BETA_T_CELL_PROLIFERATION, pathway aging index = 1.7 + 85*BLM - 25*RASAL3 - 9.5*DOCK2 - 8.9*TGFBR2 + 3.2*CD28 - 0.99*IL18; and / or

[0037] GO_LEUKOCYTE_ACTIVATION_INVOLVED_IN_INFLAMMATO RY_RESPONSE, pathway aging index = 2 + 42 * MIR124 - 3 - 9.2 * TYROBP - 6.4 * ITGAM - 5.3 * CCL3 - 4.5 * AIF1 + 3.1 * C5AR1 - 1 * TLR6; and / or KEGG_NOD_LIKE_RECEPTOR_SIGNALING_PATHWAY, pathway aging index = 15 - 77 * CASP8 + 38 * RELA - 16 * NOD2 + 14 * NLRP3 - 10 * MEFV - 4.1 * TRIP6 - 2.7 * MAPK10 - 2.2 * CASP5 - 0.27 * CARD6; and / or GO_DEACETYLASE_ACTIVITY, pathway aging index = 11 - 27 * ESD + 15 * SIRT1 - 12 * NDST1 - 7 * HDAC9 - 5.5 * CES1; and / or

[0038] KEGG_CHRONIC_MYELOID_LEUKEMIA, pathway aging index = 6.4 + 37*MDM2 - 14*BCR + 12*RELA - 9.7*CDKN1A - 5.4*SMAD3 - 3.3*STAT5A - 1.2*BRAF; and / or

[0039] GO_TRANSCRIPTION_INITIATION_FROM_RNA_POLYMERASE_I_PROMOTER, pathway aging index = -18 + 43*POLR1C - 32*CCNH + 17*POLR2H + 15*TAF1D + 11*GTF2H1 - 8.5*RRN3P1 - 4.9*ERCC3 - 4.3*TAF1C; and / or

[0040] GO_POSITIVE_REGULATION_OF_CYTOKINE_BIOSYNTHETIC_PROCESS, pathway aging index = 5.2 + 53*RELA - 12*CD86 - 6.8*BCL10 + 5.2*CD28 - 4.1*IL12B - 3.6*TYROBP + 2.8*ELANE - 0.54*MAPKAPK2; and / or

[0041] GO_T_CELL_SELECTION, pathway aging index = 7.8 + 55 * GLI3 + 26 * SRF - 17 * IL12RB1 - 15 * IL12B - 13 * AIRE + 6.5 * CD28 - 2.1 * DOCK2; and / or GO_POSITIVE_REGULATION_OF_ALPHA_BETA_T_CELL_ACTI VATION, pathway aging index = 7.3 + 91 * GLI3 - 37 * CD83 - 25 * CBFB + 18 * AP3B1 + 17 * BLM - 17 * RASAL3 - 15 * IL12RB1 - 13 * NCKAP1L - 12 * IL6R - 6.3 * TGFBR2 - 1.8 * IL12B + 0.57 * CD28 - 0.38 * CD86; and / or GO_REGULATION_OF_RAS_PROTEIN_SIGNAL_TRANSDUCTIO N, pathway aging index = 3.2 + 54 * TRIM67 - 47 * MET - 37 * CYTH1 + 32 * RAC1 + 27 * AUTS2 - 26 * CSF1 - 25 * SQSTM1 - 25 * RDX + 25 * RASGRF2 - 21 * BC R-19*MAPKAP1+18*ITGA3-16*ABR+12*FLCN+8.4*ABL2+6.5*RASGEF1A-4.8*ARFGEF2+4.6*PSD3+4.6*KRAS+4.4*RASAL3+3.1*FA RP2-2.6*GPR20+1.1*PLEKHG6+0.52*EPHB2+0.057*MCF2L, where gene represents the DNA methylation characteristic value of the gene.

[0042] In an eighth aspect, the present invention provides a system for assessing the immunity of an infected subject, the system implementing the method described in the seventh aspect, the system comprising: a data acquisition module for acquiring DNA methylation characteristic value data of the subject;

[0043] The calculation module is used to calculate the pathway aging index based on the subject's DNA methylation characteristic data to assess immunity.

[0044] In a ninth aspect, the present invention provides a medium for assessing the immunity of an infected subject, the medium comprising a procedure for implementing the method of the seventh aspect, the medium comprising:

[0045] (1) Read the DNA methylation signature data of the subjects;

[0046] (2) Calculate the pathway aging index based on the DNA methylation characteristic value data of the subjects to assess immunity.

[0047] In a tenth aspect, the present invention provides a kit for assessing the immunity of an infected subject, the kit being used to detect the DNA methylation level of the genes described in the seventh, eighth or ninth aspects.

[0048] The beneficial effects of the present invention are that the method of the present invention can be used to assess immunity using biological functions, and can also be used to predict disease outcomes, such as the outcome after infection. Attached Figure Description

[0049] To more clearly illustrate the technical solutions in the specific embodiments of the present invention, the accompanying drawings in the specific embodiments will be briefly described below.

[0050] Figure 1 A flowchart of the method of the present invention and its specific applications are shown.

[0051] Figure 2 The distribution of the three pathway aging indices with the most significant differences between the severe and mild cases in Data 1 of Embodiment 2 of the present invention is shown.

[0052] Figure 3 The data in Data 2 of Embodiment 2 of the present invention shows the distribution of the three pathway aging indices with the most significant differences between the inpatient treatment group and the outpatient group.

[0053] Figure 4 The ROC curves of the model trained using the lasso machine learning algorithm in Embodiment 2 of the present invention are shown in the training set and the independent validation set.

[0054] Figure 5 The risk profiles of three individuals (including one mild case and two severe cases) in Embodiment 3 of the present invention are shown. Detailed Implementation

[0055] The present invention will be described in detail below with reference to the accompanying drawings. It should be understood that the following description is merely illustrative and is not intended to limit the scope of the invention; the scope of protection of the invention is defined by the appended claims. Furthermore, those skilled in the art will understand that modifications can be made to the technical solutions of the present invention without departing from its spirit and intent. Unless otherwise specified, the technical means used in the embodiments are conventional means well known to those skilled in the art.

[0056] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which the subject matter pertains. Before a detailed description of the invention, the following definitions are provided to better understand it.

[0057] In the context of this invention, many embodiments use the expressions "comprising," "including," or "basically / mainly composed of." The expressions "comprising," "including," or "basically / mainly composed of" are generally understood as open-ended expressions, indicating that they include not only the elements, components, parts, or method steps specifically listed after the expression, but also other elements, components, parts, or method steps. However, in this document, the expressions "comprising," "including," or "basically / mainly composed of" can also be understood as closed-ended expressions in certain cases, indicating that they only include the elements, components, parts, or method steps specifically listed after the expression, and do not include any other elements, components, parts, or method steps. In this case, the expression is equivalent to the expression "composed of."

[0058] In this paper, the term "methylation" refers to a form of DNA chemical modification that can alter genetic expression without changing the DNA sequence. DNA methylation refers to the covalent binding of a methyl group to the 5th carbon position of cytosine in a CpG dinucleotide of the genome by DNA methyltransferases. DNA methylation can cause changes in chromatin structure, DNA conformation, DNA stability, and the way DNA interacts with proteins, thereby controlling gene expression.

[0059] Therefore, in a first aspect, the present invention provides a method for assessing immunity, the method comprising the following steps:

[0060] (1) Data acquisition: Acquire aging gene set data, pathway gene set data, and DNA methylation data of subjects containing age information. Pathway refers to a set of genes with biological functions.

[0061] (2) Establish an age prediction model: Using the DNA methylation data, the methylation feature value of the intersection of each pathway gene set and the aging gene set is used as the independent variable, and age is used as the dependent variable. An algorithm is used to construct an age prediction model, and the pathway aging index is calculated using the model. The range of the pathway aging index is 0-1.

[0062] (3) Immunity assessment: Immunity is assessed using the pathway aging index, which includes two aspects:

[0063] Immune aging index = pathway aging index

[0064] Immune vitality index = 1 - pathway aging index.

[0065] The aging gene set is a collection of genes related to aging, which can be obtained in the following ways: by searching databases such as the Aging Atlas database (https: / / ngdc.cncb.ac.cn / aging / index) and the Cellage database (https: / / genomics.senescence.info / cells / ); by comparing transcriptome data, protein data, methylation data, etc., of young and aging groups; and by collecting aging-related genes through literature review.

[0066] In one embodiment, the methylation characteristic is the average of the methylation beta values ​​or the SIMPO value of all methylation sites in the gene body region. In a preferred embodiment, the methylation characteristic is the average of the methylation beta values ​​or the SIMPO value of all methylation sites in the gene promoter region.

[0067] The SIMPO value is calculated as follows: The inventors used a previously developed SIMPO method for EWAS (Epigenome-wide Association Study) research. This method converts the beta value of methylation sites into a SIMPO value for a gene. The principle of this method is that the difference in DNA methylation between the gene body and the promoter region is significantly correlated with gene expression, with a correlation coefficient as high as 0.67, suggesting that the difference in methylation between the gene body and the promoter region is a biologically significant feature that can be used to predict gene expression. Specifically, high-order features at the gene level are extracted based on the low-level CpG site methylation data. After this conversion, the beta values ​​of multiple methylation sites in a gene are transformed into a single SIMPO value for each gene, which is correlated with the gene's expression activity.

[0068] In one implementation, the number of genes in the intersection of the pathway gene set and the aging gene set ranges from 6 to 50.

[0069] In a second aspect, the present invention provides a system for assessing immunity, the system implementing the method of the first aspect, the system comprising:

[0070] The data acquisition module is used to acquire aging gene sets, pathway gene sets, and DNA methylation data of subjects containing age information. A pathway represents a set of genes with biological functions.

[0071] The model building module is used to construct an age prediction model using the DNA methylation data, with the methylation feature value of the intersection of each pathway gene set and the aging gene set as the independent variable and age as the dependent variable, and to calculate the pathway aging index using the model. The pathway aging index is in the range of 0-1.

[0072] The immunity assessment module is used to assess immunity using the pathway aging index. This immunity assessment includes two aspects:

[0073] Immune aging index = pathway aging index

[0074] Immune vitality index = 1 - pathway aging index.

[0075] In a third aspect, the present invention provides a medium for assessing immunity, the medium comprising a procedure for implementing the method of the first aspect, the medium comprising:

[0076] (1) Obtain aging gene set, pathway gene set data, and subject DNA methylation data containing age information. Pathway represents a set of genes with biological functions. (2) Using the DNA methylation data, with the methylation feature value of the intersection of each pathway gene set and the aging gene set as the independent variable and age as the dependent variable, construct an age prediction model using an algorithm. Use the model to calculate the pathway aging index, which is in the range of 0-1.

[0077] (3) Immunity assessment: Immunity is assessed using the pathway aging index, which includes two aspects:

[0078] Immune aging index = pathway aging index

[0079] Immune vitality index = 1 - pathway aging index.

[0080] In a fourth aspect, the present invention provides a method for obtaining phenotype-related pathway aging characteristics, the method comprising:

[0081] (1) Obtain phenotypic data, wherein the phenotypic data is DNA methylation feature value × sample matrix;

[0082] (2) Calculate the pathway aging index of the phenotypic data using the method of any one of claims 1-3, and obtain the pathway aging index × sample matrix;

[0083] (3) Based on the pathway aging index × sample matrix, obtain pathway aging features that are significantly correlated with the phenotype, for example, through machine learning algorithms.

[0084] In a fifth aspect, the present invention provides a system for obtaining phenotype-related pathway aging characteristics, the system comprising:

[0085] A data acquisition device is used to acquire phenotypic data, which is DNA methylation feature values ​​× sample matrix;

[0086] The pathway aging index calculator is used to calculate the pathway aging index of the phenotypic data using the method described in any one of claims 1-3, and obtain the pathway aging index × sample matrix.

[0087] A phenotypic-pathway analyzer is used to obtain pathway aging features that are significantly correlated with the phenotype based on the pathway aging index × sample matrix, for example, through machine learning algorithms.

[0088] In a sixth aspect, the present invention provides a medium for obtaining phenotype-related pathway aging characteristics, the medium comprising a program for implementing the method of the fourth aspect, the medium comprising:

[0089] (1) Obtain phenotypic data, wherein the phenotypic data is DNA methylation feature value × sample matrix;

[0090] (2) Calculate the pathway aging index of the phenotypic data using the method of any one of claims 1-3, and obtain the pathway aging index × sample matrix;

[0091] (3) Based on the pathway aging index × sample matrix, obtain pathway aging features that are significantly correlated with the phenotype, for example, through machine learning algorithms.

[0092] In a seventh aspect, the present invention provides a method for assessing the immunity of an infected subject, the method using 10 pathways to assess immunity using the method described in the first aspect, comprising:

[0093] GO_ALPHA_BETA_T_CELL_PROLIFERATION, pathway aging index = 1.7 + 85*BLM - 25*RASAL3 - 9.5*DOCK2 - 8.9*TGFBR2 + 3.2*CD28 - 0.99*IL18; and / or

[0094] GO_LEUKOCYTE_ACTIVATION_INVOLVED_IN_INFLAMMATO RY_RESPONSE, pathway aging index = 2 + 42 * MIR124 - 3 - 9.2 * TYROBP - 6.4 * ITGAM - 5.3 * CCL3 - 4.5 * AIF1 + 3.1 * C5AR1 - 1 * TLR6; and / or KEGG_NOD_LIKE_RECEPTOR_SIGNALING_PATHWAY, pathway aging index = 15 - 77 * CASP8 + 38 * RELA - 16 * NOD2 + 14 * NLRP3 - 10 * MEFV - 4.1 * TRIP6 - 2.7 * MAPK10 - 2.2 * CASP5 - 0.27 * CARD6; and / or GO_DEACETYLASE_ACTIVITY, pathway aging index = 11 - 27 * ESD + 15 * SIRT1 - 12 * NDST1 - 7 * HDAC9 - 5.5 * CES1; and / or

[0095] KEGG_CHRONIC_MYELOID_LEUKEMIA, pathway aging index = 6.4 + 37*MDM2 - 14*BCR + 12*RELA - 9.7*CDKN1A - 5.4*SMAD3 - 3.3*STAT5A - 1.2*BRAF; and / or

[0096] GO_TRANSCRIPTION_INITIATION_FROM_RNA_POLYMERASE_I_PROMOTER, pathway aging index = -18 + 43*POLR1C - 32*CCNH + 17*POLR2H + 15*TAF1D + 11*GTF2H1 - 8.5*RRN3P1 - 4.9*ERCC3 - 4.3*TAF1C; and / or

[0097] GO_POSITIVE_REGULATION_OF_CYTOKINE_BIOSYNTHETIC_PROCESS, pathway aging index = 5.2 + 53*RELA - 12*CD86 - 6.8*BCL10 + 5.2*CD28 - 4.1*IL12B - 3.6*TYROBP + 2.8*ELANE - 0.54*MAPKAPK2; and / or

[0098] GO_T_CELL_SELECTION, pathway aging index = 7.8 + 55 * GLI3 + 26 * SRF - 17 * IL12RB1 - 15 * IL12B - 13 * AIRE + 6.5 * CD28 - 2.1 * DOCK2; and / or GO_POSITIVE_REGULATION_OF_ALPHA_BETA_T_CELL_ACTI VATION, pathway aging index = 7.3 + 91 * GLI3 - 37 * CD83 - 25 * CBFB + 18 * AP3B1 + 17 * BLM - 17 * RASAL3 - 15 * IL12RB1 - 13 * NCKAP1L - 12 * IL6R - 6.3 * TGFBR2 - 1.8 * IL12B + 0.57 * CD28 - 0.38 * CD86; and / or GO_REGULATION_OF_RAS_PROTEIN_SIGNAL_TRANSDUCTIO N, pathway aging index = 3.2 + 54 * TRIM67 - 47 * MET - 37 * CYTH1 + 32 * RAC1 + 27 * AUTS2 - 26 * CSF1 - 25 * SQSTM1 - 25 * RDX + 25 * RASGRF2 - 21 * BC R-19*MAPKAP1+18*ITGA3-16*ABR+12*FLCN+8.4*ABL2+6.5*RASGEF1A-4.8*ARFGEF2+4.6*PSD3+4.6*KRAS+4.4*RASAL3+3.1*FA RP2-2.6*GPR20+1.1*PLEKHG6+0.52*EPHB2+0.057*MCF2L, where gene represents the DNA methylation characteristic value of the gene.

[0099] This section uses the GO_T_CELL_SELECTION pathway aging index as an example to illustrate the method for assessing the immunity of infected subjects. Specifically, the pathway aging index is calculated using the DNA methylation data of two subjects, A and B, using the formula: “pathway aging index = 7.8 + 55 * GLI3 + 26 * SRF - 17 * IL12RB1 - 15 * IL12B - 13 * AIRE + 6.5 * CD28 - 2.1 * DOCK2”. If subject A's pathway aging index is greater than subject B's, then subject A's immunity is considered lower than subject B's, and subject B's recovery time after infection is shorter than subject A's.

[0100] In an eighth aspect, the present invention provides a system for assessing the immunity of an infected subject, the system implementing the method described in the seventh aspect, the system comprising: a data acquisition module for acquiring DNA methylation characteristic value data of the subject;

[0101] The calculation module is used to calculate the pathway aging index based on the subject's DNA methylation characteristic data to assess immunity.

[0102] In a ninth aspect, the present invention provides a medium for assessing the immunity of an infected subject, the medium comprising a procedure for implementing the method of the seventh aspect, the medium comprising:

[0103] (1) Read the DNA methylation signature data of the subjects;

[0104] (2) Calculate the pathway aging index based on the DNA methylation characteristic value data of the subjects to assess immunity.

[0105] In a tenth aspect, the present invention provides a kit for assessing the immunity of an infected subject, the kit being used to detect the DNA methylation level of the genes described in the seventh, eighth or ninth aspects.

[0106] Figure 1A complete data analysis and application process is described. The process includes the following steps: (1) inputting pathway sets and aging-related DNA methylation data to obtain a pathway aging index modeler; (2) inputting phenotypic (disease)-related omics data and using the pathway aging index modeler in step (1) to obtain phenotypic (disease)-specific pathway aging features. The above phenotypic (disease)-specific features can be used to obtain biomarkers, rank disease treatment targets, etc. In addition, the above process can also be applied to the calculation of individual disease risk, thereby obtaining individual disease treatment plans. In a specific implementation plan, after screening out phenotypic (disease)-specific features using the above method, they are stored in a cloud server. Data of the subjects is obtained through the individual methylation data acquisition module (e.g., collecting DNA methylation feature data from the subjects' saliva). Other symptoms of the subjects are obtained through the human-computer dialogue module. Personalized intervention measures (treatment plans) are obtained using the features stored in the server and the calculation program.

[0107] Example

[0108] It should be noted that the terminology used in this specification is for the purpose of describing specific embodiments only and is not intended to limit the invention. The foregoing summary section and the following detailed description are for illustrative purposes only and are not intended to limit the invention in any way. The scope of the invention is defined by the appended claims without departing from its spirit and intent.

[0109] Example 1: Establishment of a pathway age clock

[0110] Biological function (pathway) data were obtained from gene function annotations in the KEGG pathway, Reactome pathway, and GeneOntology databases. Pathways with ≥20 associated genes were selected for subsequent analysis.

[0111] This embodiment uses an age-related DNA methylation dataset, which is human blood cell DNA methylation data containing age information, sourced from the GEO database, accession number GSE40279 (https: / / www.ncbi.nlm.nih.gov / geo / query / acc.cgi?acc=GSE40279). The data includes whole blood DNA methylation data from 656 subjects of different ages. A DNA methylation microarray detected DNA methylation at 485,512 cytosine sites in the human genome, the vast majority of which (482,421) are CpG sites. The methylation level at each methylation site is expressed as a beta value, representing the proportion of cytosine methylated at a specific site. Multiple methylation sites are distributed within the same functional region of the same gene in the microarray, with different methylation values ​​detected at different sites. The methylation values ​​in the microarray can be normalized according to the manufacturer's instructions.

[0112] Obtain the methylation signature of each gene

[0113] Since each gene has multiple methylation sites, the methylation eigenvalue of each gene is calculated based on the normalized methylation chip data (CpG sites × Sample), resulting in a methylation eigenvalue map of the subjects (Gene × Sample). Specifically, the average methylation beta value of all methylation sites in the gene promoter region is taken as the methylation eigenvalue of that gene. This transformation converts the beta values ​​of multiple methylation sites in a gene into a single eigenvalue for each gene, resulting in the Gene × Sample matrix. Subsequent calculations are then performed using the more than 13,000 genes involved in the methylation chip.

[0114] The methylation characteristic values ​​of all genes were calculated using a t-test to determine the degree of difference between the elderly group (age ≥ 60 years) and the control group (age < 60 years). The genes were sorted in ascending order according to the p-value of the t-test, and the top 3000 genes were selected as the set of genes related to aging for subsequent calculations.

[0115] Filter the pathway

[0116] For each pathway, the intersection of its associated gene with the aforementioned aging-related genes (3000 in total) is calculated. Only pathways with an intersection of ≥6 genes are included in subsequent calculations. Pathways containing too many genes have overly broad functions; to control pathway specificity, only pathways with an intersection of ≤50 genes are included in the subsequent analysis. The remaining 2840 pathways are then included in the subsequent analysis.

[0117] For each pathway, an aging prediction model was built using the Lasso machine learning algorithm. The 656 samples were divided into two-thirds as the training set and the remaining one-third as the model validation set. In the training set, x represents the methylation signature (i.e., the average methylation signature of the promoter region) of all aging-related genes in that pathway, and y represents the aging classification of the sample (age ≥ 60 years, classification code 1; age < 60 years, classification code 0). The Lasso algorithm (leastabsolute shrinkage and selection operator) is a commonly used method in bioinformatics for building classification models. This algorithm aims to screen variables to reduce model complexity and avoid overfitting. After variable optimization, only variables that significantly contribute to the model are retained in the final prediction model.

[0118] In this embodiment, a total of 766 aged linear fitting models for pathways were established. The predictive classification performance of the models was validated on the validation set, and the AUC values ​​were calculated. The top 10 pathway models were sorted in descending order of AUC value, and the establishment status is shown in Table 1.

[0119] Table 1: Pathway aging prediction models with top 10 AUC values

[0120]

[0121]

[0122]

[0123] In Table 1, the first column is the pathway name; the second column is the number of genes belonging to the pathway; the third column is the number of predictive variables selected after lasso modeling, which is also the number of genes included in the model parameters in the fifth column; the fourth column is the AUC value of the model in the test set; and the fifth column is the model parameters, i.e., the final linear model. The value before the gene name is the weighted value of the gene's methylation feature value. The larger the absolute value of this value, the greater its contribution to the prediction model.

[0124] Example 2: Predicting Outcomes Based on Biological Functional Aging

[0125] Using the biological functional aging model in Example 1, the degree of aging of each biological function can be obtained for each individual by substituting the data into the model. The predictive power of the observed degree of biological functional aging on the outcome was calculated in two (SARS-CoV-2) related whole blood DNA methylation datasets.

[0126] Data 1, GSE179325 (https: / / www.ncbi.nlm.nih.gov / geo / query / acc.cgi?acc=GSE179325), includes whole blood DNA methylation microarray data from 473 SARS-CoV-2 positive patients and 101 negative subjects. Among the positive patients, there were 360 ​​mild cases and 113 severe cases. Patients were classified according to the World Health Organization Clinical Rating Scale (WHO COVID-19 Therapeutic Trial Synopsis. R&D Blueprint (2020)). Mild cases were defined as those with a rating of 1-4, and severe cases were defined as those with a rating of 5-8 or those who died. DNA methylation of whole blood samples was detected using the Illumina Infinium Human Methylation EPIC BeadChip.

[0127] Data 2, GSE167202 (https: / / www.ncbi.nlm.nih.gov / geo / query / acc.cgi?acc=GSE167202), includes whole blood DNA methylation microarray data from 525 individuals. Of these, 164 were infected, 296 were not infected, and 65 had other infections. Based on the severity score (SS), the groups were categorized as follows: 0 for uninfected; 1 for home isolation; 2 for hospitalization; 3 for ICU admission; and 4 for death. DNA methylation of whole blood samples was detected using the Illumina Infinium Human Methylation EPIC BeadChip.

[0128] The data analysis process is as follows:

[0129] Gene feature extraction. DNA methylation microarrays detect DNA methylation at 485,512 cytosine sites in the human genome, the vast majority (482,421) of which are CpG sites. The methylation level of each methylation site is expressed as a beta value, representing the proportion of cytosine methylated at a specific site. Multiple methylation sites are distributed within the same functional region of the same gene in the microarray, and different sites will show different methylation values. The methylation values ​​in the microarray are normalized according to the manufacturer's instructions. Since each gene has multiple methylation sites, the methylation feature value of each gene is calculated based on the above-mentioned normalized methylation microarray data, i.e., methylation sites × subject sample matrix (CpG sites × Sample), to obtain the subject's methylation feature value map, i.e., gene × subject sample matrix (Gene × Sample). The average methylation beta value of all methylation sites in the promoter region is taken as the methylation feature value of that gene. After this transformation, the beta value of multiple methylation sites in a gene is transformed into a single feature value for each gene, forming a gene × sample matrix.

[0130] The pathway aging index was calculated based on the 766 pathway aging linear fitting models established in Example 1. The input x of the model is the methylation characteristic value of the gene (the average methylation characteristic value of the methylation site in the promoter region), and the output y is a score, which is used here as a value characterizing the degree of aging.

[0131] In Data 1, methylation characteristics of 473 SARS-CoV-2 positive patients were used as input, substituted into 766 pathway aging models. After the above calculations, the methylation data were transformed into a matrix of 766 pathway aging indices × 473 patients. For each pathway, the differences in pathway aging indices between different phenotypes were calculated. The differences in the 766 pathway aging indices between the severe and mild groups were calculated using a t-test. The top 40 records with the most significant differences between the severe and mild groups, sorted in ascending order of the p-value of the t-test, are shown in Table 2.

[0132] Table 2: Top 40 records with the most significant differences between the severe and mild cases in Data 1

[0133]

[0134]

[0135]

[0136]

[0137]

[0138] Table 2 lists the pathway names in column 1, the t-test p-values ​​in column 2 (a smaller p-value indicates a higher degree of significance in the pathway aging index between the severe and mild cases), the odds ratio (OR) in column 3, and the model parameters in column 4. The Odds Ratio is calculated to evaluate the classification performance (i.e., predictive ability) of each pathway aging index on the phenotype (severe vs. mild). For a given aging index, let x be the pathway aging index variable of 473 samples (360 mild cases and 113 severe cases), and xm be the mean of x. A simple predictor is set up that predicts cases with x values ​​higher than the mean xm as severe cases and cases with x values ​​lower than the mean xm as mild cases. The Odds Ratio can then be calculated to evaluate the performance of this predictor. Specifically, let a be the number of subjects predicted to have severe illness and actually having severe illness; b be the number of subjects predicted to have severe illness and actually having mild illness; c be the number of subjects predicted to have mild illness and actually having severe illness; and d be the number of subjects predicted to have mild illness and actually having mild illness. Then the odds ratio (OR value) = (a / b) / (c / d).

[0139] In Table 2, all OR values ​​are greater than 1, indicating that the higher the pathway aging index, the higher the risk of severe illness. The further the odds ratio is from 1, the stronger the predictive ability of the predictor. Taking the first row of Table 2 as an example, the pathway name is KEGG_CHRONIC_MYELOID_LEUKEMIA, which is a leukemia-related pathway. Its hazard ratio is 5.6, indicating that if a leukemia-related pathway exhibits a characteristic pattern similar to that seen in aging individuals, i.e., the pathway aging index is high, then these individuals have a 5.6 times higher risk of developing severe illness than other individuals. The pathway in the third row of Table 2 is GO_ALPHA_BETA_T_CELL_PROLIFERATI ON, a biological process related to alpha beta T cell proliferation. Aging of this process may also lead to an increased risk of severe illness, with a hazard ratio of 3.3.

[0140] The distribution of the three pathway aging indices (KEGG_CHRONIC_MYELOID_LEUKEMIA, GO_CELLULAR_RESPONSE_TO_GAMMA_RADIATION, GO_ALPHA_BETA_T_CELL_PROLIFERATI ON) that showed the most significant differences between the severe and mild cases is as follows: Figure 2 As shown.

[0141] Similarly, in Data 2, using the methylation feature values ​​of 525 samples as input, and substituting them into a model of 766 pathway aging indices, the methylation data was transformed into a matrix of 766 pathway aging indices × 525 samples after the above calculations. For each pathway, the difference in pathway aging index between the outpatient and inpatient groups was calculated. This difference reflects that among positive patients with the same infection, some patients have more severe symptoms and require hospitalization, while others have milder symptoms and can recover at home. Therefore, this is equivalent to exploring whether the detection and calculation of pathway aging indices can predict mild or severe cases after infection.

[0142] The differences in 766 pathway aging indices between the inpatient and outpatient groups were calculated using a t-test. The top 40 records with the most significant differences between the inpatient and outpatient groups are listed in ascending order of the p-value of the t-test.

[0143] Table 3: Top 40 records with the most significant differences between the inpatient and outpatient groups in Data 2

[0144]

[0145]

[0146]

[0147]

[0148]

[0149] The pathway aging index with the strongest predictive ability in Data 2 is REACTOME_ANTIVIRAL_MECHANISM_BY_IFN_STIMULATED_GENES, which is the antiviral mechanism activated by interferon, with a predictive risk ratio of 5.3.

[0150] The distribution of the three pathway aging indices (REACTOME_ANTIVIRAL_MECHANISM_BY_IFN_STIMULATED_GENES, GO_RESPONSE_TO_ESTROGEN, GO_PROTEIN_LOCALIZATION_TO_GOLGI_APPARATUS) that showed the most significant differences between the inpatient and outpatient groups is as follows: Figure 3 As shown.

[0151] The intersection of pathway aging indices that showed significant differences between the severe group and the mild group in Data 1 and between the inpatient group and the outpatient group in Data 2 (i.e., those with outcome predictive ability) was obtained (a total of 27 indices), as shown in Table 4.

[0152] Table 4: Intersection of pathway aging indices with outcome predictive power in Data 1 and Data 2

[0153]

[0154]

[0155]

[0156]

[0157] Using the Lasso machine learning algorithm, with Data 1 as the training set and Data 2 as the independent validation dataset, a predictive model was built and validated. An optimized predictive model was obtained, which has cross-dataset predictive capabilities. Its specific parameters are shown in Table 5. After variable optimization, a total of 10 pathway aging indices were retained in the final predictive model. The last column of Table 5 shows the weighted value of each pathway aging index in the final linear model; the larger the absolute value, the greater its contribution to the predictive model. Table 5 is sorted in descending order of the absolute value of the weighted values.

[0158] Table 5: Outcome prediction models and parameters with cross-data prediction capabilities

[0159]

[0160]

[0161] As shown in Table 5

[0162] As shown, in the model that integrates multiple pathway aging indices for outcome prediction, after feature selection using the lasso algorithm, the remaining variables show further enrichment of immune-related biological processes, such as leukocyte activation and T cell selection processes related to inflammation response.

[0163] The model exhibits good generalization ability on both the training and independent validation datasets; that is, the model obtained on the training set does not show a significant performance degradation on the independent validation dataset. Figure 4 As shown, the model has an AUC of 0.81 on the training set and an AUC of 0.78 on the independent validation set.

[0164] Example 3: Calculating the PA hazard value of an individual's impact outcome

[0165] Example 2 identified risk factors influencing the outcome, namely, aging indices of multiple biological functions. Different factors have different magnitudes of risk impact. From an individual perspective, it is necessary to rank the various influencing factors to suggest potential intervention and prevention strategies. Therefore, this example fitted the relationship between each risk factor (pathway aging index) and the phenotype (severe or mild) to make the magnitude of the impact of each risk factor on the current individual comparable.

[0166] The risk factors analyzed in this embodiment are the 10 pathways listed in Table 4. For each pathway, using the aging index of 360 mild cases and 113 severe cases from Data 1 as X, and the phenotypic label Y (Y=1 for severe cases, Y=0 for mild cases), a logistic regression model is established with X as the independent variable and Y as the predictor variable. For a case i, its aging index of the 10 pathways is substituted into the fitted model, and the result is calculated as the severe case risk value (PA risk value).

[0167] Specifically, for each sample, each pathway is enumerated, and a severe illness risk value is calculated for each pathway. That is, for this sample, each pathway serves as the evaluation model, with each pathway corresponding to a risk value. In other words, the risk is calculated across 10 pathway dimensions within a given individual. The purpose of this calculation is to normalize the impact of each pathway within that individual. Through logistic regression fitting, the effect of each pathway is normalized to a value between 0 and 1, making the aging characteristics represented by different pathways comparable. This allows for ranking and comparison of the severe illness risk caused by aging characteristics within each individual, essentially creating a risk map for each individual.

[0168] Figure 5The risk profiles for three individuals are shown. The left subplot represents the risk profile of case 5 in the mild group, where the risk values ​​for each pathway are relatively low, mostly <0.2. The middle subplot represents the risk profile of case 12 in the severe group, and the right subplot represents the risk profile of case 30 in the severe group. Overall, the risk values ​​in these two subplots are higher than those in the left subplot (mild group), with some risk values ​​exceeding 0.6.

[0169] For case 12 in the severe illness group, the aging characteristic with the highest risk of severe illness was GO_ALPHA_BETA_T_CELL_PROLIFERATION. For case 30 in the severe illness group, the highest risk value was GO_REGULATION_OF_RAS_PROTEIN_SIGNAL_TR ANSDUCTION, followed by GO_T_CELL_SELECTION. This indicates that even within the severe illness group, the most influential pathways differ among individuals. This also demonstrates that this invention can quantify and decompose the pathway aging index (specifically, the degree of aging of immune-related biological functions) for each individual, providing a reference for interventions targeting immune function.

[0170] Example 4: Using a DNA methylation capture sequencing kit to detect biological aging in saliva samples.

[0171] To address the aforementioned need for detecting biological functional aging, this embodiment selected 680 genes to be detected (including genes involved in the pathway aging model in Table 4) and designed a liquid-phase chip reagent for sequence capture. The target region is the promoter region of the 680 genes (the first 1500 base pairs before the transcription start site, i.e., TSS 1500). Through probe design and synthesis, a liquid-phase chip for capturing this target region was prepared. This reagent contains a total of 35,600 probes, enabling sequence capture of methylation sites in the promoter regions of the aforementioned 680 genes. Compared to whole-genome methylation detection, the sequence capture design significantly reduces reagent and sequencing costs, making this reagent potentially valuable for widespread application.

[0172] Saliva samples were collected from two infected subjects using a room-temperature saliva collection kit manufactured by Wuhan Changmei. After nucleic acid extraction and quality control, 500 ng of genomic DNA was collected and purified using the ZYMO Research EZ DNA Methylation Gold™ Kit according to the kit instructions. The pre-constructed probe solution was thawed, and after library hybridization, capture, elution, and purification, the DNB rapid preparation kit was used to prepare the sequencing reagents. Sequencing was performed using a BGI Genomics DNBSEQ-T7 sequencer. Sequencing data was visualized using FastQC software, and sequencing data quality control was performed, including the removal of sequencing adapters and low-quality sequencing bases. After quality control, the data was aligned with a reference genome, and the alignment efficiency was calculated using bismark. The deduplicate_bismark tool was used to remove PCR redundancy and the redundancy of each sample was calculated.

[0173] The risk value of the aging index in the optimized model of Example 2 was calculated, and the results are shown in Table 6.

[0174] Table 6: Calculation results of aging index risk values ​​for 2 subjects

[0175]

[0176]

[0177] Column 4 of Table 6 shows the aging index risk value for Subject 1, and column 5 shows the aging index risk value for Subject 2. The aging index risk value ranges from 0 to 1, and the calculation method is described in Example 3. Based on the combined results of the 10 aging indices, Subject 1's risk of severe illness was lower than Subject 2's (except for KEGG_CHRONIC_MYELOID_LEUKEMIA, 0.11 vs 0.10). Follow-up results showed that, in the same living environment, Subject 1's recovery time was 3 days, and Subject 2's recovery time was 7 days. The diary entry from the first appearance of symptoms to the disappearance of major symptoms is recorded as "Recovery Time (days)". This demonstrates that the analysis results based on the DNA methylation-based aging index are consistent with the follow-up results.

[0178] The above provides a detailed description of the immunity assessment method, system, and applications based on pre-trained models provided by this invention. Specific embodiments have been used to illustrate the principles and implementation methods of this invention. The descriptions of the embodiments above are only for the purpose of helping to understand the method and core ideas of this invention. At the same time, for those skilled in the art, there will be changes in the specific implementation methods and application scope based on the ideas of this invention. Therefore, the content of this specification should not be construed as a limitation of this invention.

Claims

1. A system for assessing the immunity of subjects infected with SARS-CoV-2 virus, characterized in that, The system consists of the following 10 linear models of pathway aging indices: GO_ALPHA_BETA_T_CELL_PROLIFERATION, pathway aging index = 1.7 + 85 * BLM - 25 * RASAL3 - 9.5 * DOCK2 - 8.9 * TGFBR2 + 3.2 * CD28 - 0.99 * IL18; GO_LEUKOCYTE_ACTIVATION_INVOLVED_IN_INFLAMMATORY_RESPONSE, pathway aging index=2+42*MIR124-3-9.2*TYROBP-6.4*ITGAM-5.3*CCL3-4.5*AIF1+3.1*C5AR1-1*TLR6; KEGG_NOD_LIKE_RECEPTOR_SIGNALING_PATHWAY, pathway aging index=15-77*CASP8+38*RELA-16*NOD2+14*NLRP3-10*MEFV-4.1*TRIP6-2.7*MAPK10-2.2*CASP5-0.27*CARD6; GO_DEACETYLASE_ACTIVITY, pathway aging index = 11 - 27 * ESD + 15 * SIRT1 - 12 * NDST1 - 7 * HDAC9 - 5.5 * CES1; KEGG_CHRONIC_MYELOID_LEUKEMIA, pathway aging index = 6.4 + 37 * MDM2 - 14 * BCR + 12 * RELA - 9.7 * CDKN1A - 5.4 * SMAD3 - 3.3 * STAT5A - 1.2 * BRAF; GO_TRANSCRIPTION_INITIATION_FROM_RNA_POLYMERASE_I_PROMOTER, pathway aging index=-18+43*POLR1C-32*CCNH+17*POLR2H+15*TAF1D+11*GTF2H1-8.5*RRN3P1-4.9*ERCC3-4.3*TAF1C; GO_POSITIVE_REGULATION_OF_CYTOKINE_BIOSYNTHETIC_PROCESS, pathway aging index=5.2+53*RELA-12*CD86-6.8*BCL10+5.2*CD28-4.1*IL12B-3.6*TYROBP+2.8*ELANE-0.54*MAPKAPK2; GO_T_CELL_SELECTION, pathway aging index = 7.8 + 55 * GLI3 + 26 * SRF - 17 * IL12RB1 - 15 * IL12B - 13 * AIRE + 6.5 * CD28 - 2.1 * DOCK2; GO_POSITIVE_REGULATION_OF_ALPHA_BETA_T_CELL_ACTIVATION, pathway aging index = 7.3 + 91*GLI3 - 37*CD83 - 25*CBFB + 18*AP3B1 + 17*BLM - 17*RASAL3 - 15*IL12RB1 - 13*NCKAP1L - 12*IL6R - 6.3*TGFBR2 - 1.8*IL12B + 0.57*CD28 - 0.38*CD86; and GO_REGULATION_OF_RAS_PROTEIN_SIGNAL_TRANSDUCTION, pathway aging index=3.2+54*TRIM67-47*MET-37*CYTH1+32*RAC1+27*AUTS2-26*CSF1-25*SQSTM1-25*RDX+25*RASGRF2-21*BCR-1 9*MAPKAP1+18*ITGA3-16*ABR+12*FLCN+8.4*ABL2+6.5*RASGEF1A-4.8*ARFGEF2+4.6*PSD3 +4.6*KRAS+4.4*RASAL3+3.1*FARP2-2.6*GPR20+1.1*PLEKHG6+0.52*EPHB2+0.057*MCF2L, Among them, gene represents the DNA methylation characteristic value of the gene; In this study, the pathway aging index for each pathway is represented by X, and the phenotypic label is represented by Y, where Y is 1 for severe cases and Y is 0 for mild cases. A logistic regression model is established with X as the independent variable and Y as the predictor variable. The 10 pathway aging indices of the subjects are substituted into the model to obtain the severe risk value, which is then used to assess immunity.

2. The system according to claim 1, characterized in that, The immunity assessment is performed using the pathway aging index: the DNA methylation characteristics of the genes involved in the 10 pathway aging index linear models of the subject are obtained; the DNA methylation characteristics of the genes are then input into the 10 pathway aging index linear models to obtain the pathway aging index.

3. The system according to claim 1 or 2, characterized in that, The methylation characteristic value is the average of the methylation beta values ​​or the SIMPO value of all methylation sites in the gene body region.

4. The system according to claim 3, characterized in that, The methylation characteristic value is the average of the methylation beta values ​​or the SIMPO value of all methylation sites in the gene promoter region.

5. A method for assessing the immunity of subjects infected with SARS-CoV-2 virus, characterized in that, The method uses the system of any one of claims 1-4 to obtain the subject's severe illness risk value.

6. A medium for assessing the immunity of subjects infected with SARS-CoV-2 virus, characterized in that, The medium includes a program for implementing the method of claim 5, the medium comprising: (1) Read the DNA methylation signature data of the subjects; (2) Calculate the pathway aging index based on the DNA methylation characteristic data of the subjects; (3) Calculate the risk value of severe illness based on the pathway aging index and conduct an immune assessment.

Images