A plasma protein marker combination and screening method for predicting healthy longevity

Abstract

This application discloses a combination of plasma protein biomarkers for predicting healthy longevity and a screening method thereof, relating to the field of bioinformatics. The method includes: acquiring genetic association data between plasma proteins and the healthy longevity phenotype; extracting single nucleotide polymorphisms (SNPs) associated with plasma protein levels as instrumental variables; using Mendelian randomization to identify plasma proteins causally associated with the healthy longevity phenotype and performing sensitivity analysis; performing mediation analysis on the causally associated plasma proteins to identify plasma proteins that influence the healthy longevity phenotype through health-related factors; and constructing a combination of plasma protein biomarkers from the identified plasma proteins. This application screens proteins causally associated with healthy longevity and mediated by health-related factors from genetic association data to construct biomarker combinations, solving the technical problem of low prediction accuracy caused by prior art relying solely on correlation analysis, and achieving the technical effect of improving the prediction accuracy of biomarkers.

Description

Technical Field

[0001] This application relates to the field of bioinformatics technology, and in particular to a combination of plasma protein biomarkers and a screening method for predicting healthy longevity. Background Technology

[0002] With the aging population becoming increasingly severe, accurately monitoring the health status of the elderly and preventing the occurrence and development of diseases has become a key issue in the biomedical field. Plasma proteins, as direct biological effectors of the combined effects of genes and environmental factors, have advantages such as easy availability and quantifiability, and have become an important source of biomarkers for dynamically monitoring aging and predicting mortality risk. Currently, numerous studies have constructed predictive models of aging and disease by detecting plasma protein levels, aiming to identify protein biomarkers associated with healthy aging.

[0003] However, existing technologies mainly rely on correlation analysis to screen plasma protein biomarkers, that is, to identify proteins associated with healthy aging phenotypes by comparing differences in protein expression among different populations. Although these methods can identify relevant proteins, they cannot determine the causal relationship between these proteins and health outcomes, resulting in low prediction accuracy.

[0004] Existing technologies lack effective combinations and screening methods for plasma protein biomarkers to predict healthy longevity. Summary of the Invention

[0005] The purpose of this application is to provide a combination of plasma protein biomarkers for predicting healthy longevity and a screening method. This method screens proteins from the genetic association data of plasma proteins that are causally associated with healthy longevity and mediated by health-related factors to construct a combination of biomarkers. This solves the technical problem of low prediction accuracy caused by existing technologies that rely solely on correlation analysis, and achieves the technical effect of improving the prediction accuracy of biomarkers.

[0006] To achieve the above objectives, this application provides the following solution: In a first aspect, this application provides a method for screening a combination of plasma protein biomarkers for predicting healthy longevity, comprising: acquiring genetic association data of plasma proteins and genetic association data of a healthy longevity phenotype; extracting single nucleotide polymorphisms related to plasma protein levels from the genetic association data as instrumental variables; determining plasma proteins causally associated with the healthy longevity phenotype based on the instrumental variables using Mendelian randomization; performing mediation analysis on the causally associated plasma proteins to identify plasma proteins that mediate influences the healthy longevity phenotype through health-related factors, and constructing the identified plasma proteins into the combination of plasma protein biomarkers.

[0007] Optionally, the extraction of single nucleotide polymorphisms (SNPs) related to plasma protein levels includes: extracting SNPs related to plasma proteins at a genome-wide significance threshold; performing linkage disequilibrium aggregation on the extracted SNPs to retain approximately independent SNPs; and excluding SNPs directly related to healthy longevity phenotypes.

[0008] Optionally, the genome-wide significance threshold is P < 5 × 10⁻⁶. -8 The criterion for excluding those directly associated with healthy longevity phenotypes is P<0.05, where P is the probability value of genome-wide significance.

[0009] Optionally, the nearly independent single nucleotide polymorphisms are obtained through linkage disequilibrium aggregation, wherein the linkage disequilibrium aggregation uses a linkage disequilibrium coefficient r. 2 As a metric, and r 2 <0.1.

[0010] Optionally, the step of using Mendelian randomization to determine plasma proteins causally associated with the healthy longevity phenotype includes using an inverse variance weighted method to estimate the causal effect of plasma proteins on the healthy longevity phenotype.

[0011] Optionally, prior to mediating the causally associated plasma proteins, a sensitivity analysis of the identified causal association may be performed, the sensitivity analysis including at least one of heterogeneity testing, level pleiotropy testing, and outlier detection.

[0012] Optionally, the sensitivity analysis employs the following statistical robustness criteria: heterogeneity testing uses Cochran's Q statistic with a P ≥ 0.05 criterion; level pleiotropy testing uses the MR-Egger intercept test with a P ≥ 0.05 criterion; outlier detection uses the MR-PRESSO global test with a P ≥ 0.05 criterion; where P is the genome-wide significance probability value; the causally associated plasma proteins are further defined as plasma proteins that pass the sensitivity analysis.

[0013] Optionally, the mediation analysis employs a two-step Mendelian randomization-mediated analysis, including: identifying health-related factors causally associated with the healthy longevity phenotype; and quantifying the indirect effects mediated by the health-related factors for plasma proteins that are causally associated with both the healthy longevity phenotype and the health-related factors.

[0014] Optionally, the quantification of the indirect effects mediated by health-related factors includes: calculating the β coefficient of the indirect effect using the product method, wherein the β coefficient of the indirect effect is the product of the β coefficient of the causal effect of plasma protein on health-related factors and the β coefficient of the causal effect of health-related factors on the healthy longevity phenotype; the β coefficient is a regression coefficient used to quantify the magnitude of the causal effect; and calculating the standard error and confidence interval of the indirect effect using the Delta method.

[0015] In a second aspect, a combination of plasma protein biomarkers for predicting healthy longevity is characterized in that the combination of plasma protein biomarkers is obtained by the screening method described above.

[0016] According to the specific embodiments provided in this application, the following technical effects are disclosed: This application provides a combination of plasma protein biomarkers and a screening method for predicting healthy longevity. By extracting single nucleotide polymorphisms (SNPs) related to protein levels from plasma protein genetic association data as instrumental variables and subjecting them to rigorous statistical screening, it ensures that the genetic instruments entering subsequent analysis meet the core assumptions of Mendelian randomization, laying the foundation for the reliability of causal inference. The Mendelian randomization method is used to estimate the causal effect of plasma proteins on the healthy longevity phenotype, overcoming the limitations of existing technologies that rely solely on correlation analysis. This method can accurately identify the driver proteins that truly influence the process of healthy longevity, rather than merely associated companion proteins, thus improving the predictive accuracy of the biomarkers. Furthermore, by performing mediation analysis on causally associated plasma proteins, proteins that mediate effects on healthy longevity through health-related factors are identified, making the final constructed combination of plasma protein biomarkers more interpretable. Attached Figure Description

[0017] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0018] Figure 1 A schematic flowchart illustrating a screening method for a combination of plasma protein biomarkers for predicting healthy longevity, provided in Embodiment 1 of this application; Figure 2 for Figure 1 A detailed flowchart of step S2; Figure 3 for Figure 1 A detailed flowchart of step S5. Detailed Implementation

[0019] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0020] To make the above-mentioned objectives, features and advantages of this application more apparent and understandable, the application will be further described in detail below with reference to the accompanying drawings and specific embodiments.

[0021] Example 1: A screening method for a combination of plasma protein biomarkers for predicting healthy longevity.

[0022] like Figure 1 As shown, this embodiment provides a screening method for combinations of plasma protein biomarkers for predicting healthy longevity. This method is executed by a computer and includes the following steps: S1, obtain genetic association data of plasma proteins and genetic association data of healthy longevity phenotypes.

[0023] S2, extract single nucleotide polymorphisms related to plasma protein levels from genetic association data as instrumental variables.

[0024] S3, based on instrumental variables, uses Mendelian randomization to identify plasma proteins causally associated with the healthy longevity phenotype.

[0025] S4, perform sensitivity analysis on the established causal relationships.

[0026] S5 involves mediating plasma proteins with causal relationships to identify plasma proteins that mediate the effects of health-related factors on the healthy longevity phenotype, and constructing a ensemble of plasma protein biomarkers from the identified plasma proteins.

[0027] The above steps S1 to S4 are implemented as follows: Step S1 lays the data foundation for subsequent causal inference by introducing genetic variation as an instrumental variable. Utilizing its random allocation characteristic, it avoids the interference of confounding factors that are difficult to control in traditional observational studies, thus providing genetic evidence to support the screening results. Step S2 extracts single nucleotide polymorphisms related to protein levels from plasma protein genetic association data as instrumental variables, and conducts rigorous statistical screening to ensure that the genetic instruments entering subsequent analysis meet the core assumptions of Mendelian randomization, laying the foundation for the reliability of causal inference. Step S3 overcomes the limitations of existing technologies that rely solely on correlation analysis, enabling accurate inference of the causal relationship between plasma proteins and healthy longevity. Step S4 performs sensitivity analysis on the identified causal associations to ensure the reliability and reproducibility of the final screened causal associations. Step S5, based on the clarified causal associations, further analyzes the pathways by which plasma proteins influence healthy longevity, identifying plasma proteins that play a role in health-related factors, making the final constructed biomarker combination not only predictive but also interpretable.

[0028] In practice, the genetic association data for plasma proteins in step S1 is derived from aggregated statistics from the UK Biobank PharmaProteomics Project (UKB-PPP). UKB-PPP is currently one of the largest databases for plasma proteomics genetic association studies, containing genetic association information for thousands of plasma proteins.

[0029] In step S1, the genetic association data for the healthy longevity phenotype were integrated from the UKKB and FinnGen databases. UKKB is a large-scale biomedical cohort study database in the UK, containing genetic data, health records, and lifestyle information for over 500,000 participants. FinnGen (version 12) is a large-scale genomics study project in Finland aimed at discovering genetic variations associated with disease and health; version 12 contains genetic information and health records for over 500,000 participants. After removing and unifying duplicate data from the two databases, the integrated data includes a total of 9,692 genetic variations associated with 44,424 unique health-related factors, encompassing 3,657 diseases and 442 lifestyle factors.

[0030] In specific implementation, such as Figure 2 As shown, in step S2, single nucleotide polymorphisms related to plasma protein levels are extracted from genetic association data as instrumental variables, specifically including the following steps: S21, extracting single nucleotide polymorphisms associated with plasma proteins at a genome-wide significance threshold.

[0031] Specifically, the genome-wide significance threshold was P < 5 × 10⁻⁶.-8 And the P-value is the genome-wide significance probability value.

[0032] S22 performs linkage disequilibrium aggregation on the extracted single nucleotide polymorphisms, preserving nearly independent single nucleotide polymorphisms.

[0033] Specifically, near-independent single nucleotide polymorphisms are obtained through linkage disequilibrium aggregation, which is achieved using the linkage disequilibrium coefficient r. 2 As a metric, and with a threshold of r 2 <0.1. Chaining imbalance aggregation can be performed using PLINK software (version 1.9), setting the aggregation window to 100 kb, r 2 A threshold of 0.1 was set to preserve a set of nearly independent single nucleotide polymorphisms (SNPs) and to preserve the most important and significant SNPs in each linkage disequilibrium (LD) region.

[0034] S23 excludes single nucleotide polymorphisms directly associated with healthy longevity phenotypes.

[0035] Specifically, the nearly independent single nucleotide polymorphisms retained after linkage disequilibrium aggregation in step S22 are compared with the GWAS summary statistics of the results to exclude any SNPs directly related to the results, ensuring that the genetic instrument variables meet the independence assumption of Mendelian randomization analysis. The GWAS summary statistics of the results include: protein GWAS data from UKB-PPP: i.e., genome-wide association study data related to plasma protein levels from the UK Biobank Pharma Proteomics Project. This data is from the same source as the exposure data used in step S1, but is used here to examine whether the selected SNPs have other complex associations with plasma proteins themselves; and phenotypic GWAS data from UK nearlab FINN R12: i.e., genome-wide association study data related to health-related phenotypes from UK and FinnGen (version 12). This data covers various phenotypes associated with healthy longevity and is used to examine whether the selected SNPs are directly associated with these phenotypes.

[0036] Any single nucleotide polymorphism (SNP) that is significantly associated with the outcome phenotype recorded in the GWAS summary statistics (i.e., P < 0.05) is excluded.

[0037] Therefore, according to the core assumptions of Mendelian randomization analysis (MR), effective instrumental variables must influence the outcome only through exposure factors, and not be directly associated with the outcome. By excluding SNPs that are directly related to the outcome, potential violations of the independence assumption can be reduced, ensuring the reliability of subsequent causal effect estimations.

[0038] In practice, in step S3, based on the instrumental variables obtained in step S2, Mendelian randomization is used to identify plasma proteins that are causally associated with the healthy longevity phenotype. All analyses are based on the three core assumptions of Mendelian randomization: (1) the genetic instrument is closely related to the exposure (plasma protein level); (2) the genetic instrument is independent of any confounding factors in the exposure-outcome relationship; and (3) the genetic instrument can only affect the outcome through exposure.

[0039] The specific steps for determining plasma proteins causally associated with the healthy longevity phenotype using Mendelian randomization include: First, for each protein-longevity pair, the effect estimates for SNP-exposure and SNP-outcome are unified to the same effect allele to ensure consistency in the direction of causal effect estimation.

[0040] Then, based on the number of available inverse variance weighted instrumental variables, different causal effect estimation methods are selected: (1) When the number of available inverse variance weighted instrumental variables is 1, the Wald ratio method is used to estimate the causal effect of plasma proteins on the healthy longevity phenotype.

[0041] (2) When there are two or three available inverse variance weighted instrumental variables, the fixed effects inverse variance weighted model is used to estimate the causal effects.

[0042] (3) When the number of available inverse variance weighted instrumental variables is greater than 3, the multiplicative random effects inverse variance weighted model is used to estimate the causal effect.

[0043] It is understandable that the aforementioned "causal association" and "causal effect" are technically related. "Causal effect" refers to the quantitative value of the impact of exposure (plasma proteins) on the outcome (healthy longevity phenotype), estimated using Mendelian randomization methods, including the β coefficient and odds ratio. "Causal association," on the other hand, refers to the conclusion that a causal relationship exists between exposure and outcome, based on the statistical significance of the estimated causal effect (e.g., FDR ≤ 0.05). In other words, the existence of a "causal association" is determined by estimating the "causal effect."

[0044] In practice, in step S4, sensitivity analysis is performed on the causally related plasma proteins identified in step S3 to assess the robustness of the research results. Sensitivity analysis includes at least one of heterogeneity testing, pleiotropic level testing, and outlier detection.

[0045] As a preferred approach, the sensitivity analysis employs the following statistical judgment results for robustness: Heterogeneity was tested using Cochran's Q statistic, with a criterion of P ≥ 0.05, where P represents the probability of genome-wide significance. P ≥ 0.05 indicates no heterogeneity among the instrumental variables and robustness of the results.

[0046] The pleiotropic effect was tested using the MR-Egger intercept test, with a judgment criterion of P ≥ 0.05. P ≥ 0.05 indicates that the MR-Egger intercept does not significantly deviate from 0, and there is no directional pleiotropic effect.

[0047] Outlier detection was performed using the MR-PRESSO global test, with a judgment criterion of P ≥ 0.05. P ≥ 0.05 indicates that there are no outlier instrumental variables that would lead to level pleiotropy.

[0048] The plasma proteins identified through the above sensitivity analysis were further narrowed down to plasma proteins with significant MR associations, i.e., candidate proteins with robust causal associations with the healthy longevity phenotype.

[0049] In specific implementation, in step S5, the plasma proteins with robust causal associations identified in step S4 are subjected to mediation analysis to identify plasma proteins that mediate the influence of healthy longevity phenotypes through health-related factors, and the identified plasma proteins are constructed into a plasma protein biomarker ensemble.

[0050] As a preferred approach, the mediation analysis employs a two-step Mendelian randomized mediated analysis, such as... Figure 3 As shown, the specific steps include: S51, identify health-related factors that are causally associated with the healthy longevity phenotype.

[0051] Specifically, the Mendelian randomization method, the same as in step S3, is used to estimate the causal effect of health-related factors on longevity, with health-related factors as exposures and the healthy longevity phenotype as the outcome.

[0052] S52, for plasma proteins that are causally associated with the healthy longevity phenotype and causally associated with the health-related factors identified in step S51, quantify the indirect effects mediated by the health-related factors.

[0053] As a preferred method, quantifying the indirect effects mediated by health-related factors includes: calculating the β coefficient of the indirect effect using the product method, where the β coefficient of the indirect effect is the product of the β coefficient of the causal effect of plasma protein on health-related factors and the β coefficient of the causal effect of health-related factors on the healthy longevity phenotype, and the β coefficient is a regression coefficient used to quantify the magnitude of the causal effect; and calculating the standard error and confidence interval of the indirect effect using the Delta method.

[0054] Preferably, in the mediation analysis, plasma proteins that mediate the influence of health-related factors on the healthy longevity phenotype meet the significance threshold after false alarm rate (FDR) correction, and FDR ≤ 0.05. That is, only when the indirect effect of a protein remains significant after FDR correction (FDR ≤ 0.05) are they identified as plasma proteins that mediate the influence of health-related factors on the healthy longevity phenotype and included in the final biomarker combination.

[0055] Example 2: A combination of plasma protein biomarkers for predicting healthy longevity.

[0056] In this embodiment, the plasma protein biomarker combination was obtained by implementing the screening method described above. Each protein in the plasma protein biomarker combination has the following characteristics: a robust causal association with the healthy longevity phenotype; its mechanism of action influencing healthy longevity is mediated by specific health-related factors; and it has been verified by rigorous statistical tests (FDR≤0.05).

[0057] In one exemplary embodiment, a computer device is provided, which may be a server or a terminal. The computer device includes a processor, a memory, an input / output interface, and a communication interface. The database of the computer device is used to store genetic association data of plasma proteins and genetic association data of healthy longevity phenotypes. When the computer program is executed by the processor, it implements the steps of the screening method described above for predicting combinations of plasma protein biomarkers for healthy longevity.

[0058] In one exemplary embodiment, a computer-readable storage medium is provided storing a computer program that, when executed by a processor, implements the steps in the above-described method embodiments.

[0059] In one exemplary embodiment, a computer program product is provided, including a computer program that, when executed by a processor, implements the steps in the above-described method embodiments.

[0060] Those skilled in the art will understand that all or part of the processes in the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer-readable storage medium. When executed, the computer program can include the processes of the embodiments of the above methods. Furthermore, any references to memory, databases, or other media used in the embodiments provided in this application can include at least one of non-volatile and volatile memory.

[0061] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0062] This document uses specific examples to illustrate the principles and implementation methods of this application. The descriptions of the above embodiments are only for the purpose of helping to understand the methods and core ideas of this application. Furthermore, those skilled in the art will recognize that, based on the ideas of this application, there will be changes in the specific implementation methods and application scope. Therefore, the content of this specification should not be construed as a limitation of this application.

Claims

1. A method for screening combinations of plasma protein biomarkers for predicting healthy longevity, characterized in that, include: Obtain genetic association data for plasma proteins and genetic association data for healthy longevity phenotypes; Single nucleotide polymorphisms associated with plasma protein levels were extracted from genetic association data as instrumental variables; Based on the instrumental variables, Mendelian randomization was used to identify plasma proteins that are causally associated with the healthy longevity phenotype. Mediation analysis was performed on the causally related plasma proteins to identify plasma proteins that mediate the influence of healthy longevity phenotypes through health-related factors, and the identified plasma proteins were constructed into the plasma protein biomarker ensemble.

2. The screening method for combinations of plasma protein biomarkers for predicting healthy longevity according to claim 1, characterized in that, The extraction of single nucleotide polymorphisms related to plasma protein levels includes: Single nucleotide polymorphisms associated with plasma proteins were extracted at a genome-wide significance threshold. The extracted single nucleotide polymorphisms were subjected to linkage non-equilibrium aggregation, preserving nearly independent single nucleotide polymorphisms; Single nucleotide polymorphisms directly associated with healthy longevity phenotypes were excluded.

3. The screening method for combinations of plasma protein biomarkers for predicting healthy longevity according to claim 2, characterized in that, The significance threshold for the whole genome was P < 5 × 10⁻⁶. -8 The criterion for excluding those directly associated with healthy longevity phenotypes is P<0.05, where P is the probability value of genome-wide significance.

4. The screening method for combinations of plasma protein biomarkers for predicting healthy longevity according to claim 2, characterized in that, The nearly independent single nucleotide polymorphisms are obtained through linkage disequilibrium aggregation, which is achieved using a linkage disequilibrium coefficient r. 2 As a metric, and r 2 <0.

1.

5. The method for screening combinations of plasma protein biomarkers for predicting healthy longevity according to claim 1, characterized in that, The method of using Mendelian randomization to identify plasma proteins causally associated with the healthy longevity phenotype includes using an inverse variance weighted method to estimate the causal effect of plasma proteins on the healthy longevity phenotype.

6. The method for screening combinations of plasma protein biomarkers for predicting healthy longevity according to claim 1, characterized in that, Before performing mediation analysis on the causally associated plasma proteins, the method further includes performing sensitivity analysis on the identified causal association, which includes at least one of heterogeneity test, level pleiotropy test, and outlier detection.

7. The method for screening combinations of plasma protein biomarkers for predicting healthy longevity according to claim 6, characterized in that, The sensitivity analysis was conducted based on the robustness of the following statistical judgment results: Heterogeneity was tested using Cochran's Q statistic, with a judgment criterion of P ≥ 0.05; The pleiotropic effect was tested using the MR-Egger intercept test, with a judgment criterion of P ≥ 0.

05. Outlier detection was performed using the MR-PRESSO global test, with a judgment criterion of P ≥ 0.05; Wherein, P-value is the genome-wide significance probability value; The causally related plasma proteins are further defined as plasma proteins that pass the sensitivity analysis.

8. The method for screening combinations of plasma protein biomarkers for predicting healthy longevity according to claim 1, characterized in that, The mediation analysis employs a two-step Mendelian randomized mediated analysis, including: Identify health-related factors that are causally associated with the healthy longevity phenotype; For plasma proteins that are causally associated with a healthy longevity phenotype and are also causally associated with the aforementioned health-related factors, quantify the indirect effects mediated by those health-related factors.

9. The method for screening combinations of plasma protein biomarkers for predicting healthy longevity according to claim 8, characterized in that, The indirect effects mediated by the quantified health-related factors include: The β coefficient of the indirect effect is calculated using the product method. The β coefficient of the indirect effect is the product of the β coefficient of the causal effect of plasma protein on health-related factors and the β coefficient of the causal effect of health-related factors on the healthy longevity phenotype. The β coefficient is a regression coefficient used to quantify the magnitude of the causal effect. The standard error and confidence interval of the indirect effect were calculated using the Delta method.

10. A combination of plasma protein biomarkers for predicting healthy longevity, characterized in that, The plasma protein biomarker combination is obtained by the screening method according to any one of claims 1-9.

Images