A quantitative method for individual health status homeostasis based on multi-omics data

Abstract

This invention discloses a method for quantifying the homeostasis of an individual's health state based on multi-omics data, specifically including the following steps: 101) Data acquisition step: acquiring multi-omics data of the individual; 102) Homeostasis index acquisition step: quantifying the individual's homeostasis by calculating the curvature of the potential landscape function of each molecule in the individual; 103) Individual health state homeostasis evaluation step: assessing the stability of the individual in the current state and the risk of state transition based on the magnitude of the individual's homeostasis index. This invention provides a method for quantifying the homeostasis of an individual's health state based on multi-omics data and giving corresponding early warnings.

Description

Technical Field

[0001] This invention relates to the fields of computational biology and bioinformatics, and more specifically, to a method for quantifying the homeostasis of individual health status based on multi-omics data. Background Technology

[0002] Modern society's production and lifestyles are undergoing profound changes, posing new challenges to human health. Lack of exercise, poor lifestyle habits, unbalanced diets, and work stress leading to decreased bodily functions have become major factors endangering human health. The incidence of chronic diseases caused by these unhealthy lifestyles, primarily characterized by overeating or an unbalanced diet and insufficient exercise, is increasing. These chronic diseases mainly include hypertension, diabetes, dyslipidemia, obesity, and related cardiovascular and cerebrovascular diseases, tumors, etc. Domestic and international expert guidelines recommend lifestyle intervention as a first-line method for preventing and treating obesity and diabetes. Intervention studies have also confirmed that prediabetes is a key point in controlling the development of clinical diabetes. Currently, people are using various methods, such as health checkups, diet, and exercise, to reduce their risk of developing diseases or delay their progression.

[0003] Therefore, quantifying an individual's health homeostasis and providing disease early warning will help in the early detection, intervention, and treatment of diseases, thereby achieving the long-term goal of controlling chronic diseases.

[0004] Most existing methods for quantifying individual health status involve collecting tissue samples from subjects during physical examinations and measuring static biomarkers and clinical characteristics, such as blood glucose and blood lipids. When the measured value of this biomarker is within the normal range, the subject is considered healthy; otherwise, the subject is considered ill. Existing methods for measuring static biomarkers can only distinguish between healthy and diseased individuals, but cannot differentiate between individuals in a pre-disease state (those about to develop the disease). They lack early warning indicators for disease. For example, individuals in the same pre-disease state may exhibit differences in the stability of that state; some individuals may maintain the pre-disease state for a long time, while others may progress towards disease in a shorter period. This necessitates an indicator to measure the stability of an individual's health status. Most existing methods only quantify whether an individual is diseased, neglecting the quantification of the homeostasis of an individual's health status. Living organisms possess an internal homeostasis mechanism. Homeostasis refers to the body's ability to maintain equilibrium and a relatively stable state in response to physiological loads. Individuals with better homeostasis have a stronger ability to maintain stability and are less prone to shifting to other states; conversely, individuals with poor homeostasis have a weaker ability to maintain stability, are more susceptible to disturbances, and have a higher risk of disease. Quantifying an individual's homeostasis allows for the assessment of their health status and is crucial for disease risk assessment and early warning. From a dynamic perspective, an individual's health status can be described using a potential energy landscape, but quantifying an individual's homeostasis in their current state remains an unsolved problem. Summary of the Invention

[0005] This invention overcomes the shortcomings of existing technologies and provides a method for quantifying the homeostasis of individual health status based on multi-omics data.

[0006] The technical solution of the present invention is as follows:

[0007] A quantitative method for individual health status homeostasis based on multi-omics data includes the following steps:

[0008] 101) Data acquisition steps: Obtain multi-omics data of subjects by collecting blood samples, including but not limited to genomics, metabolomics, proteomics, etc.;

[0009] 102) Steps for obtaining the homeostasis index: The homeostasis index U of a subject is calculated using multi-omics data sampled from multiple points on a single subject to assess the subject’s stability in the current state.

[0010] First, for each subject's multi-omics data, for each feature x, the potential landscape function U(x) of the feature is calculated by applying the potential landscape theory U(x) = -logP(x); the minimum principal curvature of the potential landscape function U(x) of the feature at the mean is the internal stability index of the feature.

[0011] Repeat the above data processing method for each feature to calculate the homeostasis index of each feature of the subject. Sort all features from smallest to largest according to the size of the homeostasis index, select the top k features, and calculate the mean of the homeostasis index of these k features, which is the homeostasis index U of the subject.

[0012] 103) Steps for assessing individual health homeostasis: The stability of a subject's health status is evaluated based on the magnitude of their homeostasis index U. A high homeostasis index indicates that the subject is relatively stable in their current state and unlikely to transition to other states. Conversely, a low homeostasis index indicates poor stability in the subject's current state and a higher risk of transitioning to a disease state. In this case, a risk warning is issued to the subject.

[0013] Furthermore, in step 101), blood samples were collected from the subjects at at least three time points, and the subjects' gene expression profiles and omics data such as metabolomics and proteomics were measured using next-generation high-throughput sequencing technology.

[0014] Furthermore, the specific method for calculating the potential landscape function U(x) of the omics feature x in step 102) is as follows:

[0015] a) For feature x and any other feature y, apply the Pearson correlation coefficient to calculate the correlation between them. The formula for calculating the Pearson correlation coefficient is r(x,y)=cov(x,y) / (σ(x)σ(y)), where cov(x,y) is the covariance of x and y, and σ(x) and σ(y) are the standard deviations of x and y, respectively.

[0016] b) Based on the Pearson correlation coefficient, features highly correlated with feature x are selected and termed feature module M. Feature module M is a local module centered on feature x in the global control network. Principal component analysis (PCA) is used to reduce the dimensionality of feature module M, yielding the first and second principal components, thus reducing the high-dimensional features to two dimensions. The mean μ1 and variance of the first principal component are calculated. Calculate the mean μ2 and variance of the second principal component. Calculate the Pearson correlation coefficient ρ between the first principal component and the second principal component; assuming the expression of the omics feature follows a normal distribution, apply the calculated mean, variance, and correlation coefficient to the bivariate normal distribution probability function to construct the distribution function of this feature module; the probability density function of the bivariate normal distribution is:

[0017] The calculated binary normal distribution function P(X) is used to calculate the potential energy landscape function U(X) using the formula U(X)=-logP(X). This is the potential energy landscape function of the feature module, which characterizes the landscape potential energy function of feature x.

[0018] Furthermore, the method for calculating the internal stability index of feature x is as follows: the landscape potential function of feature x is characterized by the potential landscape function U(X) of feature module M, and the potential landscape function U(X) of feature module M at the mean X is calculated according to the eigenvalue solution method. μ The eigenvalues ​​of the Hessian matrix at (μ1, μ2) are given, where the smallest eigenvalue is the minimum principal curvature of U(X) at the mean (μ1, μ2), which is the internal stability index of the characteristic x. The formula for the Hessian matrix is:

[0019]

[0020] The greater the minimum principal curvature, the greater the curvature of the potential energy landscape U(X) at the mean, and the less likely feature x is to transition to other states. That is, the higher the internal stability index of feature x, the more stable feature x is in the current state.

[0021] The advantages of this invention compared to the prior art are:

[0022] This invention presents a method for quantifying the homeostasis of individual health status based on multi-omics data. It quantifies the stability of individual health through the magnitude of a homeostasis index. Compared to traditional health assessment methods that distinguish between healthy and diseased individuals but neglect the evaluation of health status stability, this invention addresses this aspect, providing a more accurate and comprehensive assessment of individual health status, thus offering more accurate disease warnings and aiding in early disease detection and treatment. Secondly, the molecules within an individual are typically high-dimensional and interacting, making direct calculation of the potential energy landscape function extremely difficult. This invention constructs the potential energy landscape function separately for each local module of the individual's molecules, calculates the minimum principal curvature of the potential energy function for each module, and then sums them to obtain the individual's global homeostasis index. This method considers the interactions between molecules and increases computational feasibility, facilitating practical clinical application. This invention characterizes the stability of the potential energy landscape by calculating curvature, thereby quantifying individual homeostasis and providing an accurate and feasible method for homeostasis quantification. Attached Figure Description

[0023] Figure 1 This is the overall flowchart of the present invention. Detailed Implementation

[0024] The embodiments of the present invention are described in detail below, wherein the same or similar reference numerals denote the same or similar elements or elements with similar functions throughout. The embodiments described below with reference to the accompanying drawings are exemplary and are only used to explain the present invention and should not be construed as limiting the present invention.

[0025] It will be understood by those skilled in the art that, unless otherwise defined, all terms used herein (including technical and scientific terms) have the same meaning as commonly understood by one of ordinary skill in the art to which this invention pertains. It should also be understood that terms such as those defined in general dictionaries should be understood to have the same meaning as in the context of the prior art, and should not be interpreted in an idealized or overly formal sense unless defined as herein.

[0026] The numbering of steps mentioned in the various embodiments is merely for descriptive convenience and does not imply a sequential relationship. Different steps in various specific embodiments can be combined in different orders to achieve the inventive objective of this invention.

[0027] The present invention will be further described below with reference to the accompanying drawings and specific embodiments.

[0028] Example:

[0029] like Figure 1 As shown, the method for quantifying the homeostasis of individual health status based on multi-omics data specifically includes the following steps:

[0030] 101) Data Acquisition Steps: Multi-omics data, including but not limited to genomics, metabolomics, and proteomics, are obtained from subjects by collecting blood samples. Blood samples are collected at at least three time points, and the subjects' gene expression profiles and omics data such as metabolomics and proteomics are measured using next-generation high-throughput sequencing technology.

[0031] The homeostasis index U of the subject is calculated using the acquired data. The homeostasis index U is calculated using the principal curvature method of the homeostasis potential energy function, employing multi-omics data sampled from multiple points on a single subject to assess the subject's stability in the current state. The specific homeostasis index U of each subject is processed in subsequent step 102).

[0032] 102) Steps for obtaining the homeostasis index: The homeostasis index U of a subject is calculated using multi-omics data sampled from multiple points on a single subject to assess the subject's stability in the current state. For each subject's multi-omics data, the potential landscape function U(x) of each feature x is calculated using the potential landscape theory U(x) = -logP(x); the minimum principal curvature of the potential landscape function U(x) at the mean is the homeostasis index of that feature.

[0033] The specific method for calculating the potential landscape function U(x) is as follows:

[0034] a) For feature x and any other feature y, apply the Pearson correlation coefficient to calculate the correlation between them. The formula for calculating the Pearson correlation coefficient is r(x,y)=cov(x,y) / (σ(x)σ(y)), where cov(x,y) is the covariance of x and y, and σ(x) and σ(y) are the standard deviations of x and y, respectively.

[0035] (b) Since physiological functions in organisms are maintained by the interaction of multiple molecules, the omics characteristics within an organism are not independent of each other. Therefore, for each characteristic, a local network centered on that characteristic is first constructed based on the Pearson correlation coefficient. The specific method for constructing the local network is as follows: Calculate the Pearson correlation coefficient between characteristic x and all other characteristics according to step (a), and screen out features that are highly correlated with characteristic x. The screening criterion is: r(x,y)≥0.7. The features that are highly correlated with characteristic x are called feature modules M.

[0036] Principal Component Analysis (PCA) is used to reduce the dimensionality of feature module M, yielding the first and second principal components, effectively reducing the high-dimensional features to two dimensions. The mean μ1 and variance of the first principal component are then calculated. Calculate the mean μ2 and variance of the second principal component. Calculate the Pearson correlation coefficient ρ between the first principal component and the second principal component; assuming the expression of the omics feature follows a normal distribution, apply the calculated mean, variance, and correlation coefficient to the bivariate normal distribution probability function to construct the distribution function of this feature module; the probability density function of the bivariate normal distribution is:

[0037]

[0038] The calculated binary normal distribution function P(X) is used to calculate the potential energy landscape function U(X) using the formula U(X)=-logP(X). This is the potential energy landscape function of the feature module, which characterizes the landscape potential energy function of feature x.

[0039] The minimum principal curvature of the potential landscape function U(X) of this feature at the mean (μ1, μ2) is the internal stability index of this feature. Specifically:

[0040] The potential energy landscape function U(X) is calculated using the eigenvalue solution method at the mean X. μ The eigenvalues ​​of the Hessian matrix at (μ1, μ2) are given, where the smallest eigenvalue is the minimum principal curvature of U(X) at the mean (μ1, μ2), which is the internal stability index of the characteristic x. The formula for the Hessian matrix is:

[0041]

[0042] Repeat the above data processing method for each feature to calculate the homeostatic elasticity index of each feature of the subject. Sort all features from smallest to largest according to the size of the homeostatic index, select the top k features, and calculate the mean of the homeostatic index of these k features, which is the homeostatic index U of the subject.

[0043] 103) Steps for assessing individual health homeostasis: The stability of a subject's health status is evaluated based on the magnitude of their homeostasis index U. A high homeostasis index indicates that the subject is relatively stable in their current state and unlikely to transition to other states. Conversely, a low homeostasis index indicates poor stability in the subject's current state and a higher risk of transitioning to a disease state. In this case, a risk warning is issued to the subject.

[0044] In summary, this approach employs a novel multi-omics data-based method for assessing individual health homeostasis. It quantifies health stability using a homeostasis index and provides disease warnings for individuals at higher risk. This addresses the shortcomings of traditional health assessment methods, offering a more comprehensive and accurate quantification of individual health. Secondly, this invention constructs potential energy landscape functions for each local module of an individual's molecules, calculates the minimum principal curvature of each module's potential energy function, and then sums these to obtain the individual's global homeostasis index. This overcomes the difficulty of calculating potential energy functions for high-dimensional molecules. Our method considers the interactions between molecules and improves computational feasibility, facilitating practical clinical applications. This invention characterizes the stability of the potential energy landscape by calculating curvature, thereby quantifying individual homeostasis and providing an accurate and feasible method for homeostasis quantification.

[0045] The prior description of this disclosure is provided to enable any person skilled in the art to make or use this disclosure. Various modifications to this disclosure will be apparent to those skilled in the art, and the general principles defined herein may be applied to other variations without departing from the spirit or scope of this disclosure. Therefore, this disclosure is not intended to be limited to the examples and designs described herein, but should be accorded the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A quantitative method for individual health status homeostasis based on multi-omics data, characterized by: Specifically, the steps include the following: 101) Data acquisition steps: Obtain multi-omics data from subjects by collecting blood samples, including genomics, metabolomics, and proteomics. 102) Steps for obtaining the homeostasis index: The homeostasis index U of a subject is calculated using multi-omics data sampled from multiple points on a single subject to assess the subject’s stability in the current state; First, for each subject's multi-omics data, for each feature x, the potential landscape function U(x) of the feature is calculated by applying the potential landscape theory U(x) = -logP(x); the minimum principal curvature of the potential landscape function U(x) of the feature at the mean is the internal stability index of the feature. Repeat the above data processing method for each feature to calculate the homeostasis index of each feature of the subject. Sort all features from smallest to largest according to the size of the homeostasis index, select the top k features, and calculate the mean of the homeostasis index of these k features, which is the homeostasis index U of the subject. 103) Individual Health Status Homeostasis Assessment Steps: The homeostasis of the subject's health status is evaluated based on the magnitude of the subject's homeostasis index U. A high homeostasis index indicates that the subject is relatively stable in the current state and is not prone to transitioning to other states. Conversely, a low homeostasis index indicates poor stability in the current state and a higher likelihood of transitioning to a disease state. In this case, the subject's health status is considered to be in a disease warning state, and a risk warning is issued. The specific method for calculating the potential landscape function U(x) of omics feature x is as follows: a) For feature x and any other feature y, apply the Pearson correlation coefficient to calculate the correlation between them. The formula for calculating the Pearson correlation coefficient is r(x,y)=cov(x,y) / (σ(x)σ(y)), where cov(x,y) is the covariance of x and y, and σ(x) and σ(y) are the standard deviations of x and y, respectively. b) Based on the Pearson correlation coefficient, features highly correlated with feature x are selected and termed feature module M. Feature module M is a local module centered on omics feature x in the global regulatory network. Principal component analysis (PCA) is used to reduce the dimensionality of feature module M, yielding the first and second principal components. The mean μ1 and variance of the first principal component are calculated. Calculate the mean μ2 and variance of the second principal component. Calculate the Pearson correlation coefficient ρ between the first principal component and the second principal component; assuming the expression of the omics feature follows a normal distribution, apply the calculated mean, variance, and correlation coefficient to the bivariate normal distribution probability function to construct the distribution function of this feature module; the probability density function of the bivariate normal distribution is: The potential energy landscape function U(X) is calculated using the formula U(X)=-logP(X) after calculating the bivariate normal distribution function P(X).

2. The method for quantifying the homeostasis of individual health status based on multi-omics data according to claim 1, characterized in that: In step 101), blood samples were collected from the subjects at at least three time points, and the subjects' gene expression profiles and metabolomics and proteomics data were measured using next-generation high-throughput sequencing technology.

3. The method for quantifying the homeostasis of individual health status based on multi-omics data according to claim 1, characterized in that: The specific method for calculating the internal stability index of feature x is as follows: The landscape potential function of feature x is characterized by the potential landscape function U(X) of feature module M. The potential landscape function U(X) of feature module M is calculated based on the eigenvalue solving method at the mean X. μ The eigenvalues ​​of the Hessian matrix at (μ1, μ2) are given, where the smallest eigenvalue is the minimum principal curvature of U(X) at the mean (μ1, μ2); the formula for the Hessian matrix is: The greater the minimum principal curvature, the greater the curvature of the potential energy landscape U(X) at the mean, and the less likely the feature x is to transition to other states.

Images