Cardiovascular state-oriented circadian rhythm assessment method and system

Abstract

The invention provides a circadian rhythm evaluation method and system oriented to cardiovascular states, and belongs to the field of transcriptional gene analysis, and the method comprises the steps: obtaining transcriptome data at a single time point; gene expression data are positioned and extracted from the transcriptome data through a preset rhythm homeostasis gene module, and the rhythm homeostasis gene module is constructed through a weighted gene co-expression network analysis algorithm; and calculating the collaboration, the volatility and the network concentration degree of the gene expression data based on a dynamic network biomarker algorithm, and determining the circadian rhythm disorder index of the to-be-evaluated user according to the collaboration, the volatility and the network concentration degree. According to the method, collaborative expression characteristics are systematically captured by using a rhythm homeostasis gene module, dependence on a small number of core clock gene signals is avoided, and quantitative calculation is carried out on the collaboration, volatility and network concentration ratio of gene expression data in the module based on a dynamic network biomarker algorithm; and an objective and quantifiable molecular evaluation index is provided for an individual rhythm state.

Description

TECHNICAL FIELD

[0001] The application belongs to the field of transcriptional gene analysis, and particularly relates to a circadian rhythm evaluation method and system for cardiovascular state. BACKGROUND

[0002] Circadian rhythm is a core biological process for maintaining physiological homeostasis, and is widely involved in the regulation of metabolism, immunity, hemodynamics and vascular wall function. Its basic mechanism relies on the coupling and synchronization of multi-level molecular clock networks in different tissues and cell types. In recent years, a large number of studies have shown that circadian rhythm disorders are closely related to the occurrence and development of many chronic diseases, especially the formation of atherosclerotic cardiovascular disease (ASCVD), the triggering of acute events and the prognosis of the disease. Specifically, in the management of cardiovascular disease, the mismatch of circadian rhythm has important timing significance. The process of ASCVD is not linear, but goes through several stages of transition, among which the non-ST segment elevation myocardial infarction (NSTEMI) stage is a key node for rapid transition from stability to instability. If the aggravation trend of rhythm disorder can be identified in this stage, it is possible to provide a scientific basis for early intervention, treatment timing selection and perioperative management. In addition, the stability of atherosclerotic plaques is closely related to the rhythm remodeling of the local microenvironment, and the functional changes of smooth muscle cells, fibroblasts and immune cells and other populations in the state of rhythm imbalance will directly affect the integrity of the fibrous cap and the inflammatory response.

[0003] In view of the above-mentioned needs, a number of epidemiological and genetic studies have shown that shift work can significantly disrupt the biological clock rhythm and increase the risk of acute cardiovascular events through metabolic, inflammatory and vascular function disorders. On this basis, the existing methods mostly rely on macro-level proxy indicators such as shift work, sleep rhythm disorder, jet lag to evaluate the circadian rhythm of users, which ignores the systematic dynamic changes at the molecular level, and is difficult to truly reflect the mismatch degree of endogenous rhythm in the progress of the disease, thereby leading to low accuracy of circadian rhythm evaluation. SUMMARY

[0004] In order to solve the problem of low evaluation accuracy of existing circadian rhythm relying on macro indicators, the application provides a circadian rhythm evaluation method and system for cardiovascular state.

[0005] In order to achieve the above-mentioned purpose, the application provides the following technical scheme: A circadian rhythm evaluation method for cardiovascular state, comprising: obtaining the transcriptome data of a single time point of a user to be evaluated; The gene expression data is located and extracted from the transcriptome data by a preset rhythm steady-state gene module, wherein the rhythm steady-state gene module is constructed by a weighted gene co-expression network analysis algorithm, and the weighted gene co-expression network analysis algorithm is trained based on training transcriptome data extracted from the cardiovascular state sample; The synergy, fluctuation and network concentration of the gene expression data are calculated based on a dynamic network biomarker algorithm, and the circadian rhythm disorder index of the to-be-evaluated user is determined according to the synergy, fluctuation and network concentration, wherein the circadian rhythm disorder index is used to quantify the disorder degree of the circadian rhythm of the to-be-evaluated user.

[0006] Optionally, the cardiovascular state-oriented circadian rhythm evaluation method provided by the present application further comprises: The correlation strength of each gene module constructed by the weighted gene co-expression network analysis algorithm and the preset rhythm feature is calculated, and the rhythm steady-state gene module is determined from the gene module with a correlation strength higher than a preset threshold.

[0007] Optionally, the cardiovascular state-oriented circadian rhythm evaluation method provided by the present application further comprises: Obtaining training transcriptome data; Based on the single cell clustering strategy, the plurality of training transcriptome data is aggregated into one meta cell to obtain a plurality of meta cell expression matrices; The expression similarity between the plurality of meta cell expression matrices is calculated to obtain a similarity matrix; the weighted adjacency matrix is obtained by power operation on the similarity matrix, and the topological overlap measure is calculated according to the weighted adjacency matrix; Based on the topological overlap measure, the genes in the training set of the transcriptome data are clustered to obtain a cluster tree diagram; the cluster tree diagram is divided based on the gene aggregation degree to obtain a circadian rhythm gene co-expression network composed of a plurality of gene modules.

[0008] Optionally, the cardiovascular state-oriented circadian rhythm evaluation method provided by the present application further comprises: Obtaining transcriptome data of different species; obtaining transcriptome data of different tissues; The training transcriptome data is constructed from the transcriptome data of different species and the transcriptome data of different tissues.

[0009] Optionally, the cardiovascular state-oriented circadian rhythm evaluation method provided by the present application further comprises: Based on the gene expression profile, the training transcriptome data is clustered according to the cell type to obtain a plurality of cell subgroups; In each cell subgroup, a plurality of cells with similar transcriptome features are aggregated into one meta cell, wherein the gene expression amount of the meta cell is the average or median of the gene expression amounts of all cells in the cell subgroup.

[0010] Optionally, the circadian rhythm assessment method for cardiovascular status provided by the present invention further includes: Calculate the average expression correlation of gene expression data to obtain the synergy of the users to be evaluated; Calculate the expression variance or dispersion of gene expression data to obtain the volatility of the user to be evaluated; The average connection strength of the rhythm homeostatic gene module corresponding to the gene expression data and other gene modules in the diurnal rhythm gene co-expression network is calculated to obtain the network concentration of the user to be evaluated.

[0011] Optionally, the circadian rhythm assessment method for cardiovascular status provided by the present invention further includes: The overall expression levels of circadian rhythm phase features, circadian rhythm amplitude features, and rhythm homeostasis gene modules were extracted from transcriptome data. The circadian rhythm subtype of the user to be evaluated is determined by a pre-trained machine learning classifier based on the circadian rhythm disorder index, circadian rhythm phase features, circadian rhythm amplitude features, and the overall expression level of the rhythm homeostasis gene module. The circadian rhythm subtype is constructed from the training transcriptome data using an unsupervised clustering algorithm.

[0012] This invention also provides a circadian rhythm assessment system for cardiovascular status, comprising: The transcription data acquisition module is used to acquire transcriptome data from a single time point for the user to be evaluated; The gene expression extraction module is used to locate and extract gene expression data from transcriptome data through a pre-set rhythm homeostasis gene module. The rhythm homeostasis gene module is constructed using a weighted gene co-expression network analysis algorithm, which is trained based on training transcriptome data extracted from cardiovascular state samples. The disorder index assessment module is used to calculate the synergy, volatility, and network concentration of gene expression data based on the dynamic network biomarker algorithm. Based on the synergy, volatility, and network concentration, the circadian rhythm disorder index of the user to be evaluated is determined. The circadian rhythm disorder index is used to quantify the degree of disorder of the user's circadian rhythm.

[0013] The present invention also provides a computer device, including a memory, a processor, and a computer program stored in the memory, wherein the processor executes the computer program to implement any of the steps in a method for assessing circadian rhythms for cardiovascular conditions.

[0014] The present invention also provides a computer-readable storage medium storing a computer program that, when loaded by a processor, is capable of executing any step of a method for assessing circadian rhythms in relation to cardiovascular conditions.

[0015] The circadian rhythm assessment method for cardiovascular status provided by this invention has the following beneficial effects: Because the circadian rhythm assessment method provided by this invention utilizes a rhythm homeostatic gene module pre-constructed by a weighted gene co-expression network analysis algorithm, it can systematically capture the co-expression characteristics of rhythm-related genes across the entire genome, thereby eliminating dependence on a few core clock gene signals. Based on a dynamic network biomarker algorithm, it quantitatively calculates the synergy, volatility, and network concentration of gene expression data within the module, thus accurately revealing the degree of homeostatic deviation of the endogenous circadian rhythm system from the perspective of network dynamics. The final output rhythm disorder index provides an objective and quantifiable molecular assessment indicator for individual rhythm status, improving the accuracy of circadian rhythm assessment. Attached Figure Description

[0016] To more clearly illustrate the embodiments and design schemes of the present invention, the accompanying drawings required for this embodiment will be briefly described below. The drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0017] Figure 1 A schematic diagram of a circadian rhythm assessment method for cardiovascular status provided in an embodiment of the present invention; Figure 2 This is an example of the process for constructing a rhythm gene co-expression network provided in an embodiment of the present invention; Figure 3 Examples of coronary heart disease process handling at different disease stages provided in embodiments of the present invention; Figure 4 This is an example of rhythmic subtype distribution differences provided in an embodiment of the present invention; Figure 5 This is an example of a rhythm subtype determination process provided in an embodiment of the present invention; Figure 6 This is an example of classifier model performance verification provided in an embodiment of the present invention; Figure 7 Examples of various dataset verification provided in the embodiments of the present invention; Figure 8 Examples of different tissue evaluations provided in embodiments of the present invention; Figure 9 Examples of different evaluation modes provided in embodiments of the present invention; Figure 10 Examples of different queue score distributions provided in embodiments of the present invention; Figure 11 This is a visual output display example provided for an embodiment of the present invention. Detailed Implementation

[0018] To enable those skilled in the art to better understand and implement the technical solutions of the present invention, the present invention will be described in detail below with reference to the accompanying drawings and specific embodiments. The following embodiments are only used to more clearly illustrate the technical solutions of the present invention and should not be construed as limiting the scope of protection of the present invention.

[0019] To address the need to quantify the degree of endogenous circadian rhythm disruption in individuals, two approaches are taken: firstly, macroscopic indicators such as shift work, sleep rhythm disturbances, and jet lag are used to infer rhythm status; secondly, a few core genes, such as "CLOCK, BMAL1, PER, and CRY," are focused on to reflect rhythm regulation. However, while this method of inferring the state of the biological clock through a few clock gene signals can reflect local characteristics of rhythm regulation, it cannot represent the dynamic imbalance of the entire rhythm network, nor can it capture systemic rhythm mismatches under multi-gene, multi-pathway, and multi-cellular interactions.

[0020] Specifically, from a measurement perspective, existing methods generally rely on external behavioral indicators, such as sleep-wake cycles, activity rhythm curves, shift work history, and diurnal blood pressure variability. While these indicators can reflect external circadian rhythm states to some extent, they are significantly influenced by individual lifestyles, psychological states, ambient light, and medication use, exhibiting considerable randomness and individual variability. On the other hand, at the molecular level, current research largely focuses on a few core clock genes, such as CLOCK, BMAL1, PER, and CRY. Although these genes can reveal some characteristics of the rhythm regulatory network, they cannot represent the overall rhythmic activity across multiple pathways, cell types, and organ systems, thus lacking global and systematic perspectives. In other words, current technologies cannot integrate multi-gene expression characteristics, nor can they quantitatively describe the differences between circadian rhythm homeostasis and the degree of disruption at the individual level.

[0021] From a methodological perspective, there is currently no universally applicable system for quantifying molecular rhythms across different tissues and species. Some studies have attempted to reconstruct the diurnal rhythms of tissue samples using temporal phase inference algorithms (such as ZeitZeiger and CYCLOPS), but these algorithms typically rely on precise recording of sampling time and model training under specific conditions, requiring extremely rigorous data collection and thus hindering their widespread application in large-sample or clinical datasets. More importantly, the characteristic gene sets of these models vary significantly across different tissues or disease backgrounds, leading to a significant decrease in predictive performance when used across platforms. Furthermore, their outputs are mostly temporal phases or relative ordination, lacking quantitative indicators reflecting the "intensity" or "direction" of rhythm disturbances, limiting their interpretability and practicality, and making them difficult to assist physicians in disease research and clinical settings.

[0022] Furthermore, existing research often focuses solely on whether circadian rhythms are disrupted, neglecting the intensity and dynamic evolution of rhythm disturbances. Rhythm mismatches exhibit continuous, reversible, and phased characteristics in most diseases. For example, in the progression of atherosclerosis, the degree of rhythm disturbance gradually deepens with disease progression, peaking in non-ST-segment elevation myocardial infarction (NSTEMI), a critical transition point from stability to instability. However, current technologies cannot continuously characterize this dynamic change, nor can they identify the critical point where the rhythm system transitions from homeostasis to instability. This deficiency prevents molecular markers of rhythm disturbances from being used by physicians for early disease warning or intervention timing.

[0023] Furthermore, in terms of cross-species and cross-tissue validation, existing rhythm studies mostly remain at the single-species or single-tissue level, lacking a systematic validation framework. Due to significant differences in the phase, amplitude, and signal-to-noise ratio of rhythm gene expression among different species, and the potential for asynchronous rhythms in different tissues within the same species, a quantitative model capable of maintaining stable performance across multiple species and tissues is lacking. This lack of standardization and transferability makes it difficult to establish unified measurement standards for rhythm-related indicators, and even more difficult to achieve consistent application between clinical and experimental models. Simultaneously, differences in gene screening, normalization methods, and model training procedures among different studies also lead to poor comparability and insufficient reproducibility of results.

[0024] Furthermore, while behavioral monitoring and time-series sampling are feasible in laboratory settings, they are difficult to implement in clinical practice due to high costs, long cycles, and poor compliance. Although molecular detection methods exist, most require high-frequency sampling or special lighting conditions, making them unsuitable for the discontinuous and heterogeneous characteristics of clinical samples. Therefore, there is currently no standardized procedure that can assess the degree of circadian rhythm disruption in an individual using only single transcriptome data.

[0025] To address the aforementioned shortcomings, this invention provides a method and system for assessing circadian rhythms in cardiovascular conditions, namely the MyPROClock system (www.myproclock-cardioartist.com). This system aims to establish a quantitative assessment framework for circadian rhythm disorders that can be applied across tissues, species, and platforms. It integrates the expression characteristics of rhythm-related genes at the whole-genome level, combining weighted gene co-expression network analysis (WGCNA) and dynamic network biomarker (DNB) algorithms to construct a "Rhythm Disorder Index (MyPROClock score)" that can be calculated at the individual level. This index can continuously and quantitatively reflect the degree of deviation of the body's rhythm system from homeostasis to imbalance, and can be used in population studies as well as animal models and cell experiments.

[0026] Specifically, the MyPROClock system integrates a weighted gene co-expression network (WGCNA) with a dynamic network biomarker (DNB) model to screen core gene modules closely related to rhythm mismatch and calculates a rhythm disorder index that reflects the degree of rhythm homeostasis and imbalance at the individual level. The MyPROClock system enables universal assessment across tissues and species and maintains stability across different data types such as bulk RNA-seq and single-cell RNA-seq.

[0027] Example 1 This application provides a method for assessing circadian rhythms in relation to cardiovascular status, specifically as follows: Figure 1 As shown, it includes the following steps: Step 11: Obtain transcriptome data for a single time point from the user to be evaluated.

[0028] Step 12: Obtain training transcriptome data; aggregate multiple training transcriptome data into a single metacell based on a single-cell clustering strategy to obtain multiple metacell expression matrices; calculate the expression similarity between multiple metacell expression matrices to obtain a similarity matrix; perform exponentiation on the similarity matrix to obtain a weighted adjacency matrix, and calculate the topological overlap measure based on the weighted adjacency matrix.

[0029] The training transcriptome data was constructed using the following steps: Step 121: Obtain transcriptome data from different species; obtain transcriptome data from different tissues.

[0030] Step 122: Construct training transcriptome data from transcriptome data of different species and different tissues.

[0031] The metacellular expression matrix is ​​constructed using the following steps: Step 123: Based on gene expression profiles, cluster the training transcriptome data according to cell type to obtain multiple cell subpopulations.

[0032] Step 124: In each cell subpopulation, multiple cells with similar transcriptomic characteristics are aggregated into a metacell, wherein the gene expression level of the metacell is the average or median of the gene expression levels of all cells in that cell subpopulation.

[0033] Step 13: Cluster the genes in the transcriptome training set based on the topological overlap measure to obtain a clustering dendrogram; divide the clustering dendrogram based on the degree of gene aggregation to obtain a circadian rhythm gene co-expression network composed of multiple gene modules.

[0034] Step 14: Calculate the association strength between the multiple gene modules constructed by the weighted gene co-expression network analysis algorithm and the pre-set rhythm features, and determine the rhythm homeostatic gene modules by the gene modules whose association strength is higher than the preset threshold.

[0035] Step 15: Locate and extract gene expression data from transcriptome data using a pre-set rhythm homeostasis gene module. The rhythm homeostasis gene module is constructed using a weighted gene co-expression network analysis algorithm.

[0036] Specifically, taking the MyPROClock system, which employs the circadian rhythm assessment method of this invention, as an example, the circadian rhythm assessment process adopted in this invention will be further explained. The MyPROClock system includes a data input module, a rhythm network modeling module, a dynamic perturbation identification module, and a single-cell mechanism analysis module. For example... Figure 2 As shown, the single-cell transcriptome data from atherosclerotic plaques were first visualized using two-dimensional dimensionality reduction: the central t-SNE plot projects single cells from vulnerable plaques in symptomatic patients (SYM) and stable plaques in asymptomatic patients (ASM) onto the same plane and colors them according to the main cell types; concentric rings surrounding the t-SNE plot label each cell type, sample origin (SYM / ASM), and cell cycle stage from the outside to the inside. Multiple cell populations were identified in the plaque sample, including vascular smooth muscle cells (SMC), smooth muscle-derived intermediate state cells (SEM), fibrochondrocytes (FC), fibroblasts, monocytes (Mono), macrophages (Macro), endothelial cells (EC), B cells, T cells, dendritic cells (DC), and mast cells; the scRNA-seq data used were from Columbia University Irving Medical Center and were reprocessed in this workflow for subsequent analysis. In the high-resolution network analysis stage, to overcome the inherent "sparse / dropout" problem in single-cell data, cells with highly similar transcriptome expressions are aggregated into "metacells" through K-nearest neighbor (KNN). Each metacell represents a similar cell state under the diurnal rhythm expression spectrum, thereby enabling the construction of diurnal rhythm gene co-expression networks according to cell type. Through weighted gene network analysis (WGCNA), the synergistic modeling between rhythm genes is achieved, resulting in diurnal rhythm gene co-expression networks composed of multiple gene modules such as M1, M2, M7, and M8.

[0037] Once the circadian rhythm gene co-expression network is constructed, when faced with a new user who needs to have their circadian rhythm assessed, the user's transcriptome data is collected. Then, multiple rhythm homeostatic gene modules in the constructed circadian rhythm gene co-expression network locate and extract the corresponding gene expression data from the transcriptome data for subsequent circadian rhythm perturbation monitoring and circadian rhythm disorder index calculation.

[0038] Furthermore, considering the association between circadian rhythms and certain diseases, such as Figure 3 As shown, the MyPROClock system can also incorporate transcriptome data from multiple disease progression stages, such as five coronary coronary disease progression stages: stable coronary syndrome (CCS), unstable angina (UA), non-ST-segment elevation myocardial infarction (NSTEMI), ST-segment elevation myocardial infarction (STEMI), and post-PCI recovery, totaling 456 samples. It models the transcriptome rhythm characteristics of different disease stages in a network form, forming rhythm snapshots in five states, and constructing a cross-sectional diurnal rhythm spectrum, thereby achieving rhythm network partitioning of disease stages.

[0039] Specifically, based on the "tipping point / mutation transition" theory, this invention proposes a method for identifying circadian rhythm transcriptomic drivers (CTDs) and characterizing their dynamic changes. The overall process is shown in the figure, and the Cohort A (bulk RNA-seq) with n=456 is used as an example for illustration: First, the subject samples are stratified according to different clinical states of atherosclerotic cardiovascular disease (ASCVD), including chronic coronary syndrome (CCS), unstable angina (UA), non-ST-segment elevation myocardial infarction (NSTEMI), ST-segment elevation myocardial infarction (STEMI), and post-percutaneous coronary intervention (Post-PCI), and peripheral blood transcriptome sequencing data of each sample are obtained; then, a co-expression network of circadian rhythm-related genes is constructed for each clinical state, and the circadian rhythm expression profile is divided into several network modules, such as M1, M2, M7, and M8, to characterize the rhythmic coordination structure in this state.

[0040] Step 16: Calculate the synergy, volatility, and network concentration of gene expression data based on the dynamic network biomarker algorithm. Determine the circadian rhythm disorder index of the user to be evaluated based on the synergy, volatility, and network concentration. The circadian rhythm disorder index is used to quantify the degree of circadian rhythm disorder of the user to be evaluated.

[0041] Among these, synergy, volatility, and network concentration can be determined through the following steps: Step 161: Calculate the average expression correlation of gene expression data to obtain the synergy of the users to be evaluated.

[0042] Step 162: Calculate the expression variance or dispersion of gene expression data to obtain the volatility of the user to be evaluated.

[0043] Step 163: Calculate the average connection strength between the rhythm homeostatic gene module corresponding to the gene expression data and other gene modules in the diurnal rhythm gene co-expression network to obtain the network concentration of the user to be evaluated.

[0044] Specifically, when extracting gene expression data corresponding to the rhythm homeostasis gene module of the user to be evaluated, the MyPROClock system uses dynamic network biomarker (DNB) theory to monitor perturbations in the rhythm system. For example, it calculates the average expression correlation, expression variance or dispersion, and average connection strength among gene expression data to determine the synergy, volatility, and network concentration of the user to be evaluated. The system then sums the data corresponding to these three indicators to determine the user's diurnal rhythm disorder index. For example... Figure 3 This invention introduces a Dynamic Network Biomarker (DNB) analysis framework to dynamically assess each module during the "stable-progressive-deteriorating" disease evolution process. The degree of network instability is quantified by calculating the Composite Index (CI) of each module. When the CI of a module peaks in the evolution trajectory, that stage is identified as a critical window near a critical transition point, indicating the transition of the system from a relatively reversible stage to an irreversible deterioration stage. Furthermore, this invention screens and evaluates the significance of Critical Transition Signals (CTS) near the critical transition points to obtain a set of key genes reflecting critical instability. These are identified as circadian rhythm transcriptome drivers (CTDs) and their corresponding dynamic trajectory characteristics, thereby enabling the characterization of circadian rhythm network changes under different ASCVD states and providing calculable molecular evidence for disease risk warning, stratified management, and intervention timing selection.

[0045] Step 17: Extract the diurnal rhythm phase features, diurnal rhythm amplitude features, and overall expression levels of rhythm homeostasis gene modules from the transcriptome data.

[0046] Step 18: Determine the circadian rhythm subtype of the user to be evaluated using a pre-trained machine learning classifier based on the circadian rhythm disorder index, circadian rhythm phase features, circadian rhythm amplitude features, and the overall expression level of the rhythm homeostasis gene module. The circadian rhythm subtype is constructed from the training transcriptome data using an unsupervised clustering algorithm.

[0047] Specifically, such as Figure 4 As shown, different rhythm subtypes exhibit significant distributional differences. Considering that each subtype displays different characteristics in disease stage, degree of coronary artery stenosis, and prognostic risk, the MyPROClock system provided by this invention can also classify and display the transcriptome data of the user to be evaluated based on data such as the diurnal rhythm disorder index, thus providing a quantitative basis for individualized assessment of clinical rhythm status. This invention conducts comparative analyses on publicly available transcriptome datasets GSE90074, GSE20680, GSE59867, GSE62646, and the self-built CohortA dataset. For the coronary artery stenosis severity data GSE90074 and GSE20680, the upper line plots show the average MyPROClock values ​​under different stenosis grades, indicating that MyPROClock progressively increases with increasing stenosis severity. The lower box plots show the distribution of MyPROClock in each grade sample, and the Kruskal-Wallis (KW) test was used to verify the significance of differences between groups. The results showed that p=1.3×10 for GSE90074. -4 The difference was significant, with p = 5.7 × 10⁻⁶ for GSE20680. -8 The differences were significant. Specifically, the stenosis classification of GSE90074 included: Grade 1 (all major vessels stenosis <10%), Grade 2 (single vessel stenosis 10%–70%), Grade 3 (single vessel stenosis >70%), Grade 4 (two vessels stenosis >70%), and Grade 5 (three vessels stenosis >70%). The stenosis classification of GSE20680 included: Grade 1 (luminal stenosis ≤25%), Grade 2 (luminal stenosis 25%–50%), and Grade 3 (stenosis ≥70% in one or more major vessels or ≥50% in two or more arteries). Furthermore, an analysis of the dynamic changes of GSE59867 and GSE62646 at different stages of acute myocardial infarction (MI) was conducted. The line graph in the figure shows that MyPROClock reaches its highest level on day 1 of MI, and then gradually decreases on day 5, 1 month, and 6 months, suggesting that the degree of rhythmic perturbation weakens as the condition stabilizes. The corresponding box plot shows the differences in the distribution of MyPROClock between different clinical stages, and the p=2.4×10 of GSE59867 was confirmed by the KW test. -12 GSE62646 has p=5.3×10 -3 The differences were significant. Furthermore, the results were validated using the self-built CohortA dataset of this invention for different ASCVD clinical states. Line graphs showed that MyPROClock increased in more severe atherosclerotic states, and box plots showed the distribution differences among clinical states. KW test confirmed p = 8.1 × 10⁻⁶. -3The differences were significant. These results demonstrate that MyPROClock of the present invention can consistently quantify changes in coronary artery stenosis aggravation, the evolution of acute and recovery phases of myocardial infarction (MI), and different ASCVD clinical states, providing calculable circadian rhythm-related molecular indicators for disease risk stratification and progression assessment.

[0048] Specifically, such as Figure 5 As shown, the MyPROClock system introduces a consensus clustering algorithm to classify rhythm types, resulting in multiple rhythm subtypes. For example, it uses ensemble consensus clustering based on various unsupervised algorithms such as k-means, hierarchical clustering, PAM, MSCUT, and SOM to perform clustering. The clustered samples, such as ATC & skmeans samples, are then classified into four rhythm subtypes based on the systematic differences in rhythm gene regulatory network activity and temporal phase: "Clock I", "Clock II", "Clock III", "Clock IV", "Clock V", "Clock VI", "Clock VII", "Clock VII", "Clock VIII ... "Clock III" and "Clock IV" correspond to different circadian rhythm imbalance patterns. Specifically, to achieve reproducible genotyping of the diurnal rhythm state in ASCVD patients, this invention uses the MyPROClock expression spectrum of the GSE59867 meta cohort with n=436 as a basis and employs a consensus clustering strategy to genotype the samples based on diurnal rhythm. Specifically, this invention constructs 25 classification schemes, which are obtained by combining five feature screening indicators (standard deviation SD, coefficient of variation CV, median absolute deviation MAD, ATC index, QCD index) with five hierarchical clustering algorithms (hclust, k-means, sk-means, PAM, mclust). Each scheme is iteratively clustered to evaluate the stability of the genotyping; among them, PAC (proportion of The Ambiguous Clustering (PAC) score is used to reflect the stability of the clustering results; a higher PAC value indicates less stable or unstable clustering. Based on the above comparisons, this invention selects ATC-kmeans as a representative clustering scheme, and its consistent clustering results are as follows: Figure 5As shown in (a): the heatmaps present the probability distribution of each sample being assigned to a specific circadian rhythm subtype, the silhouette score used to measure the similarity between a sample and its cluster, and the consensus score matrix used to quantify the consistency and stability of subtype assignment under different iterations. Based on this classification, this invention divides the samples into four circadian rhythm subtypes: ClockI, ClockII, ClockIII, and ClockIV, and further compares the MyPROClock horizontal distribution among different subtypes, as shown in (a). Figure 5 As shown in (b), there are significant differences in MyPROClock among the four subtypes, exhibiting a gradient relationship: ClockII>ClockIII>ClockI>ClockIV. The Kruskal–Wallis (KW) test showed p<2.2×10⁻⁶. -16 The differences are significant, thus proving that the present invention can achieve stable stratification and classification of the degree of circadian rhythm abnormalities at the population scale.

[0049] Based on this, the aforementioned rhythm subtypes can be classified using a pre-trained machine learning classifier, such as the miniClock classifier. Figure 6 As shown, the MyPROClock system integrates multiple machine learning algorithms such as DNN, SVM, RandomForest, KNN, and Treebag. Specificity, recall, precision, Kappa score, F1 index, and accuracy are determined through cross-validation and independent testing. Precision > 0.85 and F1 index > 0.8, indicating that multiple machine learning algorithms achieve high classification accuracy. Furthermore, it maintains stable generalization performance across multiple external queues, including GSE12334, GSE62646, GSE20680, GSE29532, and GSE21545.

[0050] In summary, the circadian rhythm assessment method provided by this invention represents a dual innovation in methodology and theoretical framework within the field of circadian rhythm biology. Traditional studies are often limited to the periodic analysis of a few clock genes, failing to reveal the synergistic mechanisms of rhythms at the genome-wide level. The MyPROClock system, by integrating multi-gene expression characteristics, establishes a dynamic network model of rhythm homeostasis-disorder-rebalancing, achieving for the first time a quantitative characterization of rhythm disorders from a systems biology perspective. This innovation changes the previous research model that relied on external phenotypes or single-time-point measurements, making the identification of rhythm disorders more objective, reproducible, and interpretable. Furthermore, the system introduces weighted gene co-expression network analysis (WGCNA) and dynamic network biomarkers (DNB) methods, effectively capturing the "critical turning point signal" of the rhythm system on the eve of imbalance, revealing the nonlinear evolution of the rhythm regulatory network. This mechanistic discovery not only broadens the theoretical boundaries of circadian rhythm research but also provides researchers with new analytical tools to explore the causal relationship between rhythm biology and disease progression.

[0051] Furthermore, the circadian rhythm assessment method provided by this invention successfully transforms complex molecular computational processes into an operable and visualized intelligent analysis platform. The miniClock platform, developed based on the MyPROClock core algorithm, automates the process from data input to result output. This system can quickly read standardized transcriptome data on a web page or in a local operating environment, automatically completing feature extraction, model computation, and visualization—the entire process takes only a few seconds, significantly improving analytical efficiency. The platform interface provides network visualizations of rhythm disorder indices, risk stratification results, and key rhythm gene modules, allowing users to directly view rhythm homeostasis trends and potential risk levels within a graphical interface. Compared to traditional research processes requiring complex programming and statistical analysis, this invention achieves engineered encapsulation of the algorithm model and clinically friendly visualization, offering significant advantages such as ease of operation, low cost, and high scalability.

[0052] Moreover, the MyPROClock system boasts a highly compatible and portable architecture. Its algorithm modules are adaptable to various data sources, including bulk RNA-seq, single-cell transcriptomics, peripheral blood samples, and public database data. A normalization mechanism maintains consistency in rhythmic characteristics across different platforms, ensuring comparability across research scenarios. The system also features a deep learning training interface, enabling self-optimization of model performance through continuous data accumulation, allowing for dynamic learning and iterative updates in future clinical applications. This self-evolving characteristic makes this invention not just an algorithm, but a continuously optimizing intelligent analysis ecosystem.

[0053] Furthermore, the rhythm disorder index output by the MyPROClock system allows researchers to construct interpretable quantitative associations between the MyPROClock Score and disease severity. Validation on multi-stage samples of atherosclerotic cardiovascular disease (ASCVD) shows an orderly increasing trend across different disease stages, peaking in non-ST-segment elevation myocardial infarction (NSTEMI), sensitively reflecting the molecular threshold of the disease transitioning from stable to unstable. Simultaneously, this index is significantly correlated with the risk of major adverse cardiovascular events (MACE), serving as an auxiliary indicator for clinical risk prediction and intervention timing. In other words, the rhythm disorder index can help physicians identify high-risk patient groups, guide the development of individualized treatment plans, and even monitor the effects of drug intervention and lifestyle modifications on the restoration of rhythm homeostasis. Moreover, the system has good universality and can be extended to multiple fields such as metabolic diseases, neuropsychiatric disorders, sleep rhythm disorders, and cancer rhythm therapy, providing a unified quantitative tool for the clinical translation of rhythm medicine.

[0054] In summary, the MyPROClock system has achieved a transformative leap from "rhythmic biology research" to "rhythmic medicine applications," promoting the development of basic science and providing a new, quantifiable, and operable tool for clinical disease management. Its unique algorithm system, broad adaptability, and powerful potential for intelligent expansion make it of significant value for promotion and application in scientific research, medicine, and the bioinformatics industry.

[0055] Example 2 Based on Example 1, robustness was validated using nine publicly available transcriptome datasets and multi-species models. Specifically, the distribution of the MyPROClock system's circadian rhythm disruption index across different species and tissues, including human peripheral blood transcriptomes from the GSE39445 dataset, rat heart transcriptomes from the GSE124870 dataset, mouse heart transcriptomes from the GSE43073 dataset, mouse liver transcriptomes from the GSE57830 dataset, mouse adipose tissue transcriptomes from the GSE35026 dataset, zebrafish pineal gland transcriptomes from the GSE13196 dataset, fruit fly head transcriptomes from the GSE39578 dataset, and red junglefowl pineal gland transcriptomes from the GSE21915 dataset, under different rhythm interference models such as sleep deprivation, circadian rhythm reversal, and ClOCK / Per2 mutation, is shown below. Figure 7 As shown.

[0056] Furthermore, the MyPROClock system demonstrates excellent performance in detecting circadian rhythms in different tissues and species, such as... Figure 8 and Figure 9 As shown, where Figure 8The experimental results of the MyPROClock system for detecting diurnal rhythms Figure 9 The experimental results of the MyPROClock system in detecting differential diurnal rhythm patterns were analyzed using transcriptomic data from human peripheral blood, primate suprachiasmatic nucleus, and mouse heart tissue. The system detected significant diurnal oscillation signals across multiple omics datasets; for example, in the GSE113883 dataset, the corresponding P=5.38×10⁻⁶. -5 The corresponding P value for the GSE98965 dataset is 7.48 × 10⁻⁶. -3 The GSE54650 dataset corresponds to P=3.54×10 -2 The MyPROClock system has been demonstrated to accurately reconstruct rhythmic variations in single-timepoint omics data and is applicable to various biological systems and tissue types. Furthermore, in human sleep deprivation models, circadian rhythm inversion models, and mutant mouse experiments, the MyPROClock system was able to identify significant changes in basal expression levels and amplitude. For example, under sleep deprivation conditions, the system detected a significant decrease in diurnal amplitude, which returned to normal in well-slept or wild-type mice. Simultaneously, in diurnal feeding restriction experiments, the system detected significant phase and amplitude reversals. These results demonstrate that MyPROClock can sensitively capture transcriptional fluctuations caused by rhythmic perturbations, achieving cross-individual and cross-model validation of rhythmic function.

[0057] The MyPROClock system demonstrates stable performance across 9 independent queues as follows: Figure 10 As shown, a standardized comparison of the MyPROClock scores for each cohort reveals a high degree of consistency in the distribution curves across different cohorts, indicating the robustness and generalization performance of the system across different populations, platforms, and disease subtypes. Furthermore, the consistency of rhythm states across cohorts is illustrated using quantile distributions, demonstrating that the MyPROClock index maintains good statistical stability under multi-center, multi-sample conditions. This indicates that rhythm imbalance has independent prognostic value, meaning the MyPROClock system possesses stable cross-cohort generalizability.

[0058] In addition, the MyPROClock system provided by this invention can also output and display that when a user uploads a single sample data, the MyPROClock system can calculate the diurnal rhythm disorder index and the probability values ​​of the four rhythm subtypes corresponding to the sample data, thereby realizing real-time interpretation.

[0059] Example 3 This application also provides a circadian rhythm assessment system for cardiovascular status, including: The transcription data acquisition module is used to acquire transcriptome data from a single time point for the user to be evaluated; The gene expression extraction module is used to locate and extract gene expression data from transcriptome data through a pre-set rhythm homeostasis gene module. The rhythm homeostasis gene module is constructed using a weighted gene co-expression network analysis algorithm, which is trained based on training transcriptome data extracted from cardiovascular state samples. The disorder index assessment module is used to calculate the synergy, volatility, and network concentration of gene expression data based on the dynamic network biomarker algorithm. Based on the synergy, volatility, and network concentration, the circadian rhythm disorder index of the user to be evaluated is determined. The circadian rhythm disorder index is used to quantify the degree of disorder of the user's circadian rhythm.

[0060] Specifically, the MyPROClock system, such as Figure 11 As shown, based on high-throughput transcriptome data and combined with systems biology and artificial intelligence algorithms, the MyPROClock system achieves accurate identification and interpretable quantification of circadian rhythm homeostasis disruption, filling the gap in existing technologies that cannot quantitatively describe rhythm disorders at the molecular level. Starting from the whole genome level, the MyPROClock system integrates approximately 2,000 rhythm-related genes to construct a multi-dimensional rhythm feature matrix. It then utilizes weighted gene co-expression network analysis (WGCNA) and dynamic network biomarker (DNB) models to identify key nodes and co-changing patterns of rhythm homeostasis disruption. By calculating parameters such as changes in gene correlation, network density, and amplitude drift within the module, the system generates a stable and reproducible individualized quantitative index—the MyPROClock Score—achieving, for the first time, a numerical and hierarchical description of rhythm disorders.

[0061] The MyPROClock system, primarily developed using Python and R, operates its core algorithm in a modular fashion. Users can initiate analysis by uploading a standardized gene expression matrix through a graphical interface. The system includes a data input module, a feature extraction module, a core algorithm module, a validation module, and an output module. The data input module supports high-throughput data from various sources, including bulk RNA-seq, single-cell RNA-seq, peripheral blood, and tissue samples, and features automatic cleaning and standardization. The feature extraction module extracts high-dimensional features based on a core rhythm gene library and constructs tissue- and species-specific weight matrices. The core algorithm module utilizes a WGCNA+DNB joint algorithm to identify rhythm homeostasis disruption signals and calculate individual rhythm disorder scores. The validation module ensures the algorithm's stability and transferability through cross-dataset and multi-species / multi-tissue external validation. The output module integrates the deep learning subsystem miniClock, enabling real-time prediction, hierarchical analysis, and visualization of rhythm disorder states.

[0062] The present invention also provides a computer device, including a memory, a processor, and a computer program stored in the memory. The processor executes the computer program to implement the steps in an embodiment of a method for assessing circadian rhythms in cardiovascular states. Specific implementation methods can be found in the method embodiments, and will not be repeated here.

[0063] Furthermore, the present invention also provides a non-transitory computer-readable storage medium containing instructions on which a computer program is stored. For example, a memory containing instructions that can be executed by a processor of a computer device to perform the above-described method. For example, the non-transitory computer-readable storage medium may be a ROM, random access memory (RAM), CD-ROM, magnetic tape, floppy disk, and optical data storage device, etc. When the computer program is executed by the processor, it can implement the steps in an embodiment of a method for assessing circadian rhythms in cardiovascular states. Specific implementation methods can be found in the method embodiments, which will not be repeated here.

[0064] Those skilled in the art will understand that embodiments of the present invention can provide methods, systems, or computer program products. Therefore, the present invention can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present invention can take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0065] This invention is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the invention. It will be understood that each block of the flowchart illustrations and / or block diagrams, as well as combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0066] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0067] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0068] It should be noted that the specific embodiments described above enable those skilled in the art to more fully understand the present invention, but do not limit the present invention in any way. Therefore, although the present invention has been described in detail in this specification and embodiments, those skilled in the art should understand that modifications or equivalent substitutions can still be made to the present invention; and all technical solutions and improvements that do not depart from the spirit and scope of the present invention are covered within the protection scope of the present invention patent. No reference numerals in the claims should be construed as limiting the scope of the claims. Any simple variations or equivalent substitutions of technical solutions that can be readily obtained by those skilled in the art within the scope of the technology disclosed in the present invention are within the protection scope of the present invention.

Claims

1. A circadian rhythm evaluation method for cardiovascular status, characterized by, The method comprises: obtaining transcriptome data of a single time point of a user to be evaluated; locating and extracting gene expression data from the transcriptome data through a pre-set circadian steady-state gene module, wherein the circadian steady-state gene module is constructed by a weighted gene co-expression network analysis algorithm trained based on training transcriptome data extracted from cardiovascular state samples; calculating the synergy, fluctuation and network concentration of the gene expression data based on a dynamic network biomarker algorithm, and determining a circadian rhythm disorder index according to the synergy, fluctuation and network concentration, wherein the circadian rhythm disorder index is used to quantify the degree of disorder of the circadian rhythm of the user to be evaluated.

2. The cardiovascular state-oriented circadian rhythm evaluation method according to claim 1, characterized by, Before identifying the circadian steady-state gene module from the transcriptome data through the pre-set circadian rhythm gene co-expression network, the method further comprises: calculating the correlation intensity of each of a plurality of gene modules constructed by the weighted gene co-expression network analysis algorithm with a pre-set rhythm feature, and determining the circadian steady-state gene module from the gene modules with a correlation intensity higher than a pre-set threshold.

3. The cardiovascular state-oriented circadian rhythm evaluation method according to claim 2, characterized by, Before calculating the correlation intensity of each of a plurality of gene modules constructed by the weighted gene co-expression network analysis algorithm with a pre-set rhythm feature, the method further comprises: obtaining training transcriptome data; aggregating a plurality of training transcriptome data into one meta cell based on a single cell clustering strategy to obtain a plurality of meta cell expression matrices; calculating the expression similarity between a plurality of the meta cell expression matrices to obtain a similarity matrix; performing power operation on the similarity matrix to obtain a weighted adjacency matrix, and calculating a topological overlap measure based on the weighted adjacency matrix; clustering genes in the training set of transcriptome data based on the topological overlap measure to obtain a clustering tree diagram; and dividing the clustering tree diagram based on gene aggregation degree to obtain a circadian rhythm gene co-expression network composed of a plurality of gene modules.

4. The cardiovascular state-oriented circadian rhythm evaluation method according to claim 3, characterized by, Obtaining training transcriptome data comprises: obtaining transcriptome data of different species; and obtaining transcriptome data of different tissues; constructing the training transcriptome data from the transcriptome data of different species and the transcriptome data of different tissues.

5. The cardiovascular state-oriented circadian rhythm evaluation method according to claim 3, characterized by, Aggregating a plurality of training transcriptome data into one meta cell based on a single cell clustering strategy to obtain a plurality of meta cell expression matrices comprises: clustering a plurality of cell subpopulations from the training transcriptome data according to cell types based on gene expression profiles; in each of the cell subpopulations, aggregating a plurality of cells with similar transcriptome characteristics into one meta cell, wherein the gene expression amount of the meta cell is the average or median of the gene expression amounts of all cells in the cell subpopulation.

6. The cardiovascular state-oriented circadian rhythm evaluation method according to claim 3, characterized by, Calculating the synergy, fluctuation and network concentration of the gene expression data based on a dynamic network biomarker algorithm comprises: calculating the average expression correlation of the gene expression data to obtain the synergy of the user to be evaluated; calculating the expression variance or dispersion of the gene expression data to obtain the fluctuation of the user to be evaluated; and calculating the network concentration of the gene expression data based on the dynamic network biomarker algorithm. calculating an average connection strength of the circadian rhythmicity gene module corresponding to the gene expression data and other gene modules in the circadian gene co-expression network, to obtain a network concentration of the user to be evaluated.

7. The cardiovascular state-oriented circadian rhythm evaluation method according to claim 3, characterized by, After determining the circadian rhythm disorder index of the user to be evaluated according to the cooperativity, fluctuation and network concentration, the method further comprises: extracting circadian phase features, circadian amplitude features and overall expression levels of the circadian rhythmicity gene module from the transcriptome data; determining a circadian rhythm subtype of the user to be evaluated by the circadian rhythm disorder index, circadian phase features, circadian amplitude features and overall expression levels of the circadian rhythmicity gene module through a pre-trained machine learning classifier, wherein the circadian rhythm subtype is constructed by the training transcriptome data through an unsupervised clustering algorithm.

8. A cardiovascular condition-oriented circadian rhythm evaluation system characterized by, comprise: a transcript data acquisition module configured to acquire single-time-point transcriptome data of a user to be evaluated; a gene expression extraction module configured to locate and extract gene expression data from the transcriptome data through a pre-set circadian rhythmicity gene module, wherein the circadian rhythmicity gene module is constructed by a weighted gene co-expression network analysis algorithm trained based on training transcriptome data extracted from cardiovascular state samples; a disorder index evaluation module configured to calculate cooperativity, fluctuation and network concentration of the gene expression data based on a dynamic network biomarker algorithm, and determine a circadian rhythm disorder index of the user to be evaluated according to the cooperativity, fluctuation and network concentration, wherein the circadian rhythm disorder index is used to quantify the degree of disorder of the circadian rhythm of the user to be evaluated.

9. A computer device comprising a memory, a processor, and a computer program stored on the memory, wherein the computer program comprises instructions that, when executed by the processor, cause the processor to perform the method of any one of claims 1-8. The processor executes the computer program to implement the steps of the cardiovascular state-oriented circadian rhythm evaluation method according to any one of claims 1 to 7.

10. A computer-readable storage medium having stored thereon a computer program, characterized in that, The computer program, when loaded by the processor, can execute the steps of the cardiovascular state-oriented circadian rhythm evaluation method according to any one of claims 1 to 7.

Images