A method for assessing the degree of cell senescence based on proteomic data and machine learning

Abstract

This invention relates to the field of biological aging assessment technology, specifically a method for assessing cellular aging based on proteomics data and machine learning. The method includes: constructing a cell type-specific candidate feature set based on single-cell transcriptomics data; obtaining a protein feature matrix based on plasma proteomics data; constructing independent machine learning-based cell type-specific aging prediction regression models for each cell type, using age as the response variable and the protein feature matrix as input; inputting the proteomics data of the sample to be predicted into the corresponding cell type-specific aging prediction model to obtain the predicted lifespan and the lifespan difference characterizing cellular aging for different cell types. This invention achieves quantitative assessment of the aging degree of different cell types in different organs under in vivo conditions by constructing a functional mapping bridge between single-cell transcriptomics and plasma proteomics.

Description

Technical Field

[0001] This invention relates to the field of biological aging assessment technology, specifically a method for assessing the degree of cellular aging based on proteomics data and machine learning. Background Technology

[0002] Aging is a degeneration of structure and function at multiple levels, including different organs and cells throughout the body, and gradually increases the risk of developing many chronic diseases over time. Aggregates of senescent cells often appear at the sites of disease and during the aging process. Cellular senescence is a process observed in many parts of the body throughout life, such as wound repair and embryonic tissue remodeling, where aging mechanisms are involved. However, if cellular senescence persists, the proteins produced during this process can damage the structure and function of many tissues and organs, leading to disease. Current research shows that specific organs, such as the brain, kidneys, and heart, exhibit significantly different susceptibility and resistance to age-related diseases in a given population. However, research on how diseases in specific organs affect cellular state and molecular levels remains very limited. Therefore, exploring the aging process at the cellular and molecular levels is of great significance.

[0003] Meanwhile, a major challenge in the prevention and treatment of aging-related diseases is the lack of reliable individual-level assessment and prediction. Current precision medicine research primarily focuses on identifying the genomic basis of different diseases, and these results have shown initial success. However, due to significant differences in the accumulation rate of senescent cells in different tissues, overall aging assessments, lacking precise resolution, can mask certain key local or early aging signals, hindering the exploration and inference of the mechanisms and associations between aging and disease at the holistic level. This limits a deeper understanding of the relevant mechanisms, thus increasing the need for reliable individual-level risk prediction tools.

[0004] Driven by this demand, the concept of biological age has gradually come into the view of researchers in the field of precision medicine. It is an important quantitative indicator of actual aging, besides chronological age, reflecting an individual's degree of aging based on their current cellular, tissue, and physiological functional state, and is influenced by factors such as genetics, lifestyle, and environment. The concept's development relies on the widespread application of machine learning in various medical and biological fields, including clinical research and the synthesis of biomolecules. By using and processing clinical and imaging information on a large scale, population models are created to attempt to understand the processes of disease and aging.

[0005] Several models have been proposed, demonstrating great potential in predicting aging trajectories and age-related disease risks by utilizing clinical chemical biomarkers related to certain life processes. However, given the complexity of the human aging trajectory, the specific roles of these few observed clinical chemical biomarkers in the mechanisms of aging are difficult to explain, limiting our understanding of the details of human aging mechanisms. Furthermore, most of these biomarkers have low specificity, further hindering their application to interpreting organ-specific or even more precise aging patterns. In recent years, single-cell RNA-seq technology has developed rapidly and gained widespread acceptance. Several studies have performed single-cell transcriptome sequencing on multiple organs throughout the body and have published the data. For example, CN113838531B discloses a method for assessing cellular senescence based on transcriptome data and machine learning strategies. This method trains and fits a model to known transcriptome data to obtain a cellular senescence scoring model, thus achieving the goal of predicting the degree of cellular senescence using only tissue sample transcriptome sequencing data. However, this method only uses a simple classification logistic regression strategy for senescence identification. In terms of principle, it cannot effectively identify key gene expression characteristics of cell types during senescence. In terms of application, it only uses publicly available single-cell transcriptome datasets, lacking integration with gene data measurable in different clinical situations, making it difficult to ensure adaptability and integration efficiency in the application stage. Summary of the Invention

[0006] To address the problems in existing technologies, this invention provides a method for assessing cellular aging based on proteomics data and machine learning. By constructing a mapping bridge between single-cell transcriptome RNA genes and plasma proteome, it is possible to accurately trace the origins of different cell types in different organs at the molecular level, enabling precise modeling and quantitative assessment of cellular aging.

[0007] This invention is achieved through the following technical solution: A method for assessing cellular senescence based on proteomics data and machine learning, comprising: Based on single-cell transcriptomics data, a set of cell type-specific candidate features was constructed. Based on plasma proteomics data, obtain protein feature matrices; Using age as the response variable, protein feature matrix as the input, and individual biological or demographic characteristics as covariates, a mutually independent cell type-level aging prediction regression model is constructed based on a specific candidate feature set for each cell type and using machine learning model building methods. The proteomics data of the sample to be predicted is input into the cell type-specific aging prediction model corresponding to the cell type to obtain the predicted lifespan and the lifespan difference characterizing cell aging for different cell types.

[0008] Preferably, the single-cell transcriptomics data are derived from multiple organs and tissues, and contain single-cell expression information of multiple cell types; Plasma proteomics data are derived from known population samples and contain expression information for various plasma proteins.

[0009] Preferably, based on single-cell transcriptomics data, a set of cell type-specific candidate features is constructed, including: Preprocessing single-cell transcriptomics data; Select any cell type from the preprocessed single-cell transcriptomics data as the target cell type; At the level of single-cell transcriptome gene expression data, the gene expression differences between single-cell sets of target cell types and single-cell sets of other cell types are compared, and differentially expressed characteristic gene sets of target cell types are screened and obtained. By mapping and integrating the set of differentially expressed characteristic genes and the set of plasma protein genes, a set of candidate features specific to the target cell type is constructed.

[0010] Preferably, the single-cell transcriptomics data is preprocessed, and the single-cell transcriptomics data is subjected to quality control, standardization and normalization, and the single cells are classified into cell types or subtypes based on cell annotation information or automatic annotation algorithms.

[0011] Preferably, during screening, genes of the target cell type identified by single-cell transcriptome are screened according to a preset threshold, and genes with high expression levels are selected proportionally as a set of differentially expressed characteristic genes.

[0012] Preferably, the protein feature matrix is ​​obtained based on plasma proteomics data, including: standardizing the plasma proteomics data, handling outliers, and imputing missing values ​​to obtain the protein feature matrix.

[0013] Preferably, the cell type-specific aging prediction regression model based on machine learning is a Lasso regression model with regularization constraints.

[0014] A system for implementing the aforementioned method for assessing cellular senescence based on proteomics data and machine learning, characterized in that it comprises: The feature acquisition module is used to construct a cell type-specific candidate feature set based on single-cell transcriptomics data and to obtain a protein feature matrix based on plasma proteomics data. The model building module is used to construct independent, machine learning-based cell type-specific aging prediction regression models for each cell type, with age as the corresponding variable, protein features as the input, and individual biological or demographic characteristics as covariates. The prediction output unit is used to input the proteomics data of the sample to be predicted into the cell type-specific aging prediction model of the corresponding cell type, and obtain the predicted lifespan and the lifespan difference characterizing cell aging for different cell types.

[0015] An electronic device includes a memory and a processor, the memory storing a computer program, the processor executing the computer program to implement the steps of the method.

[0016] A storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps of the method.

[0017] Compared with the prior art, the present invention has the following beneficial effects: This invention presents a method for assessing cellular senescence based on proteomics data and machine learning. First, it establishes a mapping bridge between single-cell transcriptome RNA genes and plasma proteome, enabling plasma proteomics to potentially achieve precise molecular-level tracing of different cell types in different organs. Furthermore, it constructs a set of target cell type-specific candidate features based on differential gene expression characteristics between target cell types and other cell types, and on gene expression levels within a predetermined high-expression range within the target cell type. Finally, based on various cell type-specific candidate proteomics features and machine learning methods, it achieves a precise assessment of the senescence state of different cell types. Attached Figure Description

[0018] Figure 1 This is a flowchart illustrating a method for assessing cellular senescence based on proteomics data and machine learning, according to the present invention. Detailed Implementation

[0019] The following specific embodiments illustrate the implementation of the present invention. Those skilled in the art can easily understand other advantages and effects of the present invention from the content disclosed in this specification.

[0020] Exemplary embodiments of the present invention will now be described with reference to the accompanying drawings. However, the present invention may be embodied in many different forms and is not limited to the embodiments described herein. These embodiments are provided to fully and completely disclose the invention and to fully convey its scope to those skilled in the art. The terminology used in the exemplary embodiments illustrated in the drawings is not intended to limit the invention. In the drawings, the same units / elements are referred to by the same reference numerals.

[0021] Unless otherwise stated, the terms used herein (including technical terms) have their common meaning as understood by one of ordinary skill in the art. Furthermore, it is understood that terms defined in commonly used dictionaries should be understood to have a meaning consistent with the context of their relevant field, and not to be interpreted as having an idealized or overly formal meaning.

[0022] The present invention will be further described in detail below with reference to specific embodiments. These descriptions are for explanation purposes only and are not intended to limit the scope of the invention.

[0023] Reference Figure 1 This invention discloses a method for assessing cellular senescence based on proteomics data and machine learning, comprising: S100, constructing a cell type-specific candidate feature set and protein feature matrix.

[0024] S101, based on single-cell transcriptomics data, constructs a set of cell type-specific candidate features, specifically: S1011, preprocessing single-cell transcriptomics data (from multiple organs and tissues, containing single-cell expression information of multiple cell types), namely: performing quality control, standardization and normalization on single-cell transcriptomics data, and classifying single cells into cell types or subtypes based on cell annotation information or automatic annotation algorithms.

[0025] S1012, Select any cell type from the preprocessed single-cell transcriptomics data as the target cell type; S1013, at the single-cell transcriptome gene expression data level, compare the gene expression differences between single-cell sets of the target cell type and single-cell sets of other cell types, and screen to obtain differentially expressed characteristic gene sets of the target cell type.

[0026] The differential expression characteristics are obtained by calculating the logarithmic fold change between the target cell type and other cell types, but are not limited to this calculation method.

[0027] During screening, genes of the target cell type are screened according to a preset threshold, and genes with high expression levels are selected in a certain proportion (such as the top 10%, but not limited to this proportion) as a set of differentially expressed characteristic genes.

[0028] The high-expression regions are quantile intervals determined based on gene expression level ranking. During screening, a coverage control parameter is set to adjust the screening intensity, balancing the number of features, stability, and information completeness.

[0029] S1014 involves mapping and integrating the differentially expressed characteristic gene set and the plasma protein gene set to construct a target cell type-specific candidate feature set. Specifically, the "gene names" in the differentially expressed characteristic gene set are matched with the "protein names" in the plasma protein gene set, meaning that the plasma proteins consistent with the characteristic genes of various cell types are retained within the current plasma protein gene set.

[0030] S102, based on plasma proteomics data, standardizes, handles outliers and imputes missing values ​​in plasma proteomics data to obtain a protein feature matrix.

[0031] Among them, plasma proteomics data are derived from known population samples (such as the UK Biobank UK, which provides expression information of about 3,000 plasma proteins from more than 50,000 people), and include expression information of a variety of plasma proteins.

[0032] S200 uses time and age as the response variable, protein feature matrix as input, and individual biological or demographic characteristics as covariates. Based on a specific set of candidate features for each cell type, it uses machine learning model building methods to construct mutually independent cell type-level aging prediction regression models.

[0033] The regression model includes a regularized regression model or a combination of multiple regression models. In this invention, the cell type-specific aging prediction regression model based on machine learning is a Lasso regression model with regularization constraints. Combined with a bootstrap sampling strategy, it can be repeatedly trained, thereby improving the model's stability and generalization ability.

[0034] S300 inputs the proteomics data of the sample to be predicted into the cell type-specific aging prediction model corresponding to the cell type, and obtains the predicted lifespan and the lifespan difference characterizing cell aging for different cell types.

[0035] During prediction, the model's stability is improved through repeated training or resampling. The lifespan difference is obtained by subtracting the predicted lifespan from the actual lifespan. For example, based on the predicted age and the population's lowest age, the corresponding biological age or aging difference is output, thereby achieving a quantitative assessment of aging status at the cellular level.

[0036] This invention is a further study based on high-throughput proteomics technology. In proteomics, the gene translation and expression processes in living organisms, including protein expression levels, post-translational modifications, and protein-protein interactions, can be studied. These expression processes themselves have the potential to uncover the origins of specific functional cell types because proteomics can identify key genes regulating life activities and construct disease regulatory networks by interpreting the interactions between regulatory genes. Among various types of proteomics, plasma proteomics stands out because it lacks protein synthesis processes; the maintenance of its components depends on the inflow and outflow of proteins from surrounding organs and cells. Furthermore, it is non-invasive, dynamic, relatively stable, and readily available. Combining these significant advantages, if a mapping between plasma protein components and cellular molecular-level resolution can be constructed, plasma proteomics can become an ideal material for studying aging mechanisms and disease development in vivo, participating in the construction of a cellular-level biological age clock to achieve minimally invasive assessment and tracking of the aging degree of any human organ.

[0037] This invention presents a method for assessing cellular senescence based on proteomics data and machine learning. It performs quality control and standardization on single-cell transcriptome data, constructs specific gene feature sets for different cell types, and maps and integrates these cell-specific features with plasma protein expression data to form a cell-origin-related protein feature matrix. A cell type-specific biological age prediction model is established based on bootstrap sampling and regularized regression constraints. By screening cell models whose predicted age is significantly correlated with actual age, the biological age and aging acceleration indicators for different cell types are output. This invention achieves quantitative assessment of the aging degree of different cell types in different organs under in vivo conditions by constructing a functional mapping bridge between single-cell transcriptome and plasma proteome. Compared with existing holistic aging prediction models, the model proposed in this invention can characterize aging heterogeneity at the cellular resolution level, improve the stability and interpretability of model predictions, and provide a new technical path for research on multi-organ aging mechanisms and early risk identification of aging-related diseases.

[0038] Example: A cellular-level aging modeling method based on multi-organ single-cell transcriptomics and plasma proteomics data, which is described in detail below: 1. Acquire multi-organ single-cell transcriptomics data (Tabula Sapiens V2) and human plasma proteomics data, and perform preprocessing; 1.1. Download expression matrix data and cell annotation information from a publicly available human multi-organ single-cell transcriptome database. The data includes 34 organs, 75 tissue sources, and a total of approximately 1,136,218 cells. 1.2 Perform quality control on the single-cell expression matrix to remove low-quality cells, low-complexity cells, and abnormally expressed cells; 1.3. The expression matrix is ​​standardized and normalized to eliminate sequencing depth differences; 1.4. Based on the cell annotation information provided by the original data, single cells are classified into different cell types; 1.5. Obtain plasma proteomics data from approximately 50,000 individuals, with approximately 2,900 protein characteristics. 1.6 Standardize the protein expression data, remove samples or proteins with a missing value ratio higher than the preset threshold, and fill in low-proportion missing values ​​using the KNN-impute method.

[0039] The data processing program is compiled based on Python. The program comes pre-installed with packages such as pandas, numpy, and scikit-learn, and runs on a Linux operating system.

[0040] 2. Construct a set of cell type-specific features; 2.1 For each target cell type, compare the set of single cells corresponding to that cell type with the other cell types, and calculate the log fold change (logFC) of each gene between the target cell type and other cell types. 2.2 Within the target cell type, gene expression levels are ranked, and the genes with the highest expression levels (top 10%) are selected as the set of differentially expressed characteristic genes for that cell type. 2.3. Map and integrate differential expression features with high expression features to construct a set of candidate features specific to the target cell type; 2.4 Adjust the feature screening intensity by setting the coverage (in what proportion of cells belonging to the cell type of interest, a certain differentially expressed feature can be identified) control parameter to balance the number of features and the integrity of information.

[0041] 3. Establish a cell-specific biological age prediction model; 3.1 The population sample is randomly split into a training set and an independent test set at a ratio of 8:2; 3.2. The Bootstrap autosampling method was used to perform 100 resampling iterations in the training set; 3.3. Based on the scikit-learn library in Python, an Lasso regression model with L1 regularization constraint is used for age prediction modeling. The principle formula is as follows:

[0042] Where: n is the number of observed samples; p is the number of features; This represents the actual age of the i-th sample; This represents the j-th candidate marker protein in the i-th sample; Represents the regression coefficient of the j-th candidate marker protein; Represents the regularization parameter; It is the intercept term.

[0043] The model was trained using data from the training set samples. The differentially expressed plasma protein levels for each cell type were used as predictors, while the chronological age of the sample was the target variable. To reduce the influence of potential confounding factors, the model included the following covariates: sex, ethnicity, Townsend Deprivation Index (TDI), fasting time before sample collection, season of blood collection, body mass index (BMI), and smoking status. 3.4. Optimize the model regularization parameters using grid search, and evaluate the model performance using five-fold cross-validation; 3.5. Apply the trained model to an independent test set to obtain the predicted age; 3.6. Also using functions from the scikit-learn library in Python, calculate the mean squared error (MSE) and mean absolute error (MAE) between the predicted age and the actual age to evaluate the magnitude of the error in the model's age prediction. Calculate the Pearson correlation coefficient R-value, the corresponding significance P-value, and the confidence interval CI of the P-value as model evaluation indicators to show the significance of the correlation between the model's predicted age and the actual age. 3.7. The 109 cell type models with R values ​​reaching the preset threshold (R>0.3) were retained for subsequent applications.

[0044] 4. Cellular-level senescence prediction and result output; Using established cell-specific models, the input data includes protein expression data of the samples to be predicted, as well as seven covariates: sex, ethnicity, Townsend poverty index of the living community, fasting time before sample collection, season of blood collection, body mass index, and smoking habits. The model outputs biological Lasso age predictions for different cell types in different samples. For each sample, each model calculates and outputs: 1) predicted age; 2) the age difference (AgeGap) between the predicted age and the actual Lowess age of the population; and 3) the AgeGap of each sample after Z-score standardization for each model. 3) is standardized to facilitate comparisons between different organ and cell models and can serve as an indicator of the degree of acceleration or slowing of cellular aging.

[0045] In this embodiment, taking pancreas T cells and kidney epithelial cells as examples, the Pearson correlation coefficient R values ​​of the model on the independent test set are 0.56 and 0.68, respectively, and the significance p-values ​​of the association analysis are much less than 0.001, indicating that the model has good age prediction ability.

[0046] This invention also discloses a system for implementing the aforementioned method for assessing cellular senescence based on proteomics data and machine learning, characterized in that it comprises: The feature acquisition module is used to construct a cell type-specific candidate feature set based on single-cell transcriptomics data and to obtain a protein feature matrix based on plasma proteomics data. The model building module is used to construct independent, machine learning-based cell type-specific aging prediction regression models for each cell type, with age as the corresponding variable, protein features as the input, and individual biological or demographic characteristics as covariates. The prediction output unit is used to input the proteomics data of the sample to be predicted into the cell type-specific aging prediction model of the corresponding cell type, and obtain the predicted lifespan and the lifespan difference characterizing cell aging for different cell types.

[0047] This invention also discloses a computer device, including a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, it implements the steps of the method. The processor may be a Central Processing Unit (CPU), or other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. It is the computing and control core of the terminal, and is suitable for implementing one or more instructions, specifically suitable for loading and executing one or more instructions from a computer storage medium to implement the corresponding method flow or corresponding function. This invention also discloses a computer-readable storage medium storing a computer program that, when executed by a processor, implements the steps of the method. The computer-readable storage medium is a memory device in a computer device used to store programs and data. It is understood that the computer-readable storage medium here can include both built-in storage media in the computer device and extended storage media supported by the computer device. The computer-readable storage medium provides storage space containing the terminal's operating system. Furthermore, this storage space also contains one or more instructions suitable for loading and execution by a processor; these instructions can be one or more computer programs (including program code). It should be noted that the computer-readable storage medium here can be high-speed RAM or non-volatile memory, such as at least one disk storage device. The processor can load and execute one or more instructions stored in the computer-readable storage medium.

[0048] Those skilled in the art will understand that embodiments of the present invention can be provided as 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.

[0049] 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, and 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 illustrations and / or block diagrams. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0050] 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.

[0051] 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.

[0052] 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 functions specified in one or more boxes. Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some or all of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of the present invention.

[0053] The above description is merely a preferred embodiment of the present invention and is not intended to limit the technical solution of the present invention in any way. Those skilled in the art should understand that, without departing from the spirit and principles of the present invention, the technical solution can be modified and replaced in several simple ways, and these modifications and replacements are all within the scope of protection covered by the claims.

Claims

1. A method for assessing cellular senescence based on proteomics data and machine learning, characterized in that, include: Based on single-cell transcriptomics data, a set of cell type-specific candidate features was constructed. Based on plasma proteomics data, obtain protein feature matrices; Using age as the response variable, protein feature matrix as the input, and individual biological or demographic characteristics as covariates, a mutually independent cell type-level aging prediction regression model is constructed based on a specific candidate feature set for each cell type and using machine learning model building methods. The proteomics data of the sample to be predicted is input into the cell type-specific aging prediction model corresponding to the cell type to obtain the predicted lifespan and the lifespan difference characterizing cell aging for different cell types.

2. The method for assessing cellular senescence based on proteomics data and machine learning according to claim 1, characterized in that, Single-cell transcriptomics data are derived from multiple organs and tissues and contain single-cell expression information of various cell types. Plasma proteomics data are derived from known population samples and contain expression information for various plasma proteins.

3. The method for assessing cellular senescence based on proteomics data and machine learning according to claim 2, characterized in that, Based on single-cell transcriptomics data, a set of cell type-specific candidate features was constructed, including: Preprocessing single-cell transcriptomics data; Select any cell type from the preprocessed single-cell transcriptomics data as the target cell type; At the level of single-cell transcriptome gene expression data, the gene expression differences between single-cell sets of target cell types and single-cell sets of other cell types are compared, and differentially expressed characteristic gene sets of target cell types are screened and obtained. By mapping and integrating the set of differentially expressed characteristic genes and the set of plasma protein genes, a set of candidate features specific to the target cell type is constructed.

4. The method for assessing cellular senescence based on proteomics data and machine learning according to claim 3, characterized in that, Preprocess single-cell transcriptomics data, perform quality control, standardization and normalization on the single-cell transcriptomics data, and classify single cells into cell types or subtypes based on cell annotation information or automatic annotation algorithms.

5. The method for assessing cellular senescence based on proteomics data and machine learning according to claim 3, characterized in that, During the screening process, genes of the target cell type identified by single-cell transcriptomics are screened according to a preset threshold, and genes with high expression levels are selected proportionally as a set of differentially expressed characteristic genes.

6. The method for assessing cellular senescence based on proteomics data and machine learning according to claim 1, characterized in that, Based on plasma proteomics data, a protein feature matrix is ​​obtained, including standardization, outlier handling, and missing value imputation of the plasma proteomics data.

7. The method for assessing cellular senescence based on proteomics data and machine learning according to claim 1, characterized in that, The cell type-specific aging prediction regression model based on machine learning is a Lasso regression model with regularization constraints.

8. A system for implementing the method for assessing cellular senescence based on proteomics data and machine learning as described in any one of claims 1 to 7, characterized in that, include: The feature acquisition module is used to construct a set of cell type-specific candidate features based on single-cell transcriptomics data; And, based on plasma proteomics data, obtain protein feature matrices; The model building module is used to construct independent cell type hierarchical aging prediction regression models based on a specific candidate feature set for each cell type, with age as the response variable, protein feature matrix as the input, and individual biological or demographic characteristics as covariates, using machine learning model building methods. The prediction output unit is used to input the proteomics data of the sample to be predicted into the cell type-specific aging prediction model of the corresponding cell type, and obtain the predicted lifespan and the lifespan difference characterizing cell aging for different cell types.

9. An electronic device comprising a memory and a processor, wherein the memory stores a computer program, characterized in that, When the processor executes the computer program, it implements the steps of the method according to any one of claims 1 to 7.

10. A storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by a processor, it implements the steps of the method according to any one of claims 1 to 7.

Images