A method of identifying a cell subpopulation associated with a disease phenotype
By constructing protein-protein interaction networks and performing network proximity analysis, we can identify cell subpopulations associated with disease phenotypes, solving the problem of inaccurate identification with small sample sizes in existing technologies and achieving precision in identifying disease pathogenesis mechanisms and targeted therapies.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- NORTHEAST FORESTRY UNIV
- Filing Date
- 2023-07-12
- Publication Date
- 2026-05-15
AI Technical Summary
Existing single-cell sequencing technologies struggle to effectively identify cell subpopulations associated with disease phenotypes, especially lacking statistical power in small sample sizes, which affects the accuracy of disease pathogenesis research and targeted therapy.
By constructing a protein-protein interaction network, the distance between cell modules and sample modules is calculated. Combined with disease phenotypic labels, cell subpopulations with significantly smaller distance distributions than the background are identified. Using data from the Gene Expression Omnibus, The Cancer Genome Atlas, and MINT databases, gene signatures are extracted and network proximity analysis is performed.
It can significantly identify cell subpopulations that drive disease phenotypic changes, improving the accuracy of disease pathogenesis research and the specificity of targeted therapy, while avoiding the inaccuracies caused by unsupervised clustering.
Smart Images
Figure CN116959562B_ABST
Abstract
Description
Technical Field
[0001] This invention belongs to the field of biomedicine, specifically relating to a method for identifying cell subpopulations associated with disease phenotypes. Background Technology
[0002] The advent of single-cell sequencing technology has revolutionized the field of biological research. Over the past decade, with the rapid development of single-cell sequencing technology and the explosive growth of single-cell data, single-cell analysis has received increasing attention in life sciences and precision medicine. Today, we can obtain genomic, transcriptomic, and epigenomic data from individual cells and conduct further research on cellular DNA, RNA, histone modifications, chromosome accessibility, and DNA methylation at the whole-genome level. Single-cell RNA sequencing is at the forefront of these methods, with throughput increasing from less than 100 cells initially to tens or even hundreds of thousands of cells in a single experiment. Single-cell RNA sequencing has also made significant progress in processing trace amounts of RNA, improving the stability of single-cell RNA sequencing on small samples.
[0003] The general approach to single-cell data analysis involves using methods such as standardization, feature selection, dimensionality reduction, and clustering to identify cell types and deduce cell differentiation trajectories in heterogeneous tissues. Single-cell sequencing enables in-depth analysis of tissue heterogeneity and identification of key cell populations influencing disease phenotypes at the individual cell level, facilitating the elucidation of disease pathogenesis and outcome-related mechanisms and deepening our understanding of disease progression. While many existing methods can correlate cell clusters with phenotypes, they cannot link individual cells to specific phenotypes. Furthermore, most current single-cell experiments involve fewer than 20 patient samples, lacking the statistical power to identify phenotype-related cell subpopulations. Studies have confirmed that changes in a small subset of key cells often drive alterations in disease phenotypes; therefore, identifying these phenotype-related cell subpopulations is crucial for understanding disease mechanisms and developing targeted therapies. Summary of the Invention
[0004] The problem this invention aims to solve is to identify phenotype-related cell subpopulations, and proposes a method for identifying disease phenotype-related cell subpopulations.
[0005] This invention is achieved through the following technical solution:
[0006] A method for identifying cell subpopulations associated with a disease phenotype includes the following steps:
[0007] A method for identifying cell subpopulations associated with a disease phenotype includes the following steps:
[0008] S1. Single-cell RNA sequencing data of diseases were collected from the Gene Expression Omnibus database to obtain single-cell expression matrices. Bulk expression matrices of diseases and corresponding phenotypic labels were collected from The Cancer Genome Atlas database. Human protein-protein interaction data were downloaded from the MINT database to construct protein-protein interaction networks.
[0009] S2. Extract the gene signatures of cells and samples from the single-cell expression matrix and the bulk expression matrix of the disease obtained in step S1, and map the gene signatures of cells and samples into the protein interaction network to form corresponding cell modules and sample modules.
[0010] S3. Based on the shortest path length in the protein-protein interaction network, calculate the distance between each cell module and each sample module in the protein-protein interaction network mapped in step S2, and determine the set of multiple sample modules as the sample module set of the disease phenotype based on the phenotypic label.
[0011] S4. Based on the sample module set of the disease phenotype obtained in step S3, calculate the distance between the cell module and the sample module set of the disease phenotype.
[0012] S5. Create a background distance distribution to assess the statistical significance of the distance between cell modules and the sample module set of disease phenotypes, and identify cells whose distance to the sample module set of disease phenotypes is significantly smaller than that of the background distance distribution.
[0013] Furthermore, the single-cell expression matrix and the bulk expression matrix of the disease mentioned in step S1 are normalized by TPM / FPKM, with genes as rows of the matrix and cells or samples as columns of the matrix.
[0014] Furthermore, the genetic signature of the cells and samples described in step S2 consists of the top 150 genes with the highest relative abundance.
[0015] Furthermore, the expression for calculating the distance between each cell module and each sample module in step S3 is as follows:
[0016]
[0017] Where S is the protein set of the sample module, C is the protein set of the cell module, d(S,C) is the path length between the cell module and the sample module, d(s,c) is the shortest path between node s and node c in the network, s is the network node in the sample module, and c is the network node in the cell module.
[0018] The shortest path length is determined by calculating the shortest path between nodes in the network.
[0019] Furthermore, the formula for calculating the distance between the cell module and the sample module set of the disease phenotype in step S4 is as follows:
[0020]
[0021] Where P is the set of sample modules related to disease phenotype, and d(P,C) is the distance between the cell module and the disease phenotype module set.
[0022] Furthermore, in step S5, the background distance distribution is created by randomly selecting a set of proteins in the network that matches the size of the original cell module.
[0023] Furthermore, in step S5, statistical significance is determined by calculating the number of random distances in the background distance distribution that are lower than the actual distance, dividing it by the total number of random cell modules, to determine the empirical P-value of the actual distance.
[0024] Furthermore, the disease phenotype labels of the samples should be matched with the binary grouping of the bulk expression matrix of the disease.
[0025] The beneficial effects of this invention are:
[0026] This invention discloses a method for identifying cell subpopulations associated with disease phenotypes. Based on network proximity, this method can significantly identify cell subpopulations driving disease phenotype changes, which is beneficial for research on disease pathogenesis and cell therapy targeting specific cell populations. This invention integrates biological molecular genetic information and biological network topology information, resulting in a more comprehensive and unbiased predictive capability. This study eliminates the need for unsupervised clustering of single-cell data, avoiding inaccuracies caused by manually setting the number of clusters. Attached Figure Description
[0027] Figure 1 This is a flowchart of a method for identifying cell subpopulations associated with disease phenotypes according to the present invention;
[0028] Figure 2 This is an unsupervised clustering diagram of single-cell RNA sequencing data from head and neck squamous cell carcinoma, used in a method for identifying cell subpopulations associated with disease phenotypes as described in this invention.
[0029] Figure 3 The cell subpopulations associated with head and neck squamous cell carcinoma identified by the method for identifying disease phenotype-related cell subpopulations described in this invention;
[0030] Figure 4 The percentage of cell types in the cell subpopulations identified by the method for identifying cell subpopulations related to disease phenotypes according to the present invention;
[0031] Figure 5 This is the ROC curve of the prediction results of the method for identifying cell subpopulations related to disease phenotypes according to the present invention. Detailed Implementation
[0032] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and specific embodiments. It should be understood that the specific embodiments described herein are only for explaining the invention and are not intended to limit the invention; that is, the described specific embodiments are merely a part of the embodiments of the invention, and not all of them. The components of the specific embodiments of the invention described and shown in the accompanying drawings can generally be arranged and designed in various different configurations, and the invention may also have other embodiments.
[0033] Therefore, the following detailed description of specific embodiments of the invention provided in the accompanying drawings is not intended to limit the scope of the claimed invention, but merely to illustrate selected specific embodiments of the invention. All other specific embodiments obtained by those skilled in the art based on these specific embodiments without inventive effort are within the scope of protection of this invention.
[0034] To further understand the invention's content, features, and effects, the following specific embodiments are provided, along with accompanying drawings. Figure 1 - Appendix Figure 5 Detailed explanation is as follows: Specific implementation method one:
[0036] A method for identifying cell subpopulations associated with a disease phenotype includes the following steps:
[0037] S1. Single-cell RNA sequencing data of diseases were collected from the Gene Expression Omnibus database to obtain single-cell expression matrices. Bulk expression matrices of diseases and corresponding phenotypic labels were collected from The Cancer Genome Atlas database. Human protein-protein interaction data were downloaded from the MINT database to construct protein-protein interaction networks.
[0038] Furthermore, the single-cell expression matrix and the bulk expression matrix of the disease mentioned in step S1 are normalized by TPM / FPKM, and genes are used as rows of the matrix, and cells or samples are used as columns of the matrix.
[0039] S2. Extract the gene signatures of cells and samples from the single-cell expression matrix and the bulk expression matrix of the disease obtained in step S1, and map the gene signatures of cells and samples into the protein interaction network to form corresponding cell modules and sample modules.
[0040] Furthermore, the genetic signature of the cells and samples described in step S2 consists of the top 150 genes with the highest relative abundance;
[0041] S3. Based on the shortest path length in the protein-protein interaction network, calculate the distance between each cell module and each sample module in the protein-protein interaction network mapped in step S2, and determine the set of multiple sample modules as the sample module set of the disease phenotype based on the phenotypic label.
[0042] Furthermore, the expression for calculating the distance between each cell module and each sample module in step S3 is as follows:
[0043]
[0044] Where S is the protein set of the sample module, C is the protein set of the cell module, d(S,C) is the path length between the cell module and the sample module, d(s,c) is the shortest path between node s and node c in the network, s is the network node in the sample module, and c is the network node in the cell module.
[0045] The shortest path length is determined by calculating the shortest path between nodes in the network;
[0046] S4. Based on the sample module set of the disease phenotype obtained in step S3, calculate the distance between the cell module and the sample module set of the disease phenotype.
[0047] Furthermore, the formula for calculating the distance between the cell module and the sample module set of the disease phenotype in step S4 is as follows:
[0048]
[0049] Where P is the set of sample modules related to disease phenotype, and d(P,C) is the distance between the cell module and the disease phenotype module set;
[0050] S5. Create a background distance distribution to assess the statistical significance of the distance between cell modules and the sample module set of disease phenotypes, and identify cells whose distance to the sample module set of disease phenotypes is significantly smaller than that of the background distance distribution.
[0051] Furthermore, in step S5, the background distance distribution is created by randomly selecting a set of proteins in the network that matches the size of the original cell module.
[0052] Furthermore, in step S5, statistical significance is determined by calculating the number of random distances in the background distance distribution that are lower than the actual distance, dividing it by the total number of random cell modules, to determine the empirical P-value of the actual distance.
[0053] Furthermore, the disease phenotype labels of the samples should be matched with the binary grouping of the bulk expression matrix of the disease.
[0054] Furthermore, a single-cell RNA sequencing dataset for head and neck squamous cell carcinoma was downloaded from the Gene Expression Omnibus (GEO: accession number: GSE103322) database. This single-cell RNA sequencing data came from 18 patients with oral tumors; after removing cells from lymph nodes and those with unidentifiable cell types, 4244 cells remained. Figure 2 As shown. Bulk expression data of TCGA-HNSC and binary phenotypic tags matched to the bulk dataset were downloaded using the GDCRNATools R package. Read counts for each gene were further converted to transcripts per million (TPM). 69,567 experimentally validated human protein-protein interaction data were downloaded from the MINT database as real-world data.
[0055] Furthermore, the method of this embodiment identified 46 cells associated with the head and neck squamous cell carcinoma phenotype, such as... Figure 3 As shown, 97.8% of them are malignant cells, and the rest are fibroblasts, such as... Figure 4 As shown in the figure. ROC analysis showed that the AUC value obtained by this method was 0.86, indicating that it has high sensitivity and specificity in identifying head and neck squamous cell carcinoma-related cells, such as... Figure 5 As shown.
[0056] It should be noted that relational terms such as "first" and "second" are used merely to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitations, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element.
[0057] Although this application has been described above with reference to specific embodiments, various modifications can be made and components can be replaced with equivalents without departing from the scope of this application. In particular, as long as there is no structural conflict, the features in the specific embodiments disclosed in this application can be combined with each other in any way. The lack of an exhaustive description of these combinations in this specification is merely for the sake of brevity and resource conservation. Therefore, this application is not limited to the specific embodiments disclosed herein, but includes all technical solutions falling within the scope of the claims.
Claims
1. A method for identifying cell subpopulations associated with a disease phenotype, characterized in that, Includes the following steps: S1. Single-cell RNA sequencing data of diseases were collected from the Gene Expression Omnibus database to obtain single-cell expression matrices. Bulk expression matrices of diseases and corresponding phenotypic labels were collected from The Cancer Genome Atlas database. Human protein-protein interaction data were downloaded from the MINT database to construct protein-protein interaction networks. The single-cell expression matrix and the bulk expression matrix of the disease mentioned in step S1 are normalized by TPM / FPKM, and genes are the rows of the matrix, and cells or samples are the columns of the matrix. S2. Extract the gene signatures of cells and samples from the single-cell expression matrix and the bulk expression matrix of the disease obtained in step S1, and map the gene signatures of cells and samples into the protein interaction network to form corresponding cell modules and sample modules. The genetic signature of the cells and samples described in step S2 consists of the top 150 genes with the highest relative abundance; S3. Based on the shortest path length in the protein-protein interaction network, calculate the distance between each cell module and each sample module in the protein-protein interaction network mapped in step S2, and determine the set of multiple sample modules as the sample module set of the disease phenotype based on the phenotypic label. The expression for calculating the distance between each cell module and each sample module in step S3 is as follows: ； in, For the protein set of the sample module, A collection of proteins for a cellular module. The path length between the cell module and the sample module. For nodes in the network and nodes The shortest path between them, where s is the network node in the sample module and c is the network node in the cell module; The shortest path length is determined by calculating the shortest path between nodes in the network; S4. Based on the sample module set of the disease phenotype obtained in step S3, calculate the distance between the cell module and the sample module set of the disease phenotype. The formula for calculating the distance between the cell module and the sample module set of the disease phenotype in step S4 is as follows: ； in, A collection of sample modules related to disease phenotypes. This represents the distance between the set of cell modules and the set of disease phenotype modules. S5. Create a background distance distribution to assess the statistical significance of the distance between cell modules and the sample module set of disease phenotypes, and identify cells whose distance to the sample module set of disease phenotypes is significantly smaller than that of the background distance distribution. In step S5, the background distance distribution is created by randomly selecting a set of proteins in the network that matches the size of the original cell module. In step S5, statistical significance is determined by calculating the number of random distances in the background distance distribution that are lower than the actual distance, dividing it by the total number of random cell modules, to determine the empirical P-value of the actual distance. The cell subpopulation in question is a subpopulation associated with head and neck squamous cell carcinoma.
2. The method for identifying cell subpopulations associated with disease phenotypes according to claim 1, characterized in that, The disease phenotype labels of the samples should match the binary groupings of the disease's bulk expression matrix.