CN116486911A - Processing method and system for respiratory disease data - Google Patents

Abstract

The invention discloses a method, a system, equipment and a computer readable storage medium for processing respiratory disease data, wherein the method comprises the following steps: acquiring single cell sequencing sequence data to be analyzed; processing the single-cell sequencing sequence data to be analyzed by adopting a machine learning method to obtain PAS scoring matrix of a cell pathway and PAS of the cell pathway; acquiring genetic association data of respiratory diseases, and processing the genetic association data to obtain path data with SNPs annotation; performing statistical analysis processing on PAS of the cell pathway and the pathway data with SNP annotation to obtain an estimation coefficient; multiplying the estimation coefficient by PAS and then summing to obtain a genetic related pathway activity score gPAS of the cell; outputting the genetically related pathway activity score gPAS.

Description

A method and system for processing respiratory system disease data

technical field

The present invention relates to the technical field of gene sequencing, and more specifically, to a method and system for processing respiratory disease data.

Background technique

Respiratory system disease is a common and frequently-occurring disease. The main lesions are in the trachea, bronchi, lungs and chest cavity. Mild cases often cause coughing, chest pain, and respiratory problems. Severe cases have dyspnea, hypoxia, and even death due to respiratory failure. The death rate in the city is the third, while in the countryside it is the first. What should be paid more attention to is that due to air pollution, smoking, population aging and other factors, the morbidity and mortality of chronic obstructive pulmonary disease (COPD for short, including chronic bronchitis, emphysema, and pulmonary heart disease), bronchial asthma, lung cancer, pulmonary diffuse interstitial fibrosis, and pulmonary infection have increased at home and abroad.

Coronaviruses are a large family of viruses known to cause more serious illnesses such as colds, Middle East Respiratory Syndrome (MERS) and Severe Acute Respiratory Syndrome (SARS). Common signs of a person infected with a coronavirus include respiratory symptoms, fever, cough, shortness of breath, and dyspnea. In more severe cases, infection can lead to pneumonia, severe acute respiratory syndrome, kidney failure, and even death. Many of the symptoms of coronavirus-caused illness are manageable, so treatment should be tailored to the patient's clinical condition. In addition, ancillary care for infected people can be very effective, and self-protection includes: maintaining basic hand and respiratory hygiene, adhering to safe eating habits, etc. Understanding the influence of host genetic composition on the immune response to severe infection could aid in the development of effective vaccines and therapeutics to control the associated respiratory disease pandemic. With the rapid development of sequencing technology, single-cell sequencing technology has brought more comprehensive opportunities to reveal the relevant mechanisms of related respiratory diseases.

Identifying key cell subpopulations associated with complex diseases or traits using single-cell RNA sequencing (scRNA-seq) technology is critical for understanding complex disease mechanisms. However, scRNA-seq data does not allow large-scale sequencing due to its high cost and low-throughput characteristics, and most current research samples based on single cells do not exceed 20, resulting in limited statistical power and cannot accurately reveal risk subsets associated with diseases or characteristics in cell subpopulations. In addition, scRNA-seq data is characterized by high sparsity, technical noise, and variance instability at the gene level. Genetic association data such as: (genome-wide association studies, GWAS) are widely used in the study of different complex diseases or traits, and correlating scRNA-seq data with phenotype-related genetic information from GWAS of large-scale samples is considered to be a practical and effective method to reveal the genetic molecular mechanisms of complex diseases or traits at single-cell resolution.

Methods that combine GWAS with scRNA-seq data to identify cell types associated with complex diseases include methods such as LDSC-SEG, MAGMA, RolyPoly, but the above methods require extensive tuning of parameters to annotate cell types with known marker genes and largely ignore the internal heterogeneity of each cell type. In addition, existing techniques can identify genes with high expression levels, but a potential pitfall is that excessive focus on highly expressed genes underestimates the functional role of relatively low expressed genes that are important for revealing cell fate.

Contents of the invention

The present invention aims to solve at least one of the technical problems existing in the prior art. To this end, the present invention provides a method and system for processing respiratory disease data; the method of the present invention infers genes, cells, etc. related to respiratory diseases by inventing a scoring method based on single-cell pathways, combining scRNA-seq data and genetic association data, digging deep into the life laws hidden behind single-cell sequencing data, and determining the potential connections between genes, cells, cell subgroups, biological pathways, etc., and respiratory diseases.

The present application discloses a method for processing respiratory disease data, including:

Obtain the single-cell sequencing sequence data to be analyzed;

Processing the single-cell sequencing sequence data to be analyzed by using a machine learning method to obtain a PAS scoring matrix of cell pathways and a PAS of cell pathways;

Obtain genetic association data of respiratory diseases, process the genetic association data, and obtain pathway data with SNPs annotation;

performing statistical analysis on the PAS of the cellular pathway and the pathway data with SNP annotations to obtain estimated coefficients;

The estimated coefficient is multiplied by PAS and then summed to obtain the cell's genetic-related pathway activity score gPAS;

Output the genetic related pathway activity score gPAS.

The step of performing statistical analysis on the PAS of the cell pathway and the pathway data with SNP annotations to obtain an estimated coefficient includes:

Obtain the genetic effect value of all SNPs in the single pathway data based on the pathway data with SNPs annotation;

Based on the multigene regression model of the genetic association data of respiratory diseases, based on the PAS and the genetic effect value, the distribution of the genetic effect value is estimated to obtain an estimated coefficient;

Optionally, the formula for obtaining the genetic effect value is: Among them, β represents the theoretical effect size vector of m SNPs, ε represents the random environmental error, R represents the LD matrix, and X ^T represents the standard genotype of the SNPs in the genetic association data sample;

Optionally, the manner of obtaining the estimated coefficients includes:

where τi _,j represents the estimated coefficient of pathway i in cell j, _τ0 represents the intercept term, ^σ2 represents the variance of the SNP effect size in the pathway, Indicates weighted PAS.

The acquisition formula of the genetic related pathway activity score gPAS (gPj) is:

Among them, the is the estimated coefficient after optimization.

The method also includes: performing a correlation analysis and sorting on the genetic related pathway activity score gPAS and the gene expression of each cell, and screening out N trait-related genes;

Optionally, the method of performing correlation analysis and sorting on the genetic correlation pathway activity score gPAS and the gene expression of each cell includes: determining the correlation between the expression of a single gene and the gPAS by Pearson correlation coefficient (PCC), and sorting the genes according to the correlation to obtain the N trait-related genes;

Optionally, the N trait-related genes are the top 1000 or last 1000 trait-related genes sorted according to the descending or ascending order of correlation;

Optionally, the N trait-related genes include one or more of the following: CALM3, PIK3R1, IL32, CD3E, B2M, PRS29, and GZMB.

The method also includes: calculating the trait-related score TRS of each cell according to the N trait-related genes; performing clustering according to the trait-related score TRS and the horizontal P value of a single cell to obtain trait-related cells related to respiratory diseases with different levels of severity;

Optionally, calculate the trait-related score TRS of the N trait genes by using a cell scoring method;

Optionally, the different grades of severity of the respiratory diseases include mild, moderate and severe.

The method further includes: obtaining trait-related cell types or subgroups based on a block bootstrap method.

The method further includes: sorting the genetic-related pathway activity score gPAS, and obtaining trait-related pathways according to the statistically important values according to the sorting results and the P values of the pathways at the cell type level.

detect new Application of products of CD8+T cell subsets in the preparation of products for diagnosing respiratory diseases.

A device for processing respiratory disease data, the device comprising: a memory and a processor;

The memory is used to store program instructions; the processor is used to call the program instructions, and when the program instructions are executed, is used to execute the above-mentioned method for processing respiratory system disease data.

A system for processing respiratory disease data, comprising:

an acquisition unit, configured to acquire single-cell sequencing sequence data to be analyzed;

The first processing unit is configured to process the single-cell sequencing sequence data to be analyzed by using a machine learning method to obtain a PAS scoring matrix of cell pathways and a PAS of cell pathways;

The second processing unit is used to obtain genetic association data of respiratory diseases, and process the genetic association data to obtain pathway data with SNPs annotation;

The third processing unit is used to perform statistical analysis on the PAS of the cell pathway and the pathway data with SNP annotations to obtain estimated coefficients;

The fourth processing unit is used to multiply the estimated coefficient by PAS and then sum to obtain the genetic-related pathway activity score gPAS of the cell;

An output unit, configured to output the genetic correlation pathway activity score gPAS.

A computer-readable storage medium, on which a computer program is stored, and when the computer program is executed by a processor, the above-mentioned method for processing respiratory disease data is realized.

The application has the following beneficial effects:

1. This application innovatively discloses a processing method for respiratory system disease data that combines single-cell sequencing data and genetic correlation data, which can infer genes, cells, cell subgroups and related biological pathways related to respiratory system diseases from a deep and multi-dimensional perspective. Understanding the impact of the above-mentioned host genetic components on the immune response to severe infection is helpful for the development of effective vaccines and therapeutic methods to control the pandemic and contribute to the research of respiratory system diseases; The functional role of different genes of the pathway to obtain a stable cell state significantly increases statistical power, biological interpretability, and reproducibility of results; it overcomes the limitations of known annotated cell types, and may discover new genetically related subpopulations and key genes or pathways of cell types, which are widely used and practical. For example: through this plan, it can be concluded that new gene drives that can be prioritized CD8+ T cell subsets may play an important role in mediating immune responses in patients with severe respiratory diseases.

2. This application innovatively discloses a single-cell pathway-based scoring method, which adopts a multigene regression model to reveal genes and cell subgroups related to traits by using scRNA-seq data transformed by pathway activity and genetic association research data; it effectively overcomes the problem that the current identification of genes and cell subgroups related to polygenic risk of complex diseases is largely hindered by the small sample size and high sparsity of scRNA-seq data, resulting in limited statistical power and the inability to accurately reveal risk subsets related to diseases or characteristics in cell subgroups. This method digs deep into the life laws hidden behind the single-cell sequencing data, and conducts in-depth analysis from multiple dimensions such as the relationship between population genetic mutations and diseases and single-cell sequencing gene abundance information, which greatly improves the accuracy and depth of data analysis.

3. This application is based on large-scale simulation and real data, and uses the above scoring method to combine scRNA-seq data and genetic association data, which can effectively overcome the problem that a large number of adjustment parameters are required in the prior art to conveniently annotate cell types with known marker genes, and the internal heterogeneity within each cell type will be ignored to a large extent; there will be no underestimation of the functional role of genes with relatively low expression levels but important for revealing cell fate due to over-focus on high-expression genes. Related key transcription factors; at the same time, it can effectively reduce the sparsity and technical noise of scRNA-seq data, and show good robustness and ability in identifying feature-related cell types and subpopulations.

Description of drawings

In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following will briefly introduce the accompanying drawings that need to be used in the description of the embodiments. Obviously, the accompanying drawings in the following description are only some embodiments of the present invention. For those skilled in the art, other drawings can also be obtained based on these drawings without creative work.

Fig. 1 is a schematic flowchart of the analysis of the processing method of respiratory system disease data provided by the embodiment of the present invention;

Fig. 2 is a schematic diagram of a processing device for respiratory disease data provided by an embodiment of the present invention;

Fig. 3 is a schematic flowchart of a processing system for respiratory system disease data provided by an embodiment of the present invention;

Fig. 4 is an overview of gPAS obtained by the single-cell pathway-based scoring method provided by the embodiment of the present invention, and output of TRS, trait-related genes, trait-related cells, trait-related cell types/subgroups, and trait-type tube pathways using gPAS.

Detailed ways

In order to enable those skilled in the art to better understand the solutions of the present invention, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention.

In some processes described in the specification and claims of the present invention and the descriptions in the above-mentioned drawings, multiple operations appearing in a specific order are included, but it should be clearly understood that these operations may not be performed in the order in which they appear herein or executed in parallel. The serial numbers of operations such as 101, 102, etc. are only used to distinguish different operations, and the serial numbers themselves do not represent any execution order. Additionally, these processes can include more or fewer operations, and these operations can be performed sequentially or in parallel. It should be noted that the descriptions of "first" and "second" in this article are used to distinguish different messages, devices, modules, etc., and do not represent a sequence, nor do they limit that "first" and "second" are different types.

The following will clearly and completely describe the technical solutions in the embodiments of the present invention with reference to the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only some, not all, embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without making creative efforts belong to the protection scope of the present invention.

Figure 1 is a schematic flowchart of a method for processing respiratory disease data provided by an embodiment of the present invention. Specifically, the method includes the following steps:

101: Obtain the single-cell sequencing sequence data to be analyzed;

In one embodiment, the single-cell sequencing data includes seven independent single-cell RNA-seq (scRNA-seq) or single-nucleus RNA-seq (snRNA-seq) datasets covering 1.39 million cells from humans (homo sapiens) and mice (mus musculus). For blood cells, two scRNA-seq datasets based on human BMMC (N = 35,582 cells) and human PBMC (N = 97,039 cells) were collected to reveal trait-associated cell subpopulations or types. For immune/metabolic related diseases/traits, a scRNA-seq dataset from human cells (HCL, N = 513,707 cells in 35 adult tissues) was utilized to construct a pseudo-bulk expression profile and priority risk tissues associated with the disease/trait for each tissue.

In one example, to discover immune cell populations associated with severe respiratory disease, a large-scale PBMC scRNA-seq dataset (N = 469,453 cells) was collected, containing 254 peripheral blood samples with different severity of respiratory disease (mild N = 109 samples, moderate N = 102 samples, severe N = 50) and 16 healthy controls. Optionally, the single-cell sequencing sequence data to be analyzed includes single-cell sequencing sequence data of a healthy control group and respiratory diseases of different degrees of severity.

102: Use machine learning to process the single-cell sequencing sequence data to be analyzed, and obtain the PAS scoring matrix of the cell pathway and the PAS of the cell pathway;

In one embodiment, the single-cell sequencing sequence data to be analyzed is processed by using a machine learning method to obtain the PAS score matrix of the cell pathway and the acquisition steps of the PAS of the cell pathway include:

Access to pathway data for respiratory diseases;

The gene-cell matrix in the single-cell sequencing sequence data is standardized to obtain the normalized gene-cell matrix; specifically, the sparse gene-cell matrix in the scRNA-seq data is standardized using the variance stabilization transformation parameter with a scaling factor of 10,000 to obtain the standardized expression of a single gene in a single cell; the normalized formula is: Among them, a _{g, j} represents the original expression of gene g in cell j, e _{g, j} represents the normalized expression of gene g in cell j;

Based on the pathway data of respiratory diseases, the normalized gene-cell matrix is transformed into a pathway-cell matrix by machine learning method, and the pathway-cell matrix is used to obtain the cell-pathway PAS score matrix, which includes the pathway activity score PAS of a single cell in a single pathway;

In one embodiment, the pathway data is KEGG pathway data, the pathway from the KEGG database is used as the default gene set for evaluating PAS, and the standardized gene-cell matrix is converted into a pathway-cell matrix by using the method of singular value decomposition SVD; P _i is used to represent the gene set in pathway i, and for each pathway i, matrix A _i is selected from the standardized gene-cell matrix A, wherein the columns of matrix A _i are all N cells, and the rows are the |P _i |genes in the pathway gene set P _i , obtained according to SVD where U represents an N×N orthogonal matrix, Σ represents a diagonal matrix with all zeros except the main diagonal elements, V ^T represents a |P _i |×|P _i | orthogonal matrix; for a right orthogonal matrix /> The t-th column vector v _t represents the t-th principal component, which reflects the collaborative expression variability of genes in the pathway in single-cell data; since the first principal component PC1 represents the largest variance variation, the projection of cell j characteristics on PC1 represents the PASs _i, j of pathway i; for cell j, the original PASs _{i ,j} is adjusted using all expression variances in pathway i;

In one embodiment, the pathway activity score PAS is optimized to obtain a weighted PAS;

Ways to obtain weighted PAS include:

in, Indicates weighted PAS, /> Indicates the standardized expression of gene g in cell i after optimization, and si _,j indicates the pathway activity score PAS of pathway i in cell j;

In one embodiment, Ways to obtain include:

in, Indicates the normalized expression of gene g in cell i, MAX(e _g,j ) indicates the maximum value of gene expression in pathway i, and MIN(eg _,j ) indicates the minimum value of gene expression in pathway i.

Optionally, the machine learning method includes a singular value decomposition (SVD) method; the singular value decomposition (SVD) method greatly improves the computational efficiency of analyzing sparse matrices, and can obtain eigenvalues without calculating the variance matrix; the standardized gene-cell matrix is replaced by a low-dimensional pathway-cell matrix using the singular value decomposition method.

103: Obtain genetic association data of respiratory diseases, process the genetic association data, and obtain pathway data with SNPs annotations;

In one embodiment, the genetic association data for respiratory diseases includes genetic association data for severe respiratory diseases;

In one embodiment, the step of processing the genetic association data to obtain pathway data with SNPs annotations includes:

Screen the SNPs of a single gene from the genetic association data, map the SNPs of a single gene to the corresponding pathway based on the pathway data of respiratory diseases, and obtain the pathway data with SNPs annotation;

Optionally, the step of obtaining the SNPs of a single gene includes: after obtaining the SNPs of the genes in the genetic association data, assigning the SNPs gene pairs respectively to obtain the assignment results;

Repeated genes corresponding to multiple genes in the assignment results were treated as independent SNP gene associations; SNPs with a minor allele frequency (MAF) greater than 0.1 in the assignment results were retained; SNPs on the sex chromosome were deleted; SNPs of a single gene were obtained;

After summarizing the SNPs of a single gene, it is the SNPs of all genes. Specifically, the genetic association data is GWAS data, and 20kb is used as the default parameter to assign SNPs in the GWAS summary statistics data to related genes; the symbol g(k) is used to represent the gene g with SNP k, and through the assignment of SNP gene pairs, there are several single SNPs corresponding to multiple genes; because the whole process needs to infer parameters from thousands of SNPs, but the above-mentioned SNPs corresponding to multiple genes have no effect on the inference process, so the above-mentioned repeated genes need to be treated as independent SNP gene associations; Keep the SNPs with a minor allele frequency (MAF) greater than 0.1, delete the SNPs on the sex chromosomes, and finally get the SNPs of related genes;

Based on the pathways in the KEGG database, the genes with associated SNPs are annotated into the pathways, and Si = Equation 2 is used to represent the SNPs set in pathway i; the linkage disequilibrium LD (linkage disequilibrium) is calculated for the SNPs extracted from the GWAS summary data using the Phase 3 data of the Thousand Genomes Project; this program provides functional gene sets such as GO, Reactome, and MSigDB as replacement options. Additionally, the major histocompatibility complex region Chr6: 25-35Mbp where extensive LD is present was deleted.

In one embodiment, GWAS data are phenotyped, and phenotype annotations for a given phenotype include dichotomies, continuous dependency features, or endophenotype and center measures.

104: Statistically analyze and process the PAS of the cell pathway and the pathway data with SNP annotations to obtain the estimated coefficient;

In one embodiment, the PAS of the cell pathway and the pathway data with SNP annotations are statistically analyzed, and the step of obtaining the estimated coefficient includes:

Get the genetic effect value of all SNPs in a single pathway data based on the pathway data with SNPs annotation;

Based on the multigene regression model of the genetic association data of respiratory diseases, based on the PAS and the genetic effect value, the parameter estimation of the distribution of the genetic effect value is carried out to obtain the estimated coefficient;

Optionally, the formula for obtaining the genetic effect value is: Among them, β represents the theoretical effect size vector of m SNPs, ε represents the random environmental error, R represents the LD matrix, and X ^T represents the standard genotype of the SNPs in the genetic association data sample;

Optionally, the methods for obtaining the estimated coefficients include:

Among them, τi _,j represents the estimated coefficient of pathway i in cell j, and the estimated coefficient reflects the influence of cell-specific PAS on the variance of the GWAS effect size, that is, the influence of genetics on the response; _τ0 represents the intercept term, and ^σ2 represents the variance of the SNP effect size in the pathway, Indicates weighted PAS.

In one embodiment, S _i represents the SNP set containing all SNPs in the mapping gene of each pathway i, and the polygenic model assumes that the effect size of all SNPs in the prior pathway i follows a multivariate normal distribution, where σ ² represents the variance of the SNPs effect size in the pathway, and I represents |S _i |×|S _i |identity matrix;

In one embodiment, based on previous assumptions, the genetic effect value The distribution of is estimated using the following formula: /> Use this formula to optimize the estimated coefficients;

In one embodiment, in order to optimize the estimated coefficient of each pathway in the polygenic regression model, the method of moments (method-of-moments approach) that can significantly improve computational efficiency and estimate consistent convergence is used to optimize the polygenic regression model; then, the observed and expected square effects of the SNPs associated with each pathway are fitted, and the expected value is estimated by the following formula: where Tr represents the matrix trace.

105: Multiply the estimated coefficient by PAS and then sum to obtain the genetic-related pathway activity score gPAS of the cell;

In one embodiment, the genetic-related pathway activity score gPAS (gPj) is gPAS related to respiratory diseases, and the acquisition formula is:

in, is the estimated coefficient after optimization.

In one embodiment, the method further includes: performing correlation analysis and sorting on the genetic related pathway activity score gPAS and the gene expression of each cell, and screening out N trait-related genes;

Optional, the method of analyzing and sorting the genetic expression of the genetic -related channel activity score and the gene expression of each cell includes: determine the expression of the single gene and the correlation between the GPAS through the Pilson correlation coefficient (PCC), and sort the genes according to the correlation to obtain N personalized genes related to respiratory diseases. Specifically, in order to maximize the efficacy, each gene G. The expression is weighted in reverse by its genetic noise level at a specific technical noise level. The level of noise is estimated by the average variance relationship between the foundation of the foundation in the SCRNA-SEQ data;

Optionally, the N trait-related genes are the first 1000 or last 1000 trait-related genes sorted according to the descending or ascending order of correlation; N is not limited to 1000, and N is a natural integer. N trait-related genes include one or more of the following: CALM3, PIK3R1, IL32, CD3E, B2M, PRS29, and GZMB;

In one embodiment, the method further includes: calculating the trait-related score TRS of each cell according to N trait-related genes; performing clustering according to the trait-related score TRS and the horizontal P value of a single cell to obtain trait-related cells related to respiratory diseases with different levels of severity; trait-related genes are significantly enriched in the following trait-related cells, including: hay bone marrow immature T16 cells (hay bone marrow T16 cells), lung naive CD8+ T cells (lung/> CD8+Tcells), liver NKT cells (liver NKT cells) and brain naive T cells (brain/> T cells); the formula for obtaining the trait-related score TRS is: TRS=average RE(GS)-average RE(CG); wherein, average RE(GS) is the average relative expression value of N trait-related gene sets in a given cell, and average RE(CG) is the average relative expression value of the same number of control gene sets randomly selected from the existing gene bank; RE is relative expression; GS is gene set; CG is control gene set;

Optionally, different grades of severity of respiratory diseases include mild, moderate and severe.

Optionally, use the cell scoring method of the AddModuleScore function in Seurat to calculate the trait correlation score TRS of N genes.

In one embodiment, the method further includes: obtaining cell types or subgroups related to traits related to respiratory diseases based on a block bootstrap method, and determining whether the cell types to which individual cells belong are related. Trait-associated cell types or subpopulations (associated with severe respiratory disease) include one or more of the following: CD8+ T cells, megakaryocytes, CD16+ monocytes; /> Highly expressed genes in CD8+ T cells include: memory effector marker genes (GZMK, AQP3, GZMA, PRF1, and GNLY) and exhaustive effect marker genes (LAG3, TIGIT, GZMA, GZMB, PRDM1, and IFNG); specifically, a group of cells is regarded as pseudo-bulk Transcriptome profiling, and average gene expression across cells within a given cell type; for associated cell types, block bootstrap method is used to estimate standard errors and calculate t-statistics for each cell type's corresponding P value. Given that the goal of block bootstrapping is to preserve the data structure when sampling from an empirical distribution, the pathways of the KEGG database are used to partition the genome into multiple biologically meaningful blocks, and the aforementioned pathway-based blocks are sampled with replacement. Under the default parameters, 200 iterations are performed for each cell type association analysis, and the default parameters can be modified for specific execution.

In one embodiment, the method further includes: sorting the genetically related pathway activity score gPAS, and obtaining the trait-related pathways related to respiratory diseases according to the sorting results (selecting the top-ranked pathways in the sorting results) and the P value of the pathways at the cell type level; the trait-related pathways include ribosomes, T cell receptor signaling pathways, primary immunodeficiency, natural killer cell-mediated cytotoxicity, and platelet activation.

Specifically, the gPAS is sorted based on the central limit theorem; the cell type t is represented by the symbol C _t , and the pathway percentage rank of each cell j within C _t is calculated using the following formula: where, /> represents the gPAS level of pathway i in cell j, and M represents the total number of pathways; similarly, the statistical importance value T _i ^t of each pathway i in cell type t is calculated using the following formula: /> where, /> Suppose: H ₀ : T _i ^t = 0 vs H ₁ : T _i ^t >0; the P value for each pathway i in cell type t is: />

In one embodiment, the statistical significance of individual cells is determined by calculating the rank distribution of trait-related genes to further assess whether the cell is significantly related to the trait of interest; specifically, the percentage rank of the trait-related genes in the cell is obtained, Among them, r _g,j represent the expression level of gene g in cell j, and G represents the number of genes related to the specified trait; the percentage level of genes follows the normal distribution U(0,1), under the null assumption that there is no correlation between the percentage levels of genes, the statistical value T _j of each cell is obtained, and the formula is as follows: />

Based on the large number of cells in the single-cell data, the distribution of _Tj was derived using the central limit theorem: Wherein N is the total number of cells; the hypothesis of significance test is: H ₀ : T _j = 0 vs H ₁ : T _j >0; the P value of each cell j is: p _j = Pr(T _j ≤ t).

106: Output the genetic related pathway activity score gPAS;

An application that detects new Application of products of CD8+ T cell subsets in the preparation of products for diagnosing respiratory diseases; new /> CD8+ T cell subsets are a newly discovered function associated with respiratory disease.

Fig. 2 is a schematic flowchart of a device for processing respiratory disease data provided by an embodiment of the present invention. The device includes: a memory and a processor; the memory is used to store program instructions; the processor is used to call the program instructions, and when the program instructions are executed, it is used to execute the above-mentioned method for processing respiratory system disease data.

Fig. 3 is a schematic flowchart of a processing system for respiratory disease data provided by an embodiment of the present invention, including:

An acquisition unit 301, configured to acquire single-cell sequencing sequence data to be analyzed;

The first processing unit 302 is configured to use a machine learning method to process the single-cell sequencing sequence data to be analyzed to obtain a PAS scoring matrix of cell pathways and a PAS of cell pathways;

The second processing unit 303 is used to obtain genetic association data of respiratory diseases, process the genetic association data, and obtain pathway data with SNPs annotations;

The third processing unit 304 is configured to perform statistical analysis on the PAS of the cell pathway and the pathway data with SNP annotations to obtain estimated coefficients;

The fourth processing unit 305 is used to multiply the estimated coefficient by PAS and then sum to obtain the genetic-related pathway activity score gPAS of the cell;

The output unit 306 is configured to output the genetic related pathway activity score gPAS.

A computer-readable storage medium, on which a computer program is stored, and when the computer program is executed by a processor, the above-mentioned method for processing respiratory system disease data is realized.

Fig. 4 is an overview of gPAS obtained by the single-cell pathway-based scoring method provided by the embodiment of the present invention, and output of TRS, trait-related genes, trait-related cells, trait-related cell types/subgroups, and trait-type tube pathways using gPAS;

Among them, A indicates that the gene-cell matrix is converted into a pathway-cell matrix by using the method of singular value decomposition, PC1 indicates the PAS of each pathway; B indicates that the SNP in the GWAS data is annotated into the corresponding pathway; C indicates the multigene regression model; among them, the figure at the top indicates that the estimated coefficient in each pathway is inferred by the multigene regression model, and then calculated using the estimated coefficient and the corresponding PAS to obtain gPAS, and the figure at the bottom indicates the Pearson correlation model, which is used to correlate the gPAS of each cell with the genes of all individual cells , in order to rank the trait-related genes; use the AddModuleScore function in Seurat to get the top N trait-related genes (the default top 1,000). To calculate the trait-related score TRS of each cell; D represents the output, including four outputs: trait-related cells, trait-related cell types, trait-related pathways and trait-related genes.

The validation results of this validation example show that assigning intrinsic weights to indications can moderately improve the performance of the method relative to the default setting.

Those skilled in the art can clearly understand that for the convenience and brevity of the description, the specific working process of the above-described system, device and unit can refer to the corresponding process in the foregoing method embodiment, which will not be repeated here.

In the several embodiments provided in this application, it should be understood that the disclosed system, device and method can be implemented in other ways. For example, the device embodiments described above are only illustrative. For example, the division of the units is only a logical function division. In actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not implemented. In another point, the mutual coupling or direct coupling or communication connection shown or discussed may be through some interfaces, and the indirect coupling or communication connection of devices or units may be in electrical, mechanical or other forms.

The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place, or may be distributed to multiple network units. Part or all of the units can be selected according to actual needs to achieve the purpose of the solution of this embodiment.

In addition, each functional unit in each embodiment of the present invention may be integrated into one processing unit, each unit may exist separately physically, or two or more units may be integrated into one unit. The above-mentioned integrated units can be implemented in the form of hardware or in the form of software functional units.

Those of ordinary skill in the art can understand that all or part of the steps in the various methods of the above-mentioned embodiments can be completed by instructing related hardware through a program, and the program can be stored in a computer-readable storage medium, and the storage medium can include: Read Only Memory (ROM, Read Only Memory), Random Access Memory (RAM, Random Access Memory), magnetic disk or optical disk, etc.

Those of ordinary skill in the art can understand that all or part of the steps in the method of the above-mentioned embodiments can be completed through a program to instruct relevant hardware, and the program can be stored in a computer-readable storage medium, and the above-mentioned storage medium can be a read-only memory, a magnetic disk or an optical disk, etc.

A computer device provided by the present invention has been introduced in detail above. For those of ordinary skill in the art, according to the ideas of the embodiments of the present invention, there will be changes in the specific implementation and application scope. In summary, the contents of this specification should not be understood as limitations on the present invention.

Claims (10)

1. A method of processing respiratory disease data, comprising:

acquiring single cell sequencing sequence data to be analyzed;

processing the single-cell sequencing sequence data to be analyzed by adopting a machine learning method to obtain PAS scoring matrix of a cell pathway and PAS of the cell pathway;

acquiring genetic association data of respiratory diseases, and processing the genetic association data to obtain path data with SNPs annotation;

performing statistical analysis processing on PAS of the cell pathway and the pathway data with SNP annotation to obtain an estimation coefficient;

multiplying the estimation coefficient by PAS and then summing to obtain a genetic related pathway activity score gPAS of the cell;

outputting the genetically related pathway activity score gPAS.

2. The method of claim 1, wherein the step of statistically analyzing the PAS of the cellular pathway and the pathway data with SNP annotations to obtain estimated coefficients comprises:

obtaining genetic effect values of all SNPs in single path data based on the path data with the SNPs annotation;

based on a polygene regression model of genetic association data of respiratory diseases, carrying out parameter estimation on the distribution of the genetic effect values based on the PAS and the genetic effect values to obtain estimation coefficients;

optionally, the obtaining formula of the genetic effect value is:wherein beta represents mThe magnitude vector of the theoretical effect of SNPs, epsilon, represents the random environmental error, R represents the LD matrix, X ^T Representing a standard genotype of SNPs in the genetically related data sample;

optionally, the obtaining manner of the estimation coefficient includes:

wherein τ _i,j Estimated coefficient, τ, representing pathway i in cell j ₀ Representing intercept term, σ ² Shows the variance of the magnitude of SNP effect in the pathway,representing a weighted PAS.

3. The method of claim 1, wherein the genetically related pathway activity score gPAS (gP _j ) The acquisition formula of (1) is:

wherein the gP _j As gPAS, saidAnd the estimated coefficient is optimized.

4. The method of processing respiratory disease data according to claim 1, wherein the method further comprises: performing correlation analysis and sequencing on the genetic related pathway activity score gPAS and the gene expression quantity of each cell, and screening N personality-related genes;

optionally, the method for performing correlation analysis and sequencing on the genetic correlation pathway activity score gPAS and the gene expression quantity of each cell comprises the following steps: determining the correlation between the expression of a single gene and the gPAS through a Pearson Correlation Coefficient (PCC), and sequencing the genes according to the correlation to obtain the N personality-related genes;

optionally, the N trait related genes are the first 1000 or the last 1000 trait related genes sequenced according to a descending or ascending rule of the relativity;

optionally, the N personality-related genes include one or more of the following: CALM3, PIK3R1, IL32, CD3E, B2M, PRS29, and GZMB.

5. The method of processing respiratory disease data according to claim 1, wherein the method further comprises: calculating a trait related score TRS for each cell according to the N trait related genes; clustering according to the trait related score TRS and the level P value of the single cells to obtain trait related cells related to respiratory diseases with different levels of severity;

optionally, calculating the trait related score TRS of the N personality genes using a cell scoring method;

optionally, the different levels of severity of the respiratory disease include mild, moderate and severe.

6. The method of processing respiratory disease data according to claim 1, wherein the method further comprises: obtaining a trait-related cell type or subpopulation based on the block boot method block bootstrap method;

optionally, the method further comprises: and sequencing the genetic related pathway activity scores gPAS, and obtaining the trait related pathway according to the sequencing result and the P value of the pathway on the cell type level.

7. Detecting newUse of a product of a cell subpopulation for the preparation of a product for diagnosing a respiratory disease.

8. A device for processing respiratory disease data, the device comprising: a memory and a processor;

the memory is used for storing program instructions; the processor is adapted to invoke program instructions for performing the method of processing respiratory disease data according to any of claims 1-6 when the program instructions are executed.

9. A system for processing respiratory disease data, comprising:

an acquisition unit for acquiring single-cell sequencing sequence data to be analyzed;

the first processing unit is used for processing the single-cell sequencing sequence data to be analyzed by adopting a machine learning method to obtain a PAS scoring matrix of a cell pathway and PAS of the cell pathway;

the second processing unit is used for acquiring genetic association data of the respiratory tract diseases, and processing the genetic association data to obtain path data with SNPs annotation;

a third processing unit for performing statistical analysis processing on PAS of the cell pathway and the pathway data with SNP annotation to obtain an estimation coefficient;

a fourth processing unit for multiplying the estimation coefficient by PAS and then summing to obtain a genetic related pathway activity score gPAS of the cell;

and an output unit for outputting the genetically related pathway activity score gPAS.

10. A computer readable storage medium having stored thereon a computer program which, when executed by a processor, implements a method of processing respiratory disease data according to any of the preceding claims 1-6.