Single-cell-based database visualization method

Abstract

The present invention relates to the technical field of single-cell data processing, and in particular to a single-cell-based database visualization method. The method comprises the following steps: obtaining a single-cell dataset; performing standardization processing on the single-cell dataset to obtain standardized single-cell data; performing high-dimensional feature extraction on the standardized single-cell data to obtain single-cell high-dimensional feature data; performing nonlinear dimensionality reduction processing on the single-cell high-dimensional feature data to obtain single-cell reduced-dimensional data; performing data clustering on the single-cell reduced-dimensional data to obtain single-cell cluster data; performing gene differential expression extraction on the single-cell cluster data to obtain differentially expressed gene data; the present invention uses data processing technology, feature mapping technology, and visualization technology to perform multimodal data fusion analysis on the single-cell data; and dynamically weighting the single-cell database to enhance the visualization effect of the single-cell database.

Description

Technical Field

[0001] The present invention relates to the technical field of single-cell data processing, and in particular to a database visualization method based on single cells. Background Art

[0002] Single-cell technology has evolved from the basic transcriptome sequencing stage to the multi-omics and spatial transcriptome expansion stage, resulting in an explosive growth in the amount of single-cell data. Various types of databases have emerged, covering different species, tissues, and cell types. Due to the huge amount of data carried by single-cell databases, conventional database display methods cannot intuitively reflect the specific situation of single-cell data. Therefore, researchers use visualization tools to visualize the single-cell database. As the complexity of single-cell data increases, researchers begin to perform dimensionality reduction and clustering operations on single-cell data to achieve visual display effects of single-cell databases. However, when faced with complex single-cell data, existing methods fail to perform multimodal data fusion analysis on single-cell data; and lack dynamic weight division of single-cell databases, resulting in poor visualization of single-cell databases. Summary of the Invention

[0003] Based on this, it is necessary to provide a single-cell-based database visualization method to solve at least one of the above technical problems.

[0004] To achieve the above-mentioned purpose, a single-cell database visualization method is provided, which comprises the following steps:

[0005] Step S1: Acquire a single-cell data set; perform standardization processing on the single-cell data set to obtain standardized single-cell data; perform high-dimensional feature extraction on the standardized single-cell data to obtain single-cell high-dimensional feature data; perform nonlinear dimensionality reduction processing on the single-cell high-dimensional feature data to obtain single-cell reduced dimensionality data;

[0006] Step S2: performing data clustering on the single-cell dimensionality reduction data to obtain single-cell cluster data; performing gene differential expression extraction on the single-cell cluster data to obtain differentially expressed gene data; performing single-cell function annotation on the differentially expressed gene data to obtain single-cell function annotation data; performing single-cell type feature identification on the single-cell cluster data based on the single-cell function annotation data to obtain single-cell type feature data; performing single-cell type classification on the single-cell cluster data based on the single-cell type feature data to obtain single-cell type data; performing multimodal data fusion on the single-cell cluster data using the single-cell type data to obtain multimodal fusion data;

[0007] Step S3: performing single-cell modality relationship analysis on the multimodal fusion data to generate single-cell relationship data; determining the database table structure of the single-cell relationship information based on the differentially expressed gene data to generate data table structure features; establishing a single-cell database index structure based on the multimodal fusion data and the data table structure information to generate database index parameters; constructing a single-cell database architecture based on the single-cell relationship data, the data table structure features, and the database index parameters to obtain a single-cell database architecture;

[0008] Step S4: Based on the single-cell database architecture, single-cell expression pattern characteristics are identified for the multimodal fusion data to obtain single-cell expression pattern characteristics; database display information is extracted for the single-cell database architecture according to the single-cell expression pattern characteristics to obtain database display data; dynamic weight marking is performed on the database display data to generate dynamic weight marking data; color marking is performed on the multimodal fusion data based on the dynamic weight marking data to obtain color marking feature data; data color depth is determined for the color marking feature data to generate data color depth feature information; the single-cell database architecture is visualized through the data color depth feature information to obtain a single-cell data visualization report.

[0009] The present invention provides basic data for subsequent analysis by acquiring single-cell data sets; standardizes the single-cell data sets to ensure data consistency and comparability; performs high-dimensional feature extraction on the standardized single-cell data to capture key information in the single-cell data; performs nonlinear dimensionality reduction on the single-cell high-dimensional feature data to effectively reduce the dimensionality of the data while retaining the most important features; performs data clustering on the single-cell dimensionality reduction data to identify single-cell populations with similar characteristics; performs gene differential expression extraction on the single-cell cluster data to identify genes with significantly different expression levels in different single-cell populations; performs single-cell functional annotation on the differentially expressed gene data; performs single-cell type feature identification on the single-cell cluster data based on the single-cell function annotation data to obtain single-cell type feature data; performs single-cell type classification on the single-cell cluster data according to the single-cell type feature data to more accurately classify single-cell types; performs multimodal data fusion on the single-cell cluster data through the single-cell type data to integrate different types of data, providing a more comprehensive perspective for subsequent analysis. Single-cell modal relationship analysis of multimodal fusion data can reveal the intrinsic connections between different data types; the database table structure of single-cell relationship information is determined based on differentially expressed gene data, providing a structured framework for subsequent data storage and management; the single-cell database index structure is established for multimodal fusion data and data table structure information to improve the efficiency of data retrieval; the single-cell database architecture is constructed based on single-cell relationship data, data table structure characteristics and database index parameters to obtain the single-cell database architecture; single-cell expression pattern feature recognition of multimodal fusion data based on the single-cell database architecture can clarify the expression behavior of single cells; database display information of the single-cell database architecture is extracted based on single-cell expression pattern features to obtain database display data; database display data is dynamically weighted to generate dynamic weighted labeled data; multimodal fusion data is color-coded based on dynamic weighted labeled data to further enhance data visualization and make key features and patterns more prominent; data color depth is determined for color-coded feature data to generate data color depth feature information; the single-cell database architecture is visualized using data color depth feature information to obtain a single-cell data visualization report. Therefore, the present invention performs multimodal data fusion analysis on single-cell data through data processing technology, feature mapping technology and visualization technology; and dynamically weights the single-cell database, thereby enhancing the visualization effect of the single-cell database.

[0010] Preferably, step S2 includes the following steps:

[0011] Step S21: performing structural feature recognition on the single-cell dimensionality reduction data to obtain single-cell structural data; determining hierarchical structure parameters on the single-cell structural data to generate single-cell hierarchical structure parameters; calculating single-cell density on the single-cell dimensionality reduction data based on the single-cell hierarchical structure parameters to obtain single-cell density data; and clustering the single-cell dimensionality reduction data based on the single-cell density data to obtain single-cell clustering data.

[0012] Step S22: determining gene expression characteristics of the single-cell clustering data to obtain gene expression characteristic data; extracting differential expression characteristics of the gene expression characteristic data to obtain differentially expressed gene data; performing differential feature identification on the differentially expressed gene data to generate differential feature identification data; and annotating the single-cell function of the single-cell clustering data based on the differential feature identification data and the gene expression characteristic data to obtain single-cell function annotation data.

[0013] Step S23: performing single-cell function mapping on the single-cell clustering data according to the single-cell function annotation data to generate single-cell function mapping data; performing type mapping feature determination on the single-cell function mapping data to obtain type mapping feature data; performing single-cell type feature extraction on the single-cell clustering data using the type mapping feature data to obtain single-cell type feature data;

[0014] Step S24: performing single-cell type classification on the single-cell clustering data according to the single-cell type feature data to obtain single-cell type data; performing multimodal data fusion on the single-cell clustering data through the single-cell type data to obtain multimodal fusion data.

[0015] The present invention identifies the structural features of single-cell dimensionality reduction data to reveal the intrinsic structure of single-cell data; determines the hierarchical structure parameters of single-cell structural data to provide an organizational hierarchy for single-cell structural data; calculates the single-cell density of single-cell dimensionality reduction data based on the single-cell hierarchical structure parameters to provide a quantitative index for evaluating the distribution density of single-cell populations; clusters the single-cell dimensionality reduction data based on the single-cell density data to help group single-cell populations with similar features; determines the gene expression features of single-cell clustering data to clarify the common pattern of gene expression in each cluster; extracts the differential expression features of gene expression feature data to refine the gene expression feature data; and performs differential expression feature analysis on differentially expressed gene data. Perform differential feature identification to help discover genes with significantly different expression levels in different single-cell populations; perform single-cell function annotation on single-cell clustering data based on differential feature identification data and gene expression feature data to obtain single-cell function annotation data; perform single-cell function mapping on single-cell clustering data based on single-cell function annotation data to determine single-cell function information; perform type mapping feature determination on single-cell function mapping data to identify and classify different single-cell types; perform single-cell type division on single-cell clustering data based on single-cell type feature data to clearly distinguish different single-cell types; perform multimodal data fusion on single-cell clustering data through single-cell type data to obtain multimodal fusion data.

[0016] Preferably, step S24 includes the following steps:

[0017] Step S241: extracting feature contours from the single-cell cluster data according to the single-cell type feature data to obtain single-cell feature contour data; determining contour boundary ranges from the single-cell feature contour data to obtain contour boundary range data;

[0018] Step S242: performing single-cell boundary division on the single-cell clustering data using the contour boundary range data to obtain single-cell boundary data; performing inter-cell relationship feature recognition on the single-cell type feature data based on the single-cell boundary data to obtain inter-cell relationship feature data; determining the degree of cell differentiation on the inter-cell relationship feature data to obtain cell differentiation degree data; and performing single-cell type division on the single-cell clustering data based on the cell differentiation degree data to obtain single-cell type data;

[0019] Step S243: extracting feature dimensions from the single cell type data to obtain type feature dimension data; dimensionally encoding the single cell type data based on the type feature dimension data to generate type dimension encoded data; and feature aligning the type dimension encoded data with the single cell clustering data to obtain single cell feature aligned data;

[0020] Step S244: performing information complementarity integration on the single-cell feature alignment data to generate information complementarity integration parameters; performing multimodal heterogeneity elimination on the single-cell feature alignment data according to the information complementarity integration parameters to obtain multimodal heterogeneity elimination data; performing fusion and collaborative calibration on the single-cell feature alignment data based on the multimodal heterogeneity elimination data to obtain multimodal fusion data.

[0021] The present invention extracts feature contours from single-cell clustering data through single-cell type feature data, which helps to identify significant features in single-cell data; determines the contour boundary range of single-cell feature contour data, and provides a clear boundary for single-cell feature contour data; divides single-cell boundaries of single-cell clustering data through contour boundary range data, divides single-cell populations into more specific sub-populations, and provides a clear definition for further analysis of each sub-population; identifies intercellular relationship features of single-cell type feature data based on single-cell boundary data, and reveals the interaction between single cells; determines the degree of cell differentiation of intercellular relationship feature data, and clarifies the degree of cell differentiation; divides single-cell clustering data into single-cell types according to cell differentiation degree data, and obtains to single-cell type data; feature dimension extraction is performed on single-cell type data to obtain type feature dimension data; dimension encoding is performed on single-cell type data based on type feature dimension data to provide standardized encoding for comparison of single-cell type data; feature alignment is performed on type dimension encoded data and single-cell clustering data to ensure consistency and comparability between different data sets; information complementarity is integrated on single-cell feature alignment data to improve data integrity and analysis accuracy; multimodal heterogeneity elimination is performed on single-cell feature alignment data according to information complementarity integration parameters to reduce differences between different data sources and improve data consistency; single-cell feature alignment data is fused and collaboratively calibrated based on multimodal heterogeneity elimination data to obtain multimodal fusion data.

[0022] Preferably, step S3 includes the following steps:

[0023] Step S31: determining the modal dimension structure of the multimodal fusion data to obtain modal dimension structure data; extracting dimensional correlation from the modal dimension structure data to generate dimensional correlation information; determining the single-cell modal relationship of the multimodal fusion data based on the dimensional correlation information to generate single-cell relationship data;

[0024] Step S32: performing differentiation feature mapping on the multimodal fusion data using the cell differentiation degree data to generate differentiation feature matching data; performing relationship constraint determination on the single-cell relationship data based on the differentiation feature matching data to obtain single-cell relationship constraint data; performing primary and foreign key setting on the differentiation feature matching data based on the single-cell relationship constraint data to obtain single-cell primary and foreign key data; determining the database table structure of the single-cell relationship data using the single-cell relationship constraint data and the single-cell primary and foreign key data to generate data table structure features;

[0025] Step S33: Establish a single-cell database index based on the data table structure characteristics and generate database index parameters; construct a single-cell database architecture based on the single-cell relationship data, data table structure characteristics and database index parameters to obtain the single-cell database architecture.

[0026] The present invention determines the modal dimension structure of multimodal fusion data to clarify the structural features in different data modalities; extracts dimensional correlation from the modal dimension structure data to reveal the connections and interactions between different dimensions; determines the single-cell modal relationship of the multimodal fusion data based on the dimensional correlation information to generate single-cell relationship data; matches the differentiation characteristics of the multimodal fusion data with the cell differentiation degree data, which helps to match the cell differentiation information with the multimodal data, thereby better understanding the characteristic changes during the cell differentiation process; determines the relationship constraints of the single-cell relationship data based on the differentiation characteristic matching data, which helps to introduce the constraint conditions of the differentiation characteristics into the single-cell relationship data; sets the primary and foreign keys of the differentiation characteristic matching data based on the single-cell relationship constraint data to obtain the single-cell primary and foreign key data; determines the database table structure of the single-cell relationship data based on the single-cell relationship constraint data and the single-cell primary and foreign key data, providing a detailed table structure design for constructing a single-cell database; establishes a single-cell database index based on the data table structure characteristics to improve the query efficiency and performance of the database; and constructs the single-cell database architecture based on the single-cell relationship data, the data table structure characteristics and the database index parameters to obtain the single-cell database architecture.

[0027] Preferably, step S33 includes the following steps:

[0028] Step S331: performing type index feature recognition on the single cell type data according to the data table structure feature to generate type index feature data; determining the index structure dimension of the type index feature data based on the multimodal fusion data to obtain index structure dimension data;

[0029] Step S332: performing table structure feature recognition on the data table structure feature to obtain table structure feature data; performing structure compatibility analysis on the table structure feature data to obtain table structure compatibility information;

[0030] Step S333: performing index structure type mapping on the index structure dimension data using the table structure compatibility information to obtain index structure mapping information; performing index feature enhancement on the index structure mapping information to generate index feature enhancement data; establishing a single-cell database index based on the data table structure features according to the index feature enhancement data to generate database index parameters;

[0031] Step S334: Assign single-cell indexes to the database index parameters according to the data table structure characteristics to generate single-cell index assignment data; construct a single-cell database architecture based on the single-cell relationship data, the data table structure characteristics and the single-cell index assignment data to obtain the single-cell database architecture.

[0032] The present invention identifies the type index features of single-cell type data through data table structure features, and can mark and distinguish the key features in single-cell type data; determines the index structure dimension of type index feature data based on multimodal fusion data, and establishes a suitable index structure for different types of single-cell data, thereby optimizing the data organization method; identifies the table structure features of data table structure features, and clarifies the structural composition of the database table; performs structural compatibility analysis on table structure feature data, evaluates the compatibility between different data table structures, and ensures smooth and efficient data integration; and performs index structure type mapping on index structure dimension data through table structure compatibility information. , effectively matching the index structure with the table structure, improving the accuracy of database operations; strengthening the index features of the index structure mapping information to enhance the index retrieval efficiency and data access speed; establishing a single-cell database index based on the data table structure features according to the index feature strengthening data, and generating database index parameters; assigning single-cell indexes to the database index parameters according to the data table structure features, ensuring the consistency of the index parameters with the actual data table structure; constructing the single-cell database architecture based on the single-cell relationship data, data table structure features and single-cell index assignment data, providing an orderly and scalable framework for the integration, storage and management of single-cell data.

[0033] Preferably, step S332 includes the following steps:

[0034] Step S3321: extracting gene quantities from the gene expression feature data using the data table structure feature to generate gene expression quantity data; determining single cell populations from the single cell clustering data using the gene expression quantity data to obtain single cell population data;

[0035] Step S3322: performing population diversity feature identification on the single-cell population data based on the single-cell relationship data to obtain population diversity feature information; determining the single-cell population structure on the population diversity feature information to obtain single-cell population structure data; performing single-cell expression sequence identification on the gene expression quantity data using the single-cell population structure data to obtain single-cell expression sequence data; performing sequence structure extraction on the single-cell expression sequence data to generate single-cell sequence structure data;

[0036] Step S3323: integrating the single-cell population structure data and the single-cell sequence structure data to obtain single-cell data table structure features; performing single-cell multi-dimensional structure processing on the data table structure features to generate multi-dimensional structured hierarchical data; performing single-cell competitive tracking on the single-cell relationship data based on the multi-dimensional structured hierarchical data to obtain single-cell competitive data;

[0037] Step S3324: Match the single-cell competitive data with the multidimensional structure hierarchical data to obtain structure matching relationship data; quantify the structure impact of single-cell gene expression on the structure matching relationship data to obtain structure impact quantification data; perform adaptability measurement of data table structure characteristics based on the structure impact quantification data to obtain table structure adaptability data; integrate the structure impact quantification data and the table structure adaptability data to obtain table structure compatibility information.

[0038] The present invention uses the data table structure feature to extract gene quantity from gene expression feature data, which can quantify the expression quantity of genes in each single cell; determines the single cell population of single cell clustering data through gene expression quantity data, and reveals the situation of single cell population; identifies the population diversity characteristics of single cell population data based on single cell relationship data, and clarifies the population diversity characteristic information of single cells; identifies the single cell expression sequence of gene expression quantity data through single cell population structure data, and obtains single cell expression sequence data; extracts the sequence structure of single cell expression sequence data, and generates single cell sequence structure data; integrates the single cell population structure data and the single cell sequence structure data, so that the data at different levels are merged to form a comprehensive single cell data view; The table structure characteristics are used to perform multi-dimensional structural processing on single cells to generate multi-dimensional hierarchical data; single-cell competition is tracked on single-cell relationship data based on the multi-dimensional hierarchical data to identify potential competitive relationships between single cells; single-cell competition data is matched with multi-dimensional hierarchical data to reveal the relationship between single-cell competitiveness and multi-dimensional structure; the structural impact of single-cell gene expression structure data is quantified to quantify the impact of gene expression structure on single-cell characteristics; the adaptability of data table structure characteristics is measured based on the structural impact quantification data to evaluate the adaptability of data table structure characteristics to actual data; the structural impact quantification data and table structure adaptability data are integrated to obtain table structure compatibility information, ensuring the rationality of database design.

[0039] Preferably, step S4 includes the following steps:

[0040] Step S41: inputting the multimodal fusion data into the single-cell database architecture to perform single-cell edge detection to generate single-cell edge data; performing edge feature highlighting on the single-cell edge data to obtain edge feature highlighting data; determining single-cell trajectories on the multimodal fusion data based on the edge feature highlighting data to obtain single-cell trajectory data; performing single-cell state matching on the single-cell trajectory data to generate single-cell state data; performing single-cell expression pattern feature recognition on the multimodal fusion data based on the single-cell state data to obtain single-cell expression pattern features;

[0041] Step S42: extracting database display information from the single-cell database architecture based on the single-cell expression pattern characteristics to obtain database display data; dynamically weight-labeling the database display data to generate dynamic weight-labeled data;

[0042] Step S43: weight mapping the multimodal fusion data using the dynamic weight marking data to generate weight mapping data; color coding the weight mapping data to obtain color coding data; and color marking the multimodal fusion data based on the color coding data to obtain color marking feature data.

[0043] Step S44: determining the data color depth of the color-marked feature data to generate data color depth feature information; visually displaying the single-cell database architecture through the data color depth feature information to obtain a single-cell data visualization report.

[0044] The present invention inputs multimodal fusion data into a single-cell database architecture to perform single-cell edge detection and identify boundary information in single-cell data; determines the single-cell trajectory of the multimodal fusion data based on edge feature highlighting data, and clearly tracks the development trajectory of individual cells; performs single-cell state matching on the single-cell trajectory data to determine the state of the cell at different time points; identifies the single-cell expression pattern characteristics of the multimodal fusion data based on the single-cell state data to reveal the cell expression pattern information; extracts database display information from the single-cell database architecture based on the single-cell expression pattern characteristics, providing a data basis for data display and analysis; dynamically weights the database display data to generate Dynamic weight labeling data; weight mapping of multimodal fusion data through dynamic weight labeling data can map the weight information of the data to the multimodal fusion data; color coding of weight mapping data can represent the weight of the data through color, making the data visualization more intuitive and easy to understand; color labeling of multimodal fusion data based on color coding data enhances the distinguishability of data visualization; data color depth is determined for color-labeled feature data, and the intensity or density of the data is represented by the depth of color, providing more dimensional information for data visualization; the single-cell database architecture is visualized through data color depth feature information to obtain a single-cell data visualization report.

[0045] Preferably, step S42 includes the following steps:

[0046] Step S421: performing cell type annotation on the single-cell expression pattern features to obtain cell type annotation data; integrating the single-cell expression pattern features with the cell type annotation data to generate a single-cell expression profile; and measuring the single-cell expression level on the single-cell expression profile to obtain single-cell expression level data;

[0047] Step S422: Determine a data hierarchical relationship for the single-cell expression pattern features based on the single-cell expression level data to generate data hierarchical relationship information; generate database records for the single-cell database architecture based on the data hierarchical relationship information to obtain single-cell database record information; and format the single-cell database record information for display information to obtain database display data;

[0048] Step S423: performing expression activity feature recognition on the single-cell expression level data to generate expression activity features; performing expression activity assessment on the single-cell expression pattern features based on the expression activity features to obtain single-cell expression activity data; performing expression dynamic range assessment on the single-cell expression activity data to generate expression dynamic range data;

[0049] Step S424: Calculating a dynamic weight factor for the expression dynamic range data to obtain a dynamic weight factor; dynamically assigning weights to the single-cell expression level data based on the dynamic weight factor; if the single-cell expression level data is higher than the preset single-cell expression level data, assigning a high weight to the single-cell expression level data to generate high-weight assigned data; if the single-cell expression level data is less than or equal to the preset single-cell expression level data, assigning a low weight to the single-cell expression level data to generate low-weight assigned data; integrating the high-weight assigned data and the low-weight assigned data to obtain dynamic weight assigned data;

[0050] Step S425: Mark the database display data with display data weights using the dynamic weight allocation data to generate dynamic weight marking data.

[0051] The present invention uses single-cell expression pattern characteristics to perform cell type annotation and obtain cell type annotation data; uses cell type annotation data to perform single-cell expression profile integration on single-cell expression pattern characteristics, so as to summarize the expression information of each cell; performs single-cell expression level determination on single-cell expression profile, quantifies the gene expression level of each cell, and provides accurate data for evaluating the expression activity of cells; determines the data hierarchy relationship of single-cell expression pattern characteristics based on single-cell expression level data, and can identify the hierarchical structure between different expression levels; generates database records for single-cell database architecture according to data hierarchy relationship information, so as to convert complex single-cell data into structured database records for easy storage and query; formats display information of single-cell database record information, converts database records into user-friendly display format, and improves data readability and usability; performs expression activity feature recognition on single-cell expression level data, so as to distinguish cells with high expression activity and low expression activity; performs expression activity assessment on single-cell expression pattern characteristics according to expression activity features, so as to evaluate the expression of each cell. reach an active state; evaluate the dynamic range of expression of single-cell expression activity data to clarify the range of changes in cell expression levels; calculate dynamic weight factors for expression dynamic range data so that weights are assigned according to the range of changes in expression levels, providing a basis for subsequent data weighting; dynamically assign weights to single-cell expression level data based on dynamic weight factors. If the single-cell expression level data is higher than the preset single-cell expression level data, high weight assignment is performed on the single-cell expression level data to generate high-weight assignment data; if the single-cell expression level data is less than or equal to the preset single-cell expression level data, low weight assignment is performed on the single-cell expression level data to generate low-weight assignment data; integrate high-weight assignment data and low-weight assignment data to obtain dynamic weight assignment data. This step ensures that data with higher expression levels obtain higher weights, and vice versa, lower weights, thereby ensuring that the importance of the data matches its weight; display data weights for database display data through dynamic weight assignment data, so that the importance of different data points is reflected in the data display, making key information more prominent.

[0052] Preferably, step S44 includes the following steps:

[0053] Step S441: performing expression activity color mapping on the single-cell expression activity data using the color marking feature data to generate activity color mapping data; performing single-cell expression color depth gradient determination on the activity color mapping data to obtain expression color depth gradient data;

[0054] Step S442: determining the color depth range of the expression color depth gradient data to obtain color depth range data; dividing the expression color depth gradient data into color depth intervals according to the color depth range data to obtain single-cell color depth interval data; and identifying interval color depth center features of the single-cell color depth interval data to generate interval color depth center features;

[0055] Step S443: performing color depth centralization processing on the color depth gradient expression data based on the interval color depth center feature to obtain color depth centralization data; performing color depth standard deviation calculation on the color depth centralization data to obtain color depth standard deviation data; determining the data color depth of the multimodal fusion data using the color depth standard deviation data to generate data color depth feature information;

[0056] Step S444: Visualize the single-cell database architecture through data color depth feature information to obtain a single-cell data visualization report.

[0057] The present invention performs expression activity color mapping on single-cell expression activity data through color marking feature data, displays the expression activity of single cells with intuitive color changes, and enhances the visual recognition of the data; determines the single-cell expression color depth gradient of the activity color mapping data, assigns different color depths to cells with different expression levels, and thus visually distinguishes the expression activity intensity of the cells; determines the color depth range of the expression color depth gradient data, which can define the full range of color depth changes and provide a benchmark for subsequent color depth interval division; divides the expression color depth gradient data into color depth intervals according to the color depth range data, which can divide the continuous color depth gradient into discrete intervals, and thus classify the cell expression activity; and determines the single-cell color depth interval data. By identifying the center feature of interval color depth, the center point of each color depth interval can be determined, providing a key reference for subsequent color depth centralization processing; based on the interval color depth center feature, the color depth gradient data is processed for color depth centralization, and the color depth data center is standardized to make the color depth of the data more accurate; the color depth standard deviation of the color depth centralized data is calculated to measure the degree of discreteness of the color depth distribution; the data color depth of multimodal fusion data is determined through the color depth standard deviation data, which helps to provide a unified color depth standard for multimodal data fusion and enhance the consistency and comparability of the data; the single-cell database architecture is visualized through the data color depth feature information, so that the single-cell data can be presented in the form of intuitive color depth changes.

[0058] Preferably, step S444 includes the following steps:

[0059] Step S4441: performing color depth mapping on the single-cell database architecture using the data color depth feature information to obtain architecture color depth mapping data; performing color space conversion on the architecture color depth mapping data to obtain color space conversion data; performing color contrast enhancement on the single-cell database architecture based on the color space conversion data to generate color contrast enhancement information;

[0060] Step S4442: performing single-cell expression color depth detail enhancement on the single-cell database architecture based on the color contrast enhancement information to obtain color depth detail enhanced data; performing visualization component setting on the single-cell database architecture based on the color depth detail enhanced data to obtain a database visualization component; performing visualization rendering on the database visualization component to generate database visualization rendering information;

[0061] Step S4443: Perform visual effect enhancement on the database visualization rendering information to generate visual effect enhancement data; visualize the single-cell database architecture based on the visual effect enhancement data to obtain a single-cell data visualization report.

[0062] The present invention performs color depth mapping on the single-cell database architecture through data color depth feature information, provides color distinction for the visualization of single-cell data, thereby enhancing the recognizability of the data; performs color space conversion on the architecture color depth mapping data, thereby improving the intuitiveness of data visualization; performs color contrast enhancement on the single-cell database architecture according to the color space conversion data, making the contrast of the data more obvious, thereby improving the distinguishability and readability of the data; performs single-cell expression color depth detail enhancement on the single-cell database architecture based on the color contrast enhancement information, which can highlight the details in the single-cell expression data, so that even tiny expression differences can be clearly displayed; sets a visualization component for the single-cell database architecture according to the color depth detail enhancement data to obtain a database visualization component; performs visual rendering on the database visualization component to convert the data in the database into visual graphics or images, thereby generating database visualization rendering information; performs visual effect enhancement on the database visualization rendering information to further optimize the visual effect of the data; and visually displays the single-cell database architecture according to the visual effect enhancement data to obtain a single-cell data visualization report. BRIEF DESCRIPTION OF THE DRAWINGS

[0063] Figure 1 Schematic diagram of the steps of the single-cell-based database visualization method;

[0064] Figure 2 for Figure 1 Detailed implementation steps of step S2 in FIG.

[0065] Figure 3 for Figure 2Detailed implementation steps of step S24 are shown in the flowchart;

[0066] The purpose, features and advantages of the present invention will be further described with reference to the accompanying drawings and in conjunction with the embodiments. DETAILED DESCRIPTION

[0067] The following is a clear and complete description of the technical method of the present invention in conjunction with the accompanying drawings. It is obvious that the embodiments described are part of the embodiments of the present invention, but not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without making any creative efforts are within the scope of protection of the present invention.

[0068] In addition, the accompanying drawings are merely schematic illustrations of the present invention and are not necessarily drawn to scale. Identical reference numerals in the figures denote identical or similar parts, and thus repetitive descriptions thereof will be omitted. Some of the block diagrams shown in the accompanying drawings are functional entities that do not necessarily correspond to physically or logically separate entities. These functional entities may be implemented in software, in one or more hardware modules or integrated circuits, or in different network and / or processor and / or microcontroller approaches.

[0069] It should be understood that although the terms "first," "second," and the like may be used herein to describe various elements, these elements should not be limited by these terms. These terms are used solely to distinguish one element from another. For example, a first element may be referred to as a second element, and similarly, a second element may be referred to as a first element, without departing from the scope of the exemplary embodiments. The term "and / or" as used herein includes any and all combinations of one or more of the listed associated items.

[0070] To achieve this, please refer to Figures 1 to 3 , a single-cell-based database visualization method, the method comprising the following steps:

[0071] Step S1: Acquire a single-cell data set; perform standardization processing on the single-cell data set to obtain standardized single-cell data; perform high-dimensional feature extraction on the standardized single-cell data to obtain single-cell high-dimensional feature data; perform nonlinear dimensionality reduction processing on the single-cell high-dimensional feature data to obtain single-cell reduced dimensionality data;

[0072] Step S2: performing data clustering on the single-cell dimensionality reduction data to obtain single-cell cluster data; performing gene differential expression extraction on the single-cell cluster data to obtain differentially expressed gene data; performing single-cell function annotation on the differentially expressed gene data to obtain single-cell function annotation data; performing single-cell type feature identification on the single-cell cluster data based on the single-cell function annotation data to obtain single-cell type feature data; performing single-cell type classification on the single-cell cluster data based on the single-cell type feature data to obtain single-cell type data; performing multimodal data fusion on the single-cell cluster data using the single-cell type data to obtain multimodal fusion data;

[0073] Step S3: performing single-cell modality relationship analysis on the multimodal fusion data to generate single-cell relationship data; determining the database table structure of the single-cell relationship information based on the differentially expressed gene data to generate data table structure features; establishing a single-cell database index structure based on the multimodal fusion data and the data table structure information to generate database index parameters; constructing a single-cell database architecture based on the single-cell relationship data, the data table structure features, and the database index parameters to obtain a single-cell database architecture;

[0074] Step S4: Based on the single-cell database architecture, single-cell expression pattern characteristics are identified for the multimodal fusion data to obtain single-cell expression pattern characteristics; database display information is extracted for the single-cell database architecture according to the single-cell expression pattern characteristics to obtain database display data; dynamic weight marking is performed on the database display data to generate dynamic weight marking data; color marking is performed on the multimodal fusion data based on the dynamic weight marking data to obtain color marking feature data; data color depth is determined for the color marking feature data to generate data color depth feature information; the single-cell database architecture is visualized through the data color depth feature information to obtain a single-cell data visualization report.

[0075] The present invention provides basic data for subsequent analysis by acquiring single-cell data sets; standardizes the single-cell data sets to ensure data consistency and comparability; performs high-dimensional feature extraction on the standardized single-cell data to capture key information in the single-cell data; performs nonlinear dimensionality reduction on the single-cell high-dimensional feature data to effectively reduce the dimensionality of the data while retaining the most important features; performs data clustering on the single-cell dimensionality reduction data to identify single-cell populations with similar characteristics; performs gene differential expression extraction on the single-cell cluster data to identify genes with significantly different expression levels in different single-cell populations; performs single-cell functional annotation on the differentially expressed gene data; performs single-cell type feature identification on the single-cell cluster data based on the single-cell function annotation data to obtain single-cell type feature data; performs single-cell type classification on the single-cell cluster data according to the single-cell type feature data to more accurately classify single-cell types; performs multimodal data fusion on the single-cell cluster data through the single-cell type data to integrate different types of data, providing a more comprehensive perspective for subsequent analysis. Single-cell modal relationship analysis of multimodal fusion data can reveal the intrinsic connections between different data types; the database table structure of single-cell relationship information is determined based on differentially expressed gene data, providing a structured framework for subsequent data storage and management; the single-cell database index structure is established for multimodal fusion data and data table structure information to improve the efficiency of data retrieval; the single-cell database architecture is constructed based on single-cell relationship data, data table structure characteristics and database index parameters to obtain the single-cell database architecture; single-cell expression pattern feature recognition of multimodal fusion data based on the single-cell database architecture can clarify the expression behavior of single cells; database display information of the single-cell database architecture is extracted based on single-cell expression pattern features to obtain database display data; database display data is dynamically weighted to generate dynamic weighted labeled data; multimodal fusion data is color-coded based on dynamic weighted labeled data to further enhance data visualization and make key features and patterns more prominent; data color depth is determined for color-coded feature data to generate data color depth feature information; the single-cell database architecture is visualized using data color depth feature information to obtain a single-cell data visualization report. Therefore, the present invention performs multimodal data fusion analysis on single-cell data through data processing technology, feature mapping technology and visualization technology; and dynamically weights the single-cell database, thereby enhancing the visualization effect of the single-cell database.

[0076] In the embodiment of the present invention, reference Figure 1 The above is a schematic flow chart of the steps of the single-cell-based database visualization method of the present invention. In this example, the single-cell-based database visualization method includes the following steps:

[0077] Step S1: Acquire a single-cell data set; perform standardization processing on the single-cell data set to obtain standardized single-cell data; perform high-dimensional feature extraction on the standardized single-cell data to obtain single-cell high-dimensional feature data; perform nonlinear dimensionality reduction processing on the single-cell high-dimensional feature data to obtain single-cell reduced dimensionality data;

[0078] In an embodiment of the present invention, single-cell sequencing technology is used and based on a single-cell platform, such as the 10x Genomics platform, to obtain a single-cell data set; wherein the data set contains gene expression information of each cell and exists in the form of a matrix, with rows representing genes and columns representing individual cells; by using the CPM (Counts Per Million) normalization method, specifically, the count of each cell in the single-cell data set is divided by the total count of the cell, and then multiplied by one million to obtain standardized single-cell data; the standardized single-cell data is used to identify highly variable genes (HVGs), specifically, using the FindVariableFeatures function in the Seurat package, to identify HVGs by calculating the expression variability of each gene, and selecting the gene data with the largest expression to obtain single-cell high-dimensional feature data; the single-cell high-dimensional feature data is subjected to nonlinear dimensionality reduction processing, using nonlinear dimensionality reduction methods such as t-SNE (t-distributed Stochastic Evolution). NeighborEmbedding) method, specifically, reconstructs the similarity probability between high-dimensional data points in low-dimensional space, and projects the single-cell high-dimensional feature data into 2D or 3D space to obtain single-cell dimensionality reduction data.

[0079] Step S2: performing data clustering on the single-cell dimensionality reduction data to obtain single-cell cluster data; performing gene differential expression extraction on the single-cell cluster data to obtain differentially expressed gene data; performing single-cell function annotation on the differentially expressed gene data to obtain single-cell function annotation data; performing single-cell type feature identification on the single-cell cluster data based on the single-cell function annotation data to obtain single-cell type feature data; performing single-cell type classification on the single-cell cluster data based on the single-cell type feature data to obtain single-cell type data; performing multimodal data fusion on the single-cell cluster data using the single-cell type data to obtain multimodal fusion data;

[0080] In an embodiment of the present invention, the FindClusters function in the Seurat package is used to cluster the single-cell dimensionality reduction data, specifically identifying high-density cell connection points to form clusters, thereby obtaining single-cell cluster data; the FindMarkers function in the Seurat package is used to perform gene differential expression analysis on the single-cell cluster data, specifically comparing the gene expression levels in different clusters, identifying genes with significantly different expression levels in specific cell populations, and obtaining differentially expressed gene data; the differentially expressed gene data is compared with a known gene set, specifically using an enrichment function to obtain single-cell functional annotation data; based on the single-cell functional annotation data, in combination with known cell type marker genes, the cell type corresponding to each cluster is specifically identified; by comparing the differentially expressed genes with cell type specific genes, cell type annotation is automatically performed to obtain single-cell functional annotation data; the single-cell cluster data is typed according to the single-cell type feature data, specifically implemented by the AddModuleScore function in the Seurat package, to obtain single-cell type feature data; the IntegrateData function in the Seurat package is used for data fusion, specifically using CCA (canonical correlation analysis), thereby obtaining multimodal fusion data.

[0081] Step S3: performing single-cell modality relationship analysis on the multimodal fusion data to generate single-cell relationship data; determining the database table structure of the single-cell relationship information based on the differentially expressed gene data to generate data table structure features; establishing a single-cell database index structure based on the multimodal fusion data and the data table structure information to generate database index parameters; constructing a single-cell database architecture based on the single-cell relationship data, the data table structure features, and the database index parameters to obtain a single-cell database architecture;

[0082] In an embodiment of the present invention, single-cell modality relationship analysis is performed on multimodal fusion data, specifically using the FindAllMarkers function in the Seurat package to identify genes whose expression levels change in different modalities, and generating single-cell relationship data; by comparing the gene expression differences between different single cells, the interaction between cells is specifically revealed, and single-cell relationship data is generated; the database table structure of the single-cell relationship information is determined based on the differentially expressed gene data, and a relational database, such as MySQL, is used to design the data table; specifically, cell ID, gene name, expression level, and cell type are included to generate data table structure features; a single-cell database index structure is established for the multimodal fusion data and data table structure information, and an index tool provided by a database management system (DBMS), such as the CREATE INDEX statement of MySQL, is used to create indexes for columns such as cell ID and gene name, and generate database index parameters; a single-cell database architecture is constructed based on the single-cell relationship data, data table structure features, and database index parameters, and a database design tool, such as ER / Studio or MySQL Workbench, is used to construct an entity relationship diagram (ER Diagram) of the database to obtain the single-cell database architecture.

[0083] Step S4: Based on the single-cell database architecture, single-cell expression pattern characteristics are identified for the multimodal fusion data to obtain single-cell expression pattern characteristics; database display information is extracted for the single-cell database architecture according to the single-cell expression pattern characteristics to obtain database display data; dynamic weight marking is performed on the database display data to generate dynamic weight marking data; color marking is performed on the multimodal fusion data based on the dynamic weight marking data to obtain color marking feature data; data color depth is determined for the color marking feature data to generate data color depth feature information; the single-cell database architecture is visualized through the data color depth feature information to obtain a single-cell data visualization report.

[0084] In the embodiment of the present invention, based on the single-cell database architecture, the FindAllMarkers function in the Seurat package is used to identify the single-cell expression pattern characteristics of the multimodal fusion data, specifically comparing the gene expression levels of different cell populations, identifying genes with significantly different expression levels in specific cell populations, and thus obtaining the single-cell expression pattern characteristics; according to the single-cell expression pattern characteristics, the AverageExpression function in the Seurat package is used to extract database display information from the single-cell database architecture, specifically calculating the average expression level of genes in each cluster, and generating database display data; the database display data is dynamically weighted, and the order function in the R language is used in combination with the order function. Gene expression data, assigning a weight value to each gene to generate dynamic weight labeling data; based on the dynamic weight labeling data, color labeling is performed on the multimodal fusion data, specifically assigning different colors according to the level of gene expression to obtain color labeling feature data; the color depth of the color labeling feature data is determined, for example, using the scale_fill_gradientn function in the ggplot2 package to define a gradient color palette containing multiple colors to generate data color depth feature information; based on the data color depth feature information, the DimPlot function in the Seurat package is used in combination with the ggplot2 package to visualize the single-cell database architecture, and finally a single-cell data visualization report is obtained.

[0085] As an example of the present invention, refer to Figure 2 As shown, in this example, step S2 includes:

[0086] Step S21: performing structural feature recognition on the single-cell dimensionality reduction data to obtain single-cell structural data; determining hierarchical structure parameters on the single-cell structural data to generate single-cell hierarchical structure parameters; calculating single-cell density on the single-cell dimensionality reduction data based on the single-cell hierarchical structure parameters to obtain single-cell density data; and clustering the single-cell dimensionality reduction data based on the single-cell density data to obtain single-cell clustering data.

[0087] Step S22: determining gene expression characteristics of the single-cell clustering data to obtain gene expression characteristic data; extracting differential expression characteristics of the gene expression characteristic data to obtain differentially expressed gene data; performing differential feature identification on the differentially expressed gene data to generate differential feature identification data; and annotating the single-cell function of the single-cell clustering data based on the differential feature identification data and the gene expression characteristic data to obtain single-cell function annotation data.

[0088] Step S23: performing single-cell function mapping on the single-cell clustering data according to the single-cell function annotation data to generate single-cell function mapping data; performing type mapping feature determination on the single-cell function mapping data to obtain type mapping feature data; performing single-cell type feature extraction on the single-cell clustering data using the type mapping feature data to obtain single-cell type feature data;

[0089] Step S24: performing single-cell type classification on the single-cell clustering data according to the single-cell type feature data to obtain single-cell type data; performing multimodal data fusion on the single-cell clustering data through the single-cell type data to obtain multimodal fusion data.

[0090] In the embodiment of the present invention, the FindNeighbors function in the Seurat package is used to identify the structural features of the single-cell dimensionality reduction data; the FindNeighbors function identifies the local structural features in the data by constructing a proximity graph, and specifically clusters the structural features based on the FindClusters function to obtain the single-cell structural data; the hierarchical structure parameters of the single-cell structural data are determined by setting different resolution parameters to generate the single-cell hierarchical structure parameters; the resolution parameter controls the granularity of the clustering to obtain the clustering results; according to the single-cell hierarchical structure parameters, the DensityPlot function in the Seurat package is used to plot the single-cell Single-cell density was calculated from cell dimensionality reduction data; single-cell density data was obtained by calculating the density distribution of cells in each cluster; based on the single-cell density data, the single-cell dimensionality reduction data was clustered using the FindClusters function in the Seurat package; the parameters of the clustering algorithm were adjusted according to the density data to obtain single-cell clustering data; gene expression characteristics were determined for the single-cell clustering data, and the average expression level of genes in each cluster was calculated using the AverageExpression function in the Seurat package to obtain gene expression feature data; differential expression features were extracted from the gene expression feature data, and the FindM function in the Seurat package was used to extract the differential expression features. arkers function is used to identify genes with significant differences in expression in different clusters, and differentially expressed gene data are obtained; differentially expressed gene data are differentially characterized, specifically by marking the multiple of expression change and statistical significance level to generate differential characteristic identification data; based on the differential characteristic identification data and gene expression characteristic data, cell marker genes are used to perform functional enrichment analysis to obtain single-cell functional annotation data; based on the single-cell functional annotation data, the AddModuleScore function in the Seurat package is used to perform single-cell functional mapping on the single-cell cluster data, specifically calculating the expression pattern of each cell for a specific gene set to generate single-cell functional mapping data; The single-cell functional mapping data was used to determine the type mapping features, and the expression patterns of different cell functions were compared to obtain the type mapping feature data. Based on the type mapping feature data, the FindAllMarkers function in the Seurat package was used to extract the single-cell type features of the single-cell clustering data, specifically identifying genes with significantly different expression levels in specific cell types to obtain the single-cell type feature data. Based on the single-cell type feature data, the AddModuleScore function in the Seurat package was used to divide the single-cell clustering data into single-cell types, specifically assigning cells to different types according to their gene expression patterns to obtain the single-cell type data.Using the single-cell type data, the IntegrateData function in the Seurat package was used to perform multimodal data fusion on the single-cell cluster data, specifically using CCA (canonical correlation analysis) to obtain multimodal fusion data.

[0091] As an example of the present invention, refer to Figure 3 As shown, in this example, step S24 includes:

[0092] Step S241: extracting feature contours from the single-cell cluster data according to the single-cell type feature data to obtain single-cell feature contour data; determining contour boundary ranges from the single-cell feature contour data to obtain contour boundary range data;

[0093] Step S242: performing single-cell boundary division on the single-cell clustering data using the contour boundary range data to obtain single-cell boundary data; performing inter-cell relationship feature recognition on the single-cell type feature data based on the single-cell boundary data to obtain inter-cell relationship feature data; determining the degree of cell differentiation on the inter-cell relationship feature data to obtain cell differentiation degree data; and performing single-cell type division on the single-cell clustering data based on the cell differentiation degree data to obtain single-cell type data;

[0094] Step S243: extracting feature dimensions from the single cell type data to obtain type feature dimension data; dimensionally encoding the single cell type data based on the type feature dimension data to generate type dimension encoded data; and feature aligning the type dimension encoded data with the single cell clustering data to obtain single cell feature aligned data;

[0095] Step S244: performing information complementarity integration on the single-cell feature alignment data to generate information complementarity integration parameters; performing multimodal heterogeneity elimination on the single-cell feature alignment data according to the information complementarity integration parameters to obtain multimodal heterogeneity elimination data; performing fusion and collaborative calibration on the single-cell feature alignment data based on the multimodal heterogeneity elimination data to obtain multimodal fusion data.

[0096] In an embodiment of the present invention, image processing technology is used in combination with microfluidic chip technology to accurately locate and extract the outline of a single cell to obtain single-cell characteristic outline data; the single-cell characteristic outline data is analyzed to determine the boundary range of the cell, and the image segmentation technology is used to specifically identify and mark the boundary of the cell to generate outline boundary range data; based on the outline boundary range data, the image segmentation algorithm is used to divide the boundary of the single-cell clustering data to obtain single-cell boundary data; based on the single-cell boundary data, the cell-to-cell interaction is analyzed using the CellChat tool to obtain cell-to-cell relationship characteristic data; the pathway of ligand and receptor interaction between cells is specifically identified, thereby revealing the communication network between cells and obtaining cell-to-cell relationship characteristic data; the cell-to-cell relationship characteristic data is analyzed, and the specific degree of cell differentiation is determined by comparing the gene expression pattern between cells and the activation state of the signal pathway to obtain cell differentiation degree data; based on the cell differentiation degree data, combined with the cell type marker gene, the single-cell clustering data is typed to obtain single-cell type data, and the FindClusters function in the Seurat package is specifically used to perform cluster analysis and annotate the cell type. Feature dimensions are extracted from single-cell type data, and type feature dimension data is obtained by analyzing the specific gene expression patterns of each cell type; based on the type feature dimension data, dimension encoding techniques, such as principal component analysis (PCA), are used to encode the single-cell type data to generate type dimension encoded data; feature alignment is performed on the type dimension encoded data with the single-cell clustering data, and single-cell feature alignment data is obtained by adjusting and matching the feature spaces of different data sets; information complementarity is integrated on the single-cell feature alignment data, specifically by generating information complementarity integration parameters through information from different data sources; multimodal heterogeneity is eliminated on the single-cell feature alignment data based on the information complementarity integration parameters, specifically by reducing the heterogeneity between different data sources to obtain multimodal heterogeneity eliminated data; based on the multimodal heterogeneity eliminated data, fusion collaborative calibration techniques, such as the IntegrateData function in the Seurat package, are used to fuse the single-cell feature alignment data to obtain multimodal fusion data.

[0097] Preferably, step S3 includes the following steps:

[0098] Step S31: determining the modal dimension structure of the multimodal fusion data to obtain modal dimension structure data; extracting dimensional correlation from the modal dimension structure data to generate dimensional correlation information; determining the single-cell modal relationship of the multimodal fusion data based on the dimensional correlation information to generate single-cell relationship data;

[0099] Step S32: performing differentiation feature mapping on the multimodal fusion data using the cell differentiation degree data to generate differentiation feature matching data; performing relationship constraint determination on the single-cell relationship data based on the differentiation feature matching data to obtain single-cell relationship constraint data; performing primary and foreign key setting on the differentiation feature matching data based on the single-cell relationship constraint data to obtain single-cell primary and foreign key data; determining the database table structure of the single-cell relationship data using the single-cell relationship constraint data and the single-cell primary and foreign key data to generate data table structure features;

[0100] Step S33: Establish a single-cell database index based on the data table structure characteristics and generate database index parameters; construct a single-cell database architecture based on the single-cell relationship data, data table structure characteristics and database index parameters to obtain the single-cell database architecture.

[0101] In the embodiment of the present invention, the FindNeighbors function in the Seurat package is used to perform modal dimension structure on the multimodal fusion data to obtain modal dimension structure data; the modal dimension structure data is subjected to dimensional correlation extraction, specifically by calculating the correlation between different modalities, and using the cor function in the R language to calculate the correlation coefficient between different modal data to generate dimensional correlation information; based on the dimensional correlation information, the FindAllMarkers function in the Seurat package is used to determine the single-cell modal relationship of the multimodal fusion data; specifically by comparing the changes in gene expression levels in different modalities, genes with significantly different expression levels in specific cell populations are identified to generate single-cell relationship data; differentiation feature correspondence is performed on the multimodal fusion data through cell differentiation degree data, specifically by comparing Generate differentiation feature matching data based on the gene expression patterns of cells in different differentiation states; based on the differentiation feature matching data, use the AddModuleScore function in the Seurat package to perform relational constraints on the single-cell relationship data, specifically calculate the expression pattern of each cell for a specific gene set, and obtain single-cell relationship constraint data; set primary and foreign keys for the differentiation feature matching data based on the single-cell relationship constraint data, specifically by setting primary key and foreign key relationships in the database to obtain single-cell primary and foreign key data; determine the database table structure of the single-cell relationship data through the single-cell relationship constraint data and the single-cell primary and foreign key data, and generate data table structure characteristics; establish a single-cell database index based on the data table structure characteristics, and create indexes for columns such as cell ID and gene name by using index tools provided by the database management system (DBMS), such as the CREATE INDEX statement of MySQL, to generate database index parameters; construct a single-cell database architecture based on the single-cell relationship data, data table structure characteristics, and database index parameters, and use a database design tool, such as ER / Studio, to build an entity relationship diagram (ER Diagram) of the database, and create a database architecture based on the entity relationship diagram to obtain the single-cell database architecture.

[0102] Preferably, step S33 includes the following steps:

[0103] Step S331: performing type index feature recognition on the single cell type data according to the data table structure feature to generate type index feature data; determining the index structure dimension of the type index feature data based on the multimodal fusion data to obtain index structure dimension data;

[0104] Step S332: performing table structure feature recognition on the data table structure feature to obtain table structure feature data; performing structure compatibility analysis on the table structure feature data to obtain table structure compatibility information;

[0105] Step S333: performing index structure type mapping on the index structure dimension data using the table structure compatibility information to obtain index structure mapping information; performing index feature enhancement on the index structure mapping information to generate index feature enhancement data; establishing a single-cell database index based on the data table structure features according to the index feature enhancement data to generate database index parameters;

[0106] Step S334: Assign single-cell indexes to the database index parameters according to the data table structure characteristics to generate single-cell index assignment data; construct a single-cell database architecture based on the single-cell relationship data, the data table structure characteristics and the single-cell index assignment data to obtain the single-cell database architecture.

[0107] In an embodiment of the present invention, the FindNeighbors function in the Seurat package is used to determine the modal dimension structure of multimodal fusion data, specifically by constructing a proximity graph to identify local structural features in the data, and clustering based on the local structural features to obtain modal dimension structure data; dimensional correlation extraction is performed on the modal dimension structure data, which is achieved by calculating the correlation between different modalities to generate dimensional correlation information; for example, the cor function in the R language is used to calculate the correlation coefficient between different modal data; based on the dimensional correlation information, the FindAllMarkers function in the Seurat package is used to determine the single-cell modal relationship of the multimodal fusion data; by comparing the changes in gene expression levels in different modalities, genes with significantly different expression levels in specific cell populations are specifically identified to generate single-cell relationship data ; Identify the structural features of the data table by analyzing the table structure in the database to obtain table structure feature data; the table structure feature data includes information such as columns, data types, primary keys and foreign keys in the table; perform structural compatibility analysis on the table structure feature data, and obtain table structure compatibility information by checking the table structures of different data sources; perform index structure type mapping on the index structure dimension data through the table structure compatibility information, specifically matching the fields and index types in the database table to obtain index structure mapping information; perform index feature enhancement on the index structure mapping information, specifically improving database query efficiency by optimizing the index structure, and generating index feature enhancement data; establish a single-cell database index for the data table structure features based on the index feature enhancement data, by using the index tools provided by the database management system (DBMS), such as MySQL's CREATE INDEX statements are used to create indexes for columns such as cell ID and gene name, and to generate database index parameters. Single-cell indexes are assigned to the database index parameters based on the structural characteristics of the data table. Specifically, a specific value is assigned to each index in the database table to generate single-cell index assignment data. The single-cell database architecture is constructed based on the single-cell relationship data, the structural characteristics of the data table, and the single-cell index assignment data. By using database design tools such as MySQL Workbench to build the database entity relationship diagram (ER Diagram), and creating the database architecture based on the entity relationship diagram, the single-cell database architecture is obtained.

[0108] Preferably, step S332 includes the following steps:

[0109] Step S3321: extracting gene quantities from the gene expression feature data using the data table structure feature to generate gene expression quantity data; determining single cell populations from the single cell clustering data using the gene expression quantity data to obtain single cell population data;

[0110] Step S3322: performing population diversity feature identification on the single-cell population data based on the single-cell relationship data to obtain population diversity feature information; determining the single-cell population structure on the population diversity feature information to obtain single-cell population structure data; performing single-cell expression sequence identification on the gene expression quantity data using the single-cell population structure data to obtain single-cell expression sequence data; performing sequence structure extraction on the single-cell expression sequence data to generate single-cell sequence structure data;

[0111] Step S3323: integrating the single-cell population structure data and the single-cell sequence structure data to obtain single-cell data table structure features; performing single-cell multi-dimensional structure processing on the data table structure features to generate multi-dimensional structured hierarchical data; performing single-cell competitive tracking on the single-cell relationship data based on the multi-dimensional structured hierarchical data to obtain single-cell competitive data;

[0112] Step S3324: Match the single-cell competitive data with the multidimensional structure hierarchical data to obtain structure matching relationship data; quantify the structure impact of single-cell gene expression on the structure matching relationship data to obtain structure impact quantification data; perform adaptability measurement of data table structure characteristics based on the structure impact quantification data to obtain table structure adaptability data; integrate the structure impact quantification data and the table structure adaptability data to obtain table structure compatibility information.

[0113] In an embodiment of the present invention, the NormalizeData function in the Seurat package is used to normalize the single-cell type data, and then the FindVariableFeatures function is used to identify highly variable genes to generate gene expression data; the expression variability of each gene in all cells is specifically calculated, and the genes with the largest expression changes are screened out; based on the gene expression data, the FindClusters function in the Seurat package is used to divide the single-cell cluster data into groups, and the granularity of the cluster is specifically controlled by the resolution parameter resolution to obtain single-cell population data; the FindAllMarkers function in the Seurat package is used to identify population diversity characteristics of the single-cell population data, and the population diversity characteristic information is obtained by comparing the gene expression differences between different populations; the population diversity characteristic information is analyzed, and the high-dimensional data is reduced to two or three dimensions through principal component analysis (PCA) to obtain single-cell population structure data; based on the single-cell population structure data, the FeaturePlot function is used to identify single-cell expression sequences of the gene expression data to obtain single-cell expression sequence data. Single-cell expression sequence data were analyzed, and the sequence structure was extracted using the RidgePlot function to generate single-cell sequence structure data; single-cell population structure data and single-cell sequence structure data were integrated to obtain single-cell data table structure characteristics; data table structure characteristics were multi-dimensionally structured to generate multi-dimensional hierarchical structure data; based on the multi-dimensional hierarchical structure data, the FindNeighbors function in the Seurat package was used to perform single-cell competitive tracking on the single-cell relationship data to obtain single-cell competitive data; single-cell competitive data were matched with multi-dimensional hierarchical structure data to obtain structure matching relationship data; for the structure matching relationship data, the information bottleneck technology was used to quantify the influence of single-cell gene expression structure to obtain structure influence quantification data; based on the structure influence quantification data, the data table structure characteristics were adaptable to obtain table structure adaptability data; using data fusion technology, the structure influence quantification data and table structure adaptability data were integrated to obtain table structure compatibility information.

[0114] Preferably, step S4 includes the following steps:

[0115] Step S41: inputting the multimodal fusion data into the single-cell database architecture to perform single-cell edge detection to generate single-cell edge data; performing edge feature highlighting on the single-cell edge data to obtain edge feature highlighting data; determining single-cell trajectories on the multimodal fusion data based on the edge feature highlighting data to obtain single-cell trajectory data; performing single-cell state matching on the single-cell trajectory data to generate single-cell state data; performing single-cell expression pattern feature recognition on the multimodal fusion data based on the single-cell state data to obtain single-cell expression pattern features;

[0116] Step S42: extracting database display information from the single-cell database architecture based on the single-cell expression pattern characteristics to obtain database display data; dynamically weight-labeling the database display data to generate dynamic weight-labeled data;

[0117] Step S43: weight mapping the multimodal fusion data using the dynamic weight marking data to generate weight mapping data; color coding the weight mapping data to obtain color coding data; and color marking the multimodal fusion data based on the color coding data to obtain color marking feature data.

[0118] Step S44: determining the data color depth of the color-marked feature data to generate data color depth feature information; visually displaying the single-cell database architecture through the data color depth feature information to obtain a single-cell data visualization report.

[0119] In an embodiment of the present invention, image processing technology is used to perform single-cell edge detection on multimodal fusion data, specifically, the edge of the cell is identified by an edge detection algorithm, and the gradient amplitude and direction of the image are calculated to identify the edge information; the edge feature of the single-cell edge data is highlighted, and the edge feature highlighting data is generated by enhancing the edge feature; the Sobel operator is used to enhance the edge feature to obtain the edge feature highlighting data; based on the edge feature highlighting data, the Monocle function in the Seurat package is used to determine the single-cell trajectory of the multimodal fusion data to generate the single-cell trajectory data; the single-cell state matching is performed on the single-cell trajectory data, and the single-cell state data is generated by comparing the changes in the cell state; specifically, the DifferentialExpression function in the Seurat package is used to identify genes whose expression levels change at different developmental stages, thereby obtaining the single-cell state data; based on the single-cell state data, the FindAllMarkers function in the Seurat package is used to perform single-cell expression pattern feature recognition on the multimodal fusion data; by comparing different cells The expression pattern characteristics are identified by the gene expression level of the state to obtain the single-cell expression pattern characteristics; based on the single-cell expression pattern characteristics, the VisualizeData function in the Seurat package is used to extract the database display information of the single-cell database architecture and generate database display data; the database display data is dynamically weighted by assigning a weight value to each gene and determining the weight using the result of the standardized value of the gene expression amount; the multimodal fusion data is weight-mapped by the dynamic weight-labeled data to generate weight mapping data, and the weight value is mapped to the cell expression pattern; the weight mapping data is color-coded and a color mapping function is used, such as the colorRampPalette function in the R language, to assign a color to the weight value; the multimodal fusion data is color-coded based on the color-coded data to generate color-labeled feature data; the data color depth of the color-labeled feature data is determined by adjusting the color depth to generate data color depth feature information; the single-cell database architecture is visualized using the data color depth feature information to generate a single-cell data visualization report.

[0120] Preferably, step S42 includes the following steps:

[0121] Step S421: performing cell type annotation on the single-cell expression pattern features to obtain cell type annotation data; integrating the single-cell expression pattern features with the cell type annotation data to generate a single-cell expression profile; and measuring the single-cell expression level on the single-cell expression profile to obtain single-cell expression level data;

[0122] Step S422: Determine a data hierarchical relationship for the single-cell expression pattern features based on the single-cell expression level data to generate data hierarchical relationship information; generate database records for the single-cell database architecture based on the data hierarchical relationship information to obtain single-cell database record information; and format the single-cell database record information for display information to obtain database display data;

[0123] Step S423: performing expression activity feature recognition on the single-cell expression level data to generate expression activity features; performing expression activity assessment on the single-cell expression pattern features based on the expression activity features to obtain single-cell expression activity data; performing expression dynamic range assessment on the single-cell expression activity data to generate expression dynamic range data;

[0124] Step S424: Calculating a dynamic weight factor for the expression dynamic range data to obtain a dynamic weight factor; dynamically assigning weights to the single-cell expression level data based on the dynamic weight factor; if the single-cell expression level data is higher than the preset single-cell expression level data, assigning a high weight to the single-cell expression level data to generate high-weight assigned data; if the single-cell expression level data is less than or equal to the preset single-cell expression level data, assigning a low weight to the single-cell expression level data to generate low-weight assigned data; integrating the high-weight assigned data and the low-weight assigned data to obtain dynamic weight assigned data;

[0125] Step S425: Mark the database display data with display data weights using the dynamic weight allocation data to generate dynamic weight marking data.

[0126] In an embodiment of the present invention, the FindClusters function in the Seurat package is used to perform cluster analysis on the single-cell expression pattern characteristics, and specifically, the clustering results are labeled with cell types based on known cell marker genes to generate cell type labeling data; based on the cell type labeling data, the average expression level of the gene in each cell population is calculated to generate a single-cell expression profile; the single-cell expression profile is analyzed, and the ScaleData function in the Seurat package is used to standardize the gene expression data to obtain single-cell expression level data; based on the single-cell expression level data, the RunPCA function in the Seurat package is used to perform principal component analysis to specifically identify the principal components in the data and generate data hierarchical relationship information; based on the data hierarchical relationship information, database records are generated for the single-cell database architecture; the single-cell database record information is formatted, and the data is organized and formatted using the dplyr package in the R language to obtain database display data. Single-cell expression level data were analyzed, and differential expression analysis was performed using the FindMarkers function in the Seurat package to generate expression activity features; expression activity was assessed for single-cell expression pattern features based on the expression activity features to obtain single-cell expression activity data; single-cell expression activity data were analyzed, and the scale function in the R language was used to standardize the data to generate expression dynamic range data; principal component analysis of the expression dynamic range data was performed using the factoextra function in the FactoMineR package in CRAN to obtain dynamic weight factors; dynamic weight allocation was performed on single-cell expression level data based on the dynamic weight factors; if the single-cell expression level data was higher than the preset threshold, high weight allocation was performed using the ifelse function in the R language to generate high-weight allocation data; if it was less than or equal to the preset threshold, low weight allocation was performed to generate low-weight allocation data; the high-weight allocation data and the low-weight allocation data were integrated to obtain dynamic weight allocation data; based on the dynamic weight allocation data, the database display data was weight-labeled using the ggplot2 package in the R language to generate dynamic weight labeling data.

[0127] Preferably, step S44 includes the following steps:

[0128] Step S441: performing expression activity color mapping on the single-cell expression activity data using the color marking feature data to generate activity color mapping data; performing single-cell expression color depth gradient determination on the activity color mapping data to obtain expression color depth gradient data;

[0129] Step S442: determining the color depth range of the expression color depth gradient data to obtain color depth range data; dividing the expression color depth gradient data into color depth intervals according to the color depth range data to obtain single-cell color depth interval data; and identifying interval color depth center features of the single-cell color depth interval data to generate interval color depth center features;

[0130] Step S443: performing color depth centralization processing on the color depth gradient expression data based on the interval color depth center feature to obtain color depth centralization data; performing color depth standard deviation calculation on the color depth centralization data to obtain color depth standard deviation data; determining the data color depth of the multimodal fusion data using the color depth standard deviation data to generate data color depth feature information;

[0131] Step S444: Visualize the single-cell database architecture through data color depth feature information to obtain a single-cell data visualization report.

[0132] In the embodiment of the present invention, the DimPlot function in the Seurat package is used in combination with the ggplot2 package to perform color mapping on the single-cell expression activity data to generate activity color mapping data; the scale_color_gradient function in the ggplot2 package is used to establish a single-cell expression color depth gradient for the activity color mapping data to obtain expression color depth gradient data; the range function in the R language is used to determine the range of the color depth, specifically to determine the maximum and minimum values of the color depth, to obtain color depth range data; based on the color depth range data, the cut function in the R language is used to divide the expression color depth gradient data into intervals to obtain single-cell color depth interval data; the single-cell color depth interval data is divided into Analysis was performed, and the ave function in R language was used to identify the color depth center feature of each interval to generate the interval color depth center feature; based on the interval color depth center feature, the scale function in R language was used to perform color depth centering on the data expressing the color depth gradient to obtain color depth centering data; the sd function in R language was used to calculate the standard deviation of the color depth to obtain the color depth standard deviation data; based on the color depth standard deviation data, the ggplot2 package was used to determine the data color depth of the multimodal fusion data to generate data color depth feature information; based on the data color depth feature information, the DimPlot function in the Seurat package was used in combination with the ggplot2 package to visualize the single-cell database architecture and generate a single-cell data visualization report.

[0133] Preferably, step S444 includes the following steps:

[0134] Step S4441: performing color depth mapping on the single-cell database architecture using the data color depth feature information to obtain architecture color depth mapping data; performing color space conversion on the architecture color depth mapping data to obtain color space conversion data; performing color contrast enhancement on the single-cell database architecture based on the color space conversion data to generate color contrast enhancement information;

[0135] Step S4442: performing single-cell expression color depth detail enhancement on the single-cell database architecture based on the color contrast enhancement information to obtain color depth detail enhanced data; performing visualization component setting on the single-cell database architecture based on the color depth detail enhanced data to obtain a database visualization component; performing visualization rendering on the database visualization component to generate database visualization rendering information;

[0136] Step S4443: Perform visual effect enhancement on the database visualization rendering information to generate visual effect enhancement data; visualize the single-cell database architecture based on the visual effect enhancement data to obtain a single-cell data visualization report.

[0137] In an embodiment of the present invention, the DimPlot function in the Seurat package is used through the single-cell database architecture, and the color and size of the points are adjusted in combination with the ggplot2 package to obtain the architecture color depth mapping data; the architecture color depth mapping data is converted into a color space, and the colorsys package in the R language is used to realize the conversion of the color space, such as converting from RGB color space to HSV color space, to generate color space conversion data; according to the color space conversion data, image processing technology, such as histogram equalization, is used to improve the color contrast and generate color contrast improvement information; based on the color contrast improvement information, the FeaturePlot function in the Seurat package is used to enhance the single-cell expression color depth details of the single-cell database architecture; specifically, by amplifying the expression of specific genes in cells The pattern enhances color depth details to obtain color depth detail enhanced data; according to the color depth detail enhanced data, the visualization components of the single-cell database architecture are set to generate database visualization components; the database visualization components are visually rendered, and Web technologies such as HTML5 and JavaScript are used to achieve dynamic visualization effects to generate database visualization rendering information; the visual effects of the database visualization rendering information are enhanced, specifically using graphics processing technologies such as shadow and lighting effects to enhance the three-dimensional sense of visualization and generate visual effect enhanced data; according to the visual effect enhanced data, the programming language Python and corresponding libraries and packages such as Scanpy and Scran are used to process and analyze the data, and visually display them through the single-cell database architecture to generate a single-cell data visualization report.

[0138] The present invention is therefore intended to be illustrative and non-restrictive in all respects, with the scope of the invention being defined by the appended claims rather than the foregoing description, and all changes that come within the meaning and range of equivalents of the application documents are intended to be embraced therein.

[0139] The foregoing description is intended only to provide specific embodiments of the present invention, which will enable those skilled in the art to understand and implement the present invention. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present invention. Therefore, the present invention is not intended to be limited to the embodiments shown herein, but is to be construed in the widest possible manner consistent with the principles and novel features disclosed herein.

Claims

1. A single-cell-based database visualization method, characterized in that: The following steps are involved: Step S1: Acquire a single-cell data set; perform standardization processing on the single-cell data set to obtain standardized single-cell data; perform high-dimensional feature extraction on the standardized single-cell data to obtain single-cell high-dimensional feature data; perform nonlinear dimensionality reduction processing on the single-cell high-dimensional feature data to obtain single-cell reduced dimensionality data; Step S2: performing data clustering on the single-cell dimensionality reduction data to obtain single-cell cluster data; performing gene differential expression extraction on the single-cell cluster data to obtain differentially expressed gene data; performing single-cell function annotation on the differentially expressed gene data to obtain single-cell function annotation data; performing single-cell type feature identification on the single-cell cluster data based on the single-cell function annotation data to obtain single-cell type feature data; performing single-cell type classification on the single-cell cluster data based on the single-cell type feature data to obtain single-cell type data; performing multimodal data fusion on the single-cell cluster data using the single-cell type data to obtain multimodal fusion data; Step S3: Perform single-cell modality relationship analysis on the multimodal fusion data to generate single-cell relationship data; Determine the database table structure of the single-cell relationship information based on the differentially expressed gene data and generate data table structure features; establish the single-cell database index structure based on the multimodal fusion data and data table structure information and generate database index parameters; construct the single-cell database architecture based on the single-cell relationship data, data table structure features and database index parameters to obtain the single-cell database architecture; Step S4: Based on the single-cell database architecture, single-cell expression pattern characteristics are identified for the multimodal fusion data to obtain single-cell expression pattern characteristics; database display information is extracted for the single-cell database architecture according to the single-cell expression pattern characteristics to obtain database display data; dynamic weight marking is performed on the database display data to generate dynamic weight marking data; color marking is performed on the multimodal fusion data based on the dynamic weight marking data to obtain color marking feature data; data color depth is determined for the color marking feature data to generate data color depth feature information; the single-cell database architecture is visualized through the data color depth feature information to obtain a single-cell data visualization report.

2. The single-cell-based database visualization method according to claim 1, characterized in that: Step S2 includes the following steps: Step S21: performing structural feature recognition on the single-cell dimensionality reduction data to obtain single-cell structural data; determining hierarchical structure parameters on the single-cell structural data to generate single-cell hierarchical structure parameters; calculating single-cell density on the single-cell dimensionality reduction data based on the single-cell hierarchical structure parameters to obtain single-cell density data; and clustering the single-cell dimensionality reduction data based on the single-cell density data to obtain single-cell clustering data. Step S22: determining gene expression characteristics of the single-cell clustering data to obtain gene expression characteristic data; extracting differential expression characteristics of the gene expression characteristic data to obtain differentially expressed gene data; performing differential feature identification on the differentially expressed gene data to generate differential feature identification data; and annotating the single-cell function of the single-cell clustering data based on the differential feature identification data and the gene expression characteristic data to obtain single-cell function annotation data. Step S23: performing single-cell function mapping on the single-cell clustering data according to the single-cell function annotation data to generate single-cell function mapping data; performing type mapping feature determination on the single-cell function mapping data to obtain type mapping feature data; performing single-cell type feature extraction on the single-cell clustering data using the type mapping feature data to obtain single-cell type feature data; Step S24: performing single-cell type classification on the single-cell clustering data according to the single-cell type feature data to obtain single-cell type data; performing multimodal data fusion on the single-cell clustering data through the single-cell type data to obtain multimodal fusion data.

3. The single-cell-based database visualization method according to claim 2, characterized in that: Step S24 includes the following steps: Step S241: extracting feature contours from the single-cell cluster data according to the single-cell type feature data to obtain single-cell feature contour data; determining contour boundary ranges from the single-cell feature contour data to obtain contour boundary range data; Step S242: performing single-cell boundary division on the single-cell clustering data using the contour boundary range data to obtain single-cell boundary data; performing inter-cell relationship feature recognition on the single-cell type feature data based on the single-cell boundary data to obtain inter-cell relationship feature data; determining the degree of cell differentiation on the inter-cell relationship feature data to obtain cell differentiation degree data; and performing single-cell type division on the single-cell clustering data based on the cell differentiation degree data to obtain single-cell type data; Step S243: extracting feature dimensions from the single cell type data to obtain type feature dimension data; dimensionally encoding the single cell type data based on the type feature dimension data to generate type dimension encoded data; and feature aligning the type dimension encoded data with the single cell clustering data to obtain single cell feature aligned data; Step S244: performing information complementarity integration on the single-cell feature alignment data to generate information complementarity integration parameters; performing multimodal heterogeneity elimination on the single-cell feature alignment data according to the information complementarity integration parameters to obtain multimodal heterogeneity elimination data; performing fusion and collaborative calibration on the single-cell feature alignment data based on the multimodal heterogeneity elimination data to obtain multimodal fusion data.

4. The single-cell-based database visualization method according to claim 1, characterized in that: Step S3 includes the following steps: Step S31: determining the modal dimension structure of the multimodal fusion data to obtain modal dimension structure data; extracting dimensional correlation from the modal dimension structure data to generate dimensional correlation information; determining the single-cell modal relationship of the multimodal fusion data based on the dimensional correlation information to generate single-cell relationship data; Step S32: performing differentiation feature mapping on the multimodal fusion data using the cell differentiation degree data to generate differentiation feature matching data; performing relationship constraint determination on the single-cell relationship data based on the differentiation feature matching data to obtain single-cell relationship constraint data; performing primary and foreign key setting on the differentiation feature matching data based on the single-cell relationship constraint data to obtain single-cell primary and foreign key data; determining the database table structure of the single-cell relationship data using the single-cell relationship constraint data and the single-cell primary and foreign key data to generate data table structure features; Step S33: Establish a single-cell database index based on the data table structure characteristics and generate database index parameters; construct a single-cell database architecture based on the single-cell relationship data, data table structure characteristics and database index parameters to obtain the single-cell database architecture.

5. The single-cell-based database visualization method according to claim 4, characterized in that: Step S33 includes the following steps: Step S331: performing type index feature recognition on the single cell type data according to the data table structure feature to generate type index feature data; determining the index structure dimension of the type index feature data based on the multimodal fusion data to obtain index structure dimension data; Step S332: performing table structure feature recognition on the data table structure feature to obtain table structure feature data; performing structure compatibility analysis on the table structure feature data to obtain table structure compatibility information; Step S333: performing index structure type mapping on the index structure dimension data using the table structure compatibility information to obtain index structure mapping information; performing index feature enhancement on the index structure mapping information to generate index feature enhancement data; establishing a single-cell database index based on the data table structure features according to the index feature enhancement data to generate database index parameters; Step S334: Assign single-cell indexes to the database index parameters according to the data table structure characteristics to generate single-cell index assignment data; construct a single-cell database architecture based on the single-cell relationship data, the data table structure characteristics and the single-cell index assignment data to obtain the single-cell database architecture.

6. The single-cell-based database visualization method according to claim 5, characterized in that: Step S332 includes the following steps: Step S3321: extracting gene quantities from the gene expression feature data using the data table structure feature to generate gene expression quantity data; determining single cell populations from the single cell clustering data using the gene expression quantity data to obtain single cell population data; Step S3322: performing population diversity feature identification on the single-cell population data based on the single-cell relationship data to obtain population diversity feature information; determining the single-cell population structure on the population diversity feature information to obtain single-cell population structure data; performing single-cell expression sequence identification on the gene expression quantity data using the single-cell population structure data to obtain single-cell expression sequence data; performing sequence structure extraction on the single-cell expression sequence data to generate single-cell sequence structure data; Step S3323: integrating the single-cell population structure data and the single-cell sequence structure data to obtain single-cell data table structure features; performing single-cell multi-dimensional structure processing on the single-cell data table structure features to generate multi-dimensional structured hierarchical data; performing single-cell competitive tracking on the single-cell relationship data based on the multi-dimensional structured hierarchical data to obtain single-cell competitive data; Step S3324: Match the single-cell competitive data with the multidimensional structure hierarchical data to obtain structure matching relationship data; quantify the structure impact of single-cell gene expression on the structure matching relationship data to obtain structure impact quantification data; perform adaptability measurement of data table structure characteristics based on the structure impact quantification data to obtain table structure adaptability data; integrate the structure impact quantification data and the table structure adaptability data to obtain table structure compatibility information.

7. The single-cell-based database visualization method according to claim 1, characterized in that: Step S4 includes the following steps: Step S41: inputting the multimodal fusion data into the single-cell database architecture to perform single-cell edge detection to generate single-cell edge data; performing edge feature highlighting on the single-cell edge data to obtain edge feature highlighting data; determining single-cell trajectories on the multimodal fusion data based on the edge feature highlighting data to obtain single-cell trajectory data; performing single-cell state matching on the single-cell trajectory data to generate single-cell state data; performing single-cell expression pattern feature recognition on the multimodal fusion data based on the single-cell state data to obtain single-cell expression pattern features; Step S42: extracting database display information from the single-cell database architecture based on the single-cell expression pattern characteristics to obtain database display data; dynamically weight-labeling the database display data to generate dynamic weight-labeled data; Step S43: weight mapping the multimodal fusion data using the dynamic weight marking data to generate weight mapping data; color coding the weight mapping data to obtain color coding data; and color marking the multimodal fusion data based on the color coding data to obtain color marking feature data. Step S44: determining the data color depth of the color-marked feature data to generate data color depth feature information; visually displaying the single-cell database architecture through the data color depth feature information to obtain a single-cell data visualization report.

8. The single-cell-based database visualization method according to claim 7, characterized in that: Step S42 includes the following steps: Step S421: performing cell type annotation on the single-cell expression pattern features to obtain cell type annotation data; integrating the single-cell expression pattern features with the cell type annotation data to generate a single-cell expression profile; and measuring the single-cell expression level on the single-cell expression profile to obtain single-cell expression level data; Step S422: Determine a data hierarchical relationship for the single-cell expression pattern features based on the single-cell expression level data to generate data hierarchical relationship information; generate database records for the single-cell database architecture based on the data hierarchical relationship information to obtain single-cell database record information; and format the single-cell database record information for display information to obtain database display data; Step S423: performing expression activity feature recognition on the single-cell expression level data to generate expression activity features; performing expression activity assessment on the single-cell expression pattern features based on the expression activity features to obtain single-cell expression activity data; performing expression dynamic range assessment on the single-cell expression activity data to generate expression dynamic range data; Step S424: Calculating a dynamic weight factor for the expression dynamic range data to obtain a dynamic weight factor; dynamically assigning weights to the single-cell expression level data based on the dynamic weight factor; if the single-cell expression level data is higher than the preset single-cell expression level data, assigning a high weight to the single-cell expression level data to generate high-weight assigned data; if the single-cell expression level data is less than or equal to the preset single-cell expression level data, assigning a low weight to the single-cell expression level data to generate low-weight assigned data; integrating the high-weight assigned data and the low-weight assigned data to obtain dynamic weight assigned data; Step S425: Mark the database display data with display data weights using the dynamic weight allocation data to generate dynamic weight marking data.

9. The single-cell-based database visualization method according to claim 7, characterized in that: Step S44 includes the following steps: Step S441: performing expression activity color mapping on the single-cell expression activity data using the color marking feature data to generate activity color mapping data; performing single-cell expression color depth gradient determination on the activity color mapping data to obtain expression color depth gradient data; Step S442: determining the color depth range of the expression color depth gradient data to obtain color depth range data; dividing the expression color depth gradient data into color depth intervals according to the color depth range data to obtain single-cell color depth interval data; and identifying interval color depth center features of the single-cell color depth interval data to generate interval color depth center features; Step S443: performing color depth centralization processing on the color depth gradient expression data based on the interval color depth center feature to obtain color depth centralization data; performing color depth standard deviation calculation on the color depth centralization data to obtain color depth standard deviation data; determining the data color depth of the multimodal fusion data using the color depth standard deviation data to generate data color depth feature information; Step S444: Visualize the single-cell database architecture through data color depth feature information to obtain a single-cell data visualization report.

10. The single-cell-based database visualization method according to claim 9, characterized in that: Step S444 includes the following steps: Step S4441: performing color depth mapping on the single-cell database architecture using the data color depth feature information to obtain architecture color depth mapping data; performing color space conversion on the architecture color depth mapping data to obtain color space conversion data; performing color contrast enhancement on the single-cell database architecture based on the color space conversion data to generate color contrast enhancement information; Step S4442: performing single-cell expression color depth detail enhancement on the single-cell database architecture based on the color contrast enhancement information to obtain color depth detail enhanced data; performing visualization component setting on the single-cell database architecture based on the color depth detail enhanced data to obtain a database visualization component; performing visualization rendering on the database visualization component to generate database visualization rendering information; Step S4443: Perform visual effect enhancement on the database visualization rendering information to generate visual effect enhancement data; visualize the single-cell database architecture based on the visual effect enhancement data to obtain a single-cell data visualization report.

Images