Single cell multi-omics integration method and device
By preprocessing the single-cell gene expression and protein expression matrices, obtaining the adjacency matrix and integrating it into the multi-omics model, the problem of not utilizing cell neighbor relationships in existing technologies is solved, and more accurate single-cell multi-omics integration is achieved.
Patent Information
- Application Number
- CN202410254787.8
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-03-06
- Publication Date
- 2025-09-16
AI Technical Summary
Existing multi-omics integration methods fail to effectively utilize cell neighbor relationships, resulting in poor single-cell multi-omics integration results.
By preprocessing the gene expression matrix and protein expression matrix of single cells, the adjacency matrix indicating the neighbor relationship between cells is obtained, and it is input into the single-cell multi-omics data integration model, the multi-omics low-dimensional feature matrix is fused, and the common feature matrix of integrated single-cell multi-omics is output.
The integration of single-cell multi-omics was achieved by utilizing cell neighbor relationships, which improved the accuracy and consistency of the integration of multi-omics data.
Smart Images
Figure CN120656533A_ABST
Abstract
Description
Technical Field
[0001] The present application relates to the field of multi-omics technology, and in particular to a single-cell multi-omics integration method and device. Background Art
[0002] Multi-omics technologies have developed rapidly in recent years, especially CITE-seq technology, which can simultaneously realize the spatial positioning and detection of protein markers and mRNA molecular information of single cells.
[0003] Existing methods for integrated multi-omics analysis include totalVI, among others. These methods, based on deep learning and variational inference techniques, build a probabilistic model to model single-cell transcriptomics and proteomics, learn low-dimensional representations, and perform joint analysis. Specifically, they use a multi-autoencoder architecture to learn latent representations for individual omics. The autoencoder consists of an encoder and a decoder. The encoder maps input data (e.g., CITE-seq) into a low-dimensional latent space, while the decoder maps the latent space representation back to the original data space. While using autoencoders to learn low-dimensional representations for individual omics, totalVI assumes that each cell has multiple latent states at different omics levels. It models each latent state using a Gaussian mixture model and estimates parameters by maximizing the likelihood function of the data. These neural networks are trained to learn more accurate latent low-dimensional representations that simultaneously represent transcriptomic and proteomic expression characteristics. However, current methods for integrating single-cell multi-omics do not leverage cell neighborhood relationships to learn the characteristics of neighboring cells.
[0004] In summary, how to use cell neighbor relationships to integrate single-cell multi-omics is an urgent problem that technicians in this field need to solve. Summary of the Invention
[0005] In view of this, the present application provides a single-cell multi-omics integration method and device, aiming to integrate single-cell multi-omics by utilizing cell neighbor relationships.
[0006] In a first aspect, the present application provides a single-cell multi-omics integration method, comprising:
[0007] Preprocess the gene expression matrix of single cells to obtain a standardized feature matrix for genomics;
[0008] Preprocess the protein expression matrix of single cells to obtain a standardized feature matrix for proteomics;
[0009] Obtaining an adjacency matrix, where the adjacency matrix is used to indicate neighbor relationships between cells;
[0010] Inputting the genomics standardized feature matrix, the proteomics standardized feature matrix, and the adjacency matrix into a single-cell multi-omics data integration model;
[0011] The output is a common feature matrix integrating single-cell multi-omics.
[0012] Optionally, the preprocessing of the single-cell gene expression matrix to obtain a genomics standardized feature matrix includes:
[0013] Performing 0-1 normalization and logarithm processing on the transcriptome data in the gene expression matrix of the single cell;
[0014] The gene expression matrix of the processed single cells is reduced in dimension by principal component analysis to obtain the standardized feature matrix of genomics.
[0015] Optionally, the single-cell protein expression matrix is preprocessed to obtain a proteomics standardized feature matrix, including:
[0016] The proteomic data in the protein expression matrix of the single cell are preprocessed by the CLR normalization method to obtain the proteomics standardized feature matrix.
[0017] Optionally, obtaining the adjacency matrix includes:
[0018] Obtain m cell samples, where m is a positive integer;
[0019] Calculate the Euclidean distance between the cell samples by principal component analysis PCA;
[0020] Selecting n cell samples with the closest Euclidean distance, where n is a positive integer and is smaller than m;
[0021] Constructing a cell neighbor network graph according to the weights of the negative correlation of the Euclidean distances between the n cell samples;
[0022] An adjacency matrix is determined according to the cell neighbor network graph.
[0023] Optionally, after constructing the cell neighbor network diagram based on the weights of the negative correlations of the Euclidean distances between the n cell samples, the method further includes:
[0024] Obtain the rough label of each cell sample through pre-clustering;
[0025] removing edge connections that do not belong to the same coarse label in the neighbor relationship of each cell sample in the cell neighbor network graph to obtain a cell type-aware neighbor network graph;
[0026] Determining an adjacency matrix according to the cell neighbor network graph includes:
[0027] An adjacency matrix is determined based on the cell type-aware neighbor network graph.
[0028] Optionally, before inputting the genomics standardized feature matrix, the proteomics standardized feature matrix, and the adjacency matrix into the single-cell multi-omics data integration model, the method further comprises:
[0029] Two graph attention auto-network encoders are used to perform dimensionality reduction on the standardized feature matrix of genomics and the standardized feature matrix of proteomics respectively;
[0030] The standardized feature matrix of genomics after dimensionality reduction and the standardized feature matrix of proteomics after dimensionality reduction are fused and reconstructed to obtain a multi-omics low-dimensional feature matrix;
[0031] The step of inputting the genomics standardized feature matrix, the proteomics standardized feature matrix, and the adjacency matrix into a single-cell multi-omics data integration model comprises:
[0032] The multi-omics low-dimensional feature matrix and the adjacency matrix are input into a single-cell multi-omics data integration model.
[0033] Optionally, after inputting the multi-omics low-dimensional feature matrix and the adjacency matrix into the single-cell multi-omics data integration model, the method further comprises:
[0034] The output is a low-dimensional common feature matrix that integrates single-cell multi-omics.
[0035] In a second aspect, the present application provides a single-cell multi-omics integrated device, comprising:
[0036] The first preprocessing module is used to preprocess the gene expression matrix of single cells to obtain a standardized feature matrix for genomics;
[0037] The second preprocessing module is used to preprocess the protein expression matrix of single cells to obtain a standardized feature matrix for proteomics;
[0038] An acquisition module is used to acquire an adjacency matrix, where the adjacency matrix is used to indicate the neighbor relationship between cells;
[0039] An input module, configured to input the genomics standardized feature matrix, the proteomics standardized feature matrix, and the adjacency matrix into a single-cell multi-omics data integration model;
[0040] The first output module is used to output the common feature matrix integrating single-cell multi-omics.
[0041] Optionally, the first preprocessing module includes:
[0042] a processing unit, configured to perform 0-1 normalization and logarithm processing on the transcriptome data in the gene expression matrix of the single cell;
[0043] The dimensionality reduction unit is used to reduce the dimensionality of the processed single cell gene expression matrix through principal component analysis to obtain the standardized feature matrix of genomics.
[0044] Optionally, the second preprocessing module includes:
[0045] The preprocessing unit is used to preprocess the proteomic data in the protein expression matrix of the single cell by using a CLR normalization method to obtain the proteomics standardized feature matrix.
[0046] Optionally, the acquisition module includes:
[0047] A first acquisition unit is configured to acquire m cell samples, where m is a positive integer;
[0048] a calculation unit, configured to calculate the Euclidean distance between the cell samples by principal component analysis (PCA);
[0049] A selection unit, configured to select n cell samples with the closest Euclidean distance, where n is a positive integer and is smaller than m;
[0050] A construction unit, configured to construct a cell neighbor network graph according to the weights of the negative correlation of the Euclidean distances between the n cell samples;
[0051] A determination unit is used to determine an adjacency matrix according to the cell neighbor network graph.
[0052] Optionally, the device further includes:
[0053] The second acquisition unit is used to obtain a rough label of each cell sample through pre-clustering;
[0054] a removal unit, configured to remove edge connections that do not belong to the same coarse label in the neighbor relationship of each cell sample in the cell neighbor network graph, to obtain a cell type-aware neighbor network graph;
[0055] The determining unit includes:
[0056] A subunit is determined, which is used to perceive the neighbor network graph according to the cell type and determine the adjacency matrix.
[0057] Optionally, the device further includes:
[0058] a dimensionality reduction module, configured to perform dimensionality reduction on the standardized feature matrix of genomics and the standardized feature matrix of proteomics using two graph attention self-network encoders respectively;
[0059] A processing module is used to fuse and reconstruct the standardized feature matrix of genomics after dimensionality reduction and the standardized feature matrix of proteomics after dimensionality reduction to obtain a multi-omics low-dimensional feature matrix;
[0060] The input module comprises:
[0061] An output unit is used to input the multi-omics low-dimensional feature matrix and the adjacency matrix into a single-cell multi-omics data integration model.
[0062] Optionally, the device further includes:
[0063] The second output module is used to output a low-dimensional common feature matrix that integrates single-cell multi-omics.
[0064] In a third aspect, an embodiment of the present application provides a computer device comprising: a memory, a processor, and a computer program stored on the memory and executable on the processor. When the processor executes the computer program, it implements the single-cell multi-omics integration method described in any one of the embodiments of the first aspect of the present application.
[0065] In a fourth aspect, an embodiment of the present application provides a computer-readable storage medium, wherein the computer-readable storage medium stores instructions. When the instructions are executed on a terminal device, the terminal device executes the single-cell multi-omics integration method as described in any one of the embodiments in the first aspect of the embodiment of the present application.
[0066] The application provides a single cell multi-omics integration method. When executing the method, the gene expression matrix of the single cell is first pre-processed to obtain the standardized feature matrix of genomics; the protein expression matrix of the single cell is pre-processed to obtain the standardized feature matrix of proteomics, and then the adjacency matrix is obtained. The adjacency matrix is used to indicate the neighbor relationship between cells, and then, the standardized feature matrix of genomics, the standardized feature matrix of proteomics and the adjacency matrix are input into the single cell multi-omics data integration model, and finally the common feature matrix of the integrated single cell multi-omics is output. In this way, the single cell multi-omics is integrated by the adjacency matrix, and it is realized that the single cell multi-omics is integrated using the cell neighbor relationship. BRIEF DESCRIPTION OF THE DRAWINGS
[0067] In order to more clearly illustrate the technical solutions in this embodiment or the prior art, the following briefly introduces the drawings required for use in the embodiment or the prior art description. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.
[0068] Figure 1 A flow chart of a single-cell multi-omics integration method provided in an embodiment of the present application;
[0069] Figure 2 A cell neighbor network diagram provided in an embodiment of the present application;
[0070] Figure 3 A schematic diagram of the structure of a single-cell multi-omics integrated device provided in an embodiment of the present application;
[0071] Figure 4 A schematic diagram of the structure of a computer device provided in an embodiment of the present application. DETAILED DESCRIPTION
[0072] The following will be combined with the accompanying drawings in the embodiments of this application to clearly and completely describe the technical solutions in the embodiments of this application. This application provides a single-cell multi-omics integration method and device for use in the field of multi-omics technology. The above is only an example and does not limit the application field of the method and device provided in this application.
[0073] Multi-omics technologies have developed rapidly in recent years, especially CITE-seq technology, which can simultaneously realize the spatial positioning and detection of protein markers and mRNA molecular information of single cells.
[0074] Existing methods for integrated multi-omics analysis include totalVI, among others. These methods, based on deep learning and variational inference techniques, build a probabilistic model to model single-cell transcriptomics and proteomics, learn low-dimensional representations, and perform joint analysis. Specifically, they use a multi-autoencoder architecture to learn latent representations for individual omics. The autoencoder consists of an encoder and a decoder. The encoder maps input data (e.g., CITE-seq) into a low-dimensional latent space, while the decoder maps the latent space representation back to the original data space. While using autoencoders to learn low-dimensional representations for individual omics, totalVI assumes that each cell has multiple latent states at different omics levels. It models each latent state using a Gaussian mixture model and estimates parameters by maximizing the likelihood function of the data. These neural networks are trained to learn more accurate latent low-dimensional representations that simultaneously represent transcriptomic and proteomic expression characteristics. However, current methods for integrating single-cell multi-omics do not leverage cell neighborhood relationships to learn the characteristics of neighboring cells.
[0075] After research, the inventors proposed the technical solution of this application. When executing the method, the gene expression matrix of the single cell is first preprocessed to obtain a standardized feature matrix for genomics; the protein expression matrix of the single cell is preprocessed to obtain a standardized feature matrix for proteomics, and then an adjacency matrix is obtained. The adjacency matrix is used to indicate the neighbor relationship between cells. Then, the standardized feature matrix of genomics, the standardized feature matrix of proteomics, and the adjacency matrix are input into the single-cell multi-omics data integration model, and finally the common feature matrix of the integrated single-cell multi-omics is output. In this way, the integration of single-cell multi-omics through the adjacency matrix realizes the integration of single-cell multi-omics using cell neighbor relationships.
[0076] In order to enable those skilled in the art to better understand the present application, the present application is further described in detail below in conjunction with the accompanying drawings and specific embodiments. Obviously, the embodiments described are only a part of the embodiments of the present application, rather than all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without making creative work are within the scope of protection of the present application. It should be noted that, for ease of description, only the parts related to the relevant invention are shown in the accompanying drawings. In the absence of conflict, the embodiments in the present application and the features in the embodiments can be combined with each other.
[0077] See also Figure 1 , Figure 1 A flowchart of a single-cell multi-omics integration method provided in an embodiment of the present application includes:
[0078] S101: Preprocess the gene expression matrix of single cells to obtain a standardized feature matrix for genomics.
[0079] Due to the dimensionality differences between genomic and proteomic data, single-cell gene expression matrices and protein expression matrices must be preprocessed. The preprocessing of the single-cell gene expression matrix is as follows: the transcriptome data in the single-cell gene expression matrix is first normalized to 0-1 and logarithmized. Principal component analysis is then used to reduce the dimensionality of the processed single-cell gene expression matrix to obtain a normalized genomic feature matrix. 0-1 normalization is a data normalization method that scales the raw data to a specific, small interval, typically [0, 1]. This removes unit constraints from the data and converts it into dimensionless, pure numerical values, facilitating comparison and weighting of metrics of different units or magnitudes. Logarithmic processing is a common data preprocessing method used to process skewed data to make it closer to a normal distribution or to compress the dynamic range of the data. Principal Component Analysis (PCA) is a commonly used statistical analysis method, mainly used for dimensionality reduction of high-dimensional data. It converts a series of possibly correlated variables in the original data into a set of linearly uncorrelated variables through orthogonal transformation. These new variables are called principal components. The number of principal components is usually much smaller than the number of original variables, and it can retain most of the information of the original data.
[0080] Performing 0-1 normalization and logarithm processing on the transcriptome data in the gene expression matrix of the single cell;
[0081] The gene expression matrix of the processed single cells is reduced in dimension by principal component analysis to obtain the standardized feature matrix of genomics.
[0082] S102: Preprocess the protein expression matrix of single cells to obtain a standardized feature matrix for proteomics.
[0083] The specific method for preprocessing the protein expression matrix of a single cell is as follows: the proteomic data in the protein expression matrix of a single cell is preprocessed by the CLR normalization method to obtain a standardized feature matrix for proteomics.
[0084] S103: Obtain an adjacency matrix, which is used to indicate the neighbor relationship between cells.
[0085] First, m cell samples are obtained, where m is a positive integer. PCA is then used to calculate the Euclidean distances between these m cell samples. Next, n cell samples with the closest Euclidean distances are selected, where n is a positive integer less than m. A cell neighbor network graph is then constructed based on the weights of the negative correlations between the Euclidean distances between the n selected cell samples. Finally, the adjacency matrix is determined based on this cell neighbor network graph.
[0086] Furthermore, after constructing the cell neighbor network graph, we can also perform pre-clustering to obtain a coarse label for each cell sample. Then, we can remove the edges that do not belong to the same coarse label in the neighbor relationships of each cell sample in the cell neighbor network graph to obtain a cell type-aware neighbor network graph. Finally, we can determine the adjacency matrix based on the cell type-aware neighbor network graph.
[0087] In some embodiments, 200 cell samples are selected, and the Euclidean distances between the 200 cell samples are calculated by PCA. Then, the 10 cell samples with the closest Euclidean distances are selected, and then a cell neighbor network diagram is constructed based on the weights of the negative correlation between the Euclidean distances between the 10 selected cell samples. Figure 2 As stated, Figure 2 This example provides a cell neighbor network diagram composed of nodes and edges, where nodes represent cells and edges represent the neighbor relationships between pairs of cells in the expression feature space. Solid lines represent the same cell type, and the neighbor relationships are preserved; dashed lines represent different cell types, and the cell neighbor relationships need to be pruned. The numbers represent the edge weights, that is, the normalized Euclidean distances between principal components.
[0088] S104: Input the standardized feature matrix of genomics, the standardized feature matrix of proteomics, and the adjacency matrix into the single-cell multi-omics data integration model.
[0089] First, two graph attention auto-network encoders are used to reduce the dimensionality of the standardized feature matrix of genomics and the standardized feature matrix of proteomics. The reduced dimensionality standardized feature matrix of genomics and the reduced dimensionality standardized feature matrix of proteomics are fused and reconstructed to obtain a multi-omics low-dimensional feature matrix. Finally, the multi-omics low-dimensional feature matrix and the adjacency matrix are input into the single-cell multi-omics data integration model.
[0090] Among them, the single-cell multiomics data integration model can be a smi-DGATE (single cell multiomics integration DGATE) model.
[0091] S105: Output is the common feature matrix integrating single-cell multi-omics.
[0092] The output is a low-dimensional common feature matrix that integrates single-cell multi-omics.
[0093] In the embodiment provided in the present application, the gene expression matrix of the single cell is first pre-processed to obtain a standardized feature matrix of genomics; the protein expression matrix of the single cell is pre-processed to obtain a standardized feature matrix of proteomics, and then an adjacency matrix is obtained. The adjacency matrix is used to indicate the neighbor relationship between cells, and then the standardized feature matrix of genomics, the standardized feature matrix of proteomics, and the adjacency matrix are input into the single cell multi-omics data integration model, and finally the common feature matrix of the integrated single cell multi-omics is output. In this way, by integrating single cell multi-omics through the adjacency matrix, it is realized to integrate single cell multi-omics using cell neighbor relationships.
[0094] The above are some specific implementations of the single-cell multi-omics integration method provided in the embodiments of this application. Based on this, the present application also provides a corresponding device. The device provided in the embodiments of this application will be introduced from the perspective of functional modularization.
[0095] See also Figure 3 , Figure 3 This is a schematic diagram of the structure of a single-cell multi-omics integrated device provided in an embodiment of the present application. The single-cell multi-omics integrated device 300 includes:
[0096] A first preprocessing module 310 is used to preprocess the gene expression matrix of a single cell to obtain a standardized feature matrix for genomics;
[0097] The second preprocessing module 320 is used to preprocess the protein expression matrix of the single cell to obtain a standardized feature matrix of proteomics;
[0098] An acquisition module 330 is used to acquire an adjacency matrix, where the adjacency matrix is used to indicate neighbor relationships between cells;
[0099] An input module 340 is used to input the genomics standardized feature matrix, the proteomics standardized feature matrix, and the adjacency matrix into a single-cell multi-omics data integration model;
[0100] The first output module 350 is used to output the common feature matrix of integrated single-cell multi-omics.
[0101] Optionally, the first preprocessing module 310 includes:
[0102] a processing unit, configured to perform 0-1 normalization and logarithm processing on the transcriptome data in the gene expression matrix of the single cell;
[0103] The dimensionality reduction unit is used to reduce the dimensionality of the processed single cell gene expression matrix through principal component analysis to obtain the standardized feature matrix of genomics.
[0104] Optionally, the second preprocessing module 320 includes:
[0105] The preprocessing unit is used to preprocess the proteomic data in the protein expression matrix of the single cell by using a CLR normalization method to obtain the proteomics standardized feature matrix.
[0106] Optionally, the acquisition module 330 includes:
[0107] A first acquisition unit is configured to acquire m cell samples, where m is a positive integer;
[0108] a calculation unit, configured to calculate the Euclidean distance between the cell samples by principal component analysis (PCA);
[0109] A selection unit, configured to select n cell samples with the closest Euclidean distance, where n is a positive integer and is smaller than m;
[0110] A construction unit, configured to construct a cell neighbor network graph according to the weights of the negative correlation of the Euclidean distances between the n cell samples;
[0111] A determination unit is used to determine an adjacency matrix according to the cell neighbor network graph.
[0112] Optionally, the apparatus 300 further includes:
[0113] The second acquisition unit is used to obtain a rough label of each cell sample through pre-clustering;
[0114] a removal unit, configured to remove edge connections that do not belong to the same coarse label in the neighbor relationship of each cell sample in the cell neighbor network graph, to obtain a cell type-aware neighbor network graph;
[0115] The determining unit includes:
[0116] A subunit is determined, which is used to perceive the neighbor network graph according to the cell type and determine the adjacency matrix.
[0117] Optionally, the apparatus 300 further includes:
[0118] a dimensionality reduction module, configured to perform dimensionality reduction on the standardized feature matrix of genomics and the standardized feature matrix of proteomics using two graph attention self-network encoders respectively;
[0119] A processing module is used to fuse and reconstruct the standardized feature matrix of genomics after dimensionality reduction and the standardized feature matrix of proteomics after dimensionality reduction to obtain a multi-omics low-dimensional feature matrix;
[0120] The input module 340 includes:
[0121] An output unit is used to input the multi-omics low-dimensional feature matrix and the adjacency matrix into a single-cell multi-omics data integration model.
[0122] Optionally, the apparatus 300 further includes:
[0123] The second output module is used to output a low-dimensional common feature matrix that integrates single-cell multi-omics.
[0124] The embodiments of the present application also provide corresponding devices and computer storage media for implementing the solutions provided by the embodiments of the present application.
[0125] like Figure 4 As shown, computer device 01 is a general-purpose computing device. Components of computer device 01 may include, but are not limited to, one or more processors or processing units 03, system memory 08, and bus 04 connecting various system components (including system memory 08 and processing unit 03).
[0126] Bus 04 represents one or more of several types of bus structures, including a memory bus or memory controller, a peripheral bus, an accelerated graphics port, a processor, or a local bus using any of a variety of bus architectures. Examples of these architectures include, but are not limited to, the Industry Standard Architecture (ISA) bus, the Micro Channel Architecture (MAC) bus, the Enhanced ISA bus, the Video Electronics Standards Association (VESA) local bus, and the Peripheral Component Interconnect (PCI) bus.
[0127] The computer device 01 typically includes a variety of computer system readable media, which can be any available media that can be accessed by the computer device 01, including volatile and non-volatile media, removable and non-removable media.
[0128] System memory 08 may include computer system readable media in the form of volatile memory, such as random access memory (RAM) 09 and / or cache memory 10. Computer device 01 may further include other removable / non-removable, volatile / non-volatile computer system storage media. By way of example only, storage system 11 may be used to read and write non-removable, non-volatile magnetic media ( Figure 4 Not shown, often called a "hard drive"). Although Figure 4Not shown, a disk drive for reading and writing to a removable non-volatile disk (e.g., a "floppy disk"), and an optical disk drive for reading and writing to a removable non-volatile optical disk (e.g., a CD-ROM, DVD-ROM, or other optical media) may be provided. In these cases, each drive may be connected to the bus 04 via one or more data medium interfaces. The memory 08 may include at least one program product having a set (e.g., at least one) of program modules configured to perform the functions of various embodiments of the present invention.
[0129] A program / utility 12 having a set (at least one) of program modules 13 may be stored, for example, in memory 08. Such program modules 13 include, but are not limited to, an operating system, one or more application programs, other program modules, and program data, each of which, or some combination thereof, may include an implementation of a network environment. The program modules 13 generally implement the functions and / or methods of the embodiments described herein.
[0130] The computer device 01 may also communicate with one or more external devices 02 (e.g., a keyboard, a pointing device, a display 07, etc.), one or more devices that enable a user to interact with the computer device 01, and / or any device that enables the computer device 01 to communicate with one or more other computing devices (e.g., a network card, a modem, etc.). Such communication may be performed via an input / output (I / O) interface 06. Furthermore, the computer device 01 may also communicate with one or more networks (e.g., a local area network (LAN), a wide area network (WAN), and / or a public network, such as the Internet) via a network adapter 05. Figure 4 As shown, the network adapter 05 communicates with other modules of the computer device 01 via the bus 04. Figure 4 Not shown, other hardware and / or software modules may be used in conjunction with the computer device 01, including but not limited to: microcode, device drivers, redundant processing units, external disk drive arrays, RAID systems, tape drives, and data backup storage systems.
[0131] The processor unit 03 executes various functional applications and data processing by running programs stored in the system memory 08, such as implementing a single-cell multi-omics integrated device method provided in an embodiment of the present application.
[0132] It should be noted that, in this document, relational terms such as first and second, etc., are used only to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply the existence of any such actual relationship or order between these entities or operations. Moreover, the terms "comprises," "comprising," or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article, or device comprising a series of elements includes not only those elements, but also other elements not explicitly listed, or elements inherent to such process, method, article, or device. In the absence of further limitations, an element defined by the phrase "comprising a ..." does not exclude the presence of other identical elements in the process, method, article, or device comprising the element.
[0133] Through the description of the above embodiments, it can be known that those skilled in the art can clearly understand that all or part of the steps in the above embodiment methods can be implemented by means of software plus a general hardware platform. Based on this understanding, the technical solution of the present application can be embodied in the form of a software product, which can be stored in a storage medium, such as a read-only memory (ROM) / RAM, a magnetic disk, an optical disk, etc., and includes a number of instructions for enabling a computer device (which can be a personal computer, a server, or a network communication device such as a router) to execute the methods described in each embodiment or certain parts of the embodiments of the present application.
[0134] Each embodiment in this specification is described in a progressive manner. The same or similar parts between the embodiments can be referred to each other. Each embodiment focuses on the differences from other embodiments. In particular, for the device embodiment, since it is basically similar to the method embodiment, the description is relatively simple. For the relevant parts, refer to the partial description of the method embodiment. Some or all of the modules can be selected according to actual needs to achieve the purpose of the embodiment. Those of ordinary skill in the art can understand and implement it without paying any creative work.
[0135] The above description is merely an exemplary embodiment of the present application and is not intended to limit the scope of protection of the present application.
Claims
1. A single-cell multi-omics integration method, characterized in that: include: Preprocess the gene expression matrix of single cells to obtain a standardized feature matrix for genomics; Preprocess the protein expression matrix of single cells to obtain a standardized feature matrix for proteomics; Obtaining an adjacency matrix, where the adjacency matrix is used to indicate neighbor relationships between cells; Inputting the genomics standardized feature matrix, the proteomics standardized feature matrix, and the adjacency matrix into a single-cell multi-omics data integration model; The output is a common feature matrix integrating single-cell multi-omics.
2. The method according to claim 1, characterized in that The preprocessing of the gene expression matrix of the single cell to obtain a standardized feature matrix of genomics includes: Performing 0-1 normalization and logarithm processing on the transcriptome data in the gene expression matrix of the single cell; The gene expression matrix of the processed single cells is reduced in dimension by principal component analysis to obtain the standardized feature matrix of genomics.
3. The method according to claim 1, characterized in that The protein expression matrix of the single cell is preprocessed to obtain a standardized feature matrix of proteomics, including: The proteomic data in the protein expression matrix of the single cell are preprocessed by the CLR normalization method to obtain the proteomics standardized feature matrix.
4. The method according to claim 1, wherein The obtaining of the adjacency matrix comprises: Obtain m cell samples, where m is a positive integer; Calculate the Euclidean distance between the cell samples by principal component analysis PCA; Selecting n cell samples with the closest Euclidean distance, where n is a positive integer and is smaller than m; Constructing a cell neighbor network graph according to the weights of the negative correlation of the Euclidean distances between the n cell samples; An adjacency matrix is determined according to the cell neighbor network graph.
5. The method according to claim 4, characterized in that After constructing the cell neighbor network diagram based on the weights of the negative correlations of the Euclidean distances between the n cell samples, the method further includes: Obtain the rough label of each cell sample through pre-clustering; removing edge connections that do not belong to the same coarse label in the neighbor relationship of each cell sample in the cell neighbor network graph to obtain a cell type-aware neighbor network graph; Determining an adjacency matrix according to the cell neighbor network graph includes: An adjacency matrix is determined based on the cell type-aware neighbor network graph.
6. The method according to claim 1, characterized in that Before inputting the genomics standardized feature matrix, the proteomics standardized feature matrix, and the adjacency matrix into the single-cell multi-omics data integration model, the method further comprises: Two graph attention auto-network encoders are used to perform dimensionality reduction on the standardized feature matrix of genomics and the standardized feature matrix of proteomics respectively; The standardized feature matrix of genomics after dimensionality reduction and the standardized feature matrix of proteomics after dimensionality reduction are fused and reconstructed to obtain a multi-omics low-dimensional feature matrix; The step of inputting the genomics standardized feature matrix, the proteomics standardized feature matrix, and the adjacency matrix into a single-cell multi-omics data integration model comprises: The multi-omics low-dimensional feature matrix and the adjacency matrix are input into a single-cell multi-omics data integration model.
7. The method according to claim 6, characterized in that After inputting the multi-omics low-dimensional feature matrix and the adjacency matrix into the single-cell multi-omics data integration model, the method further comprises: The output is a low-dimensional common feature matrix that integrates single-cell multi-omics.
8. A single-cell multi-omics integrated device, characterized in that: include: The first preprocessing module is used to preprocess the gene expression matrix of single cells to obtain a standardized feature matrix for genomics; The second preprocessing module is used to preprocess the protein expression matrix of single cells to obtain a standardized feature matrix for proteomics; An acquisition module is used to acquire an adjacency matrix, where the adjacency matrix is used to indicate the neighbor relationship between cells; An input module, configured to input the genomics standardized feature matrix, the proteomics standardized feature matrix, and the adjacency matrix into a single-cell multi-omics data integration model; The first output module is used to output the common feature matrix integrating single-cell multi-omics.
9. A computer device, characterized in that: include: A memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the computer program, the single-cell multi-omics integration method according to any one of claims 1 to 7 is implemented.
10. A computer-readable storage medium, characterized in that The computer-readable storage medium stores instructions, and when the instructions are executed on a terminal device, the terminal device executes the single-cell multi-omics integration method according to any one of claims 1 to 7.