Spatial transcriptome spatial domain identification method and system based on multi-scale neighborhood

Abstract

The invention belongs to the field of spatial transcriptome sequencing data processing, and provides a spatial transcriptome spatial domain identification method and system based on a multi-scale neighborhood, and the technical scheme is that based on spatial transcriptome data, graphs of different-scale neighborhoods are constructed, and height variable genes corresponding to the graphs of the different-scale neighborhoods are calculated; generating a mask pattern based on the multi-scale height variable gene and the corresponding spatial position information, extracting to obtain a multi-scale spatial embedding feature based on the mask pattern, and performing adaptive weighting on the multi-scale spatial embedding feature to obtain a spatial embedding feature; extracting a multi-scale gene embedding feature based on the multi-scale highly variable gene, and performing adaptive weighting on the multi-scale gene embedding feature to obtain a gene embedding feature; and splicing the spatial embedding features and the gene embedding features to obtain embedding features, and performing iterative clustering on the embedding features to obtain a spatial domain recognition result. The problem of poor spatial continuity caused by insufficient feature capture is solved, and accurate recognition of the spatial domain is realized.

Description

Technical Field

[0001] The present invention belongs to the field of spatial transcriptome sequencing data processing, and in particular relates to a spatial transcriptome spatial domain identification method and system based on multi-scale neighborhoods. Background Art

[0002] The statements in this section merely provide background information related to the present invention and do not necessarily constitute prior art.

[0003] With the development of biotechnology, spatial transcriptomics technology can retain spatial location information while providing comprehensive gene expression profiles. This technology provides a new perspective for understanding tissue structure, tissue function, and disease mechanisms. Spatial domain identification is one of the most important research contents of spatial transcriptomics. However, due to the characteristics of data such as high discreteness, sparsity, and multimodality, accurately identifying spatial domains from these sequencing data is a challenging task.

[0004] Traditional spatial domain recognition methods include non-spatial clustering methods (such as Louvain), which only use gene expression as input and ignore spatial location information and histological pathology image information, resulting in a lack of spatial continuity in the clustering results and an inability to accurately map the true structure of the tissue. Although some methods have improved the accuracy of spatial domain recognition to a certain extent by utilizing spatial location information or further integrating histological pathology images, there are still problems such as poor spatial continuity caused by insufficient feature capture, high algorithm spatial complexity and time complexity. Summary of the invention

[0005] In order to solve at least one technical problem existing in the above-mentioned background technology, the present invention provides a spatial transcriptome spatial domain identification method and system based on multi-scale neighborhood, which obtains multi-scale highly variable genes based on spatial transcriptome data identification, extracts spatial embedding features and gene embedding features based on the multi-scale highly variable genes, makes full use of spatial features, and improves the accuracy and efficiency of spatial domain identification.

[0006] In order to achieve the above object, the present invention adopts the following technical solution: A first aspect of the present invention provides a spatial transcriptome spatial domain identification method based on multi-scale neighborhood, comprising the following steps: Obtain spatial transcriptome data; Based on the spatial transcriptome data, we construct graphs of neighborhoods at different scales and calculate the highly variable genes corresponding to the graphs of neighborhoods at different scales. A mask map is generated based on multi-scale highly variable genes and their corresponding spatial position information, multi-scale spatial embedding features are extracted based on the mask map, and spatial embedding features are obtained by adaptively weighting the multi-scale spatial embedding features; Based on multi-scale highly variable genes, multi-scale gene embedding features are extracted, and the multi-scale gene embedding features are adaptively weighted to obtain gene embedding features; The spatial embedding features and gene embedding features are concatenated to obtain the embedded features, and the embedded features are iteratively clustered to obtain the spatial domain recognition results.

[0007] Furthermore, after obtaining the spatial transcriptome data, the spatial transcriptome data is preprocessed, including: quality control of the gene expression profile data, screening out sites and genes that meet the conditions, and then normalizing, logarithmic transformation and standardization to obtain the preprocessed gene expression profile data.

[0008] Furthermore, the process of constructing the graphs of neighborhoods of different scales includes: Based on the spatial location information in the spatial transcriptome data, the Euclidean distance between any two sites is calculated for subsequent nearest Neighbor judgment; Set a simple graph ,in yes indivual The set of , if and only if yes Recently Neighbor or yes Recently Neighboring time, and There is an edge between ,Depend on The collection composition ,Will Pick , , , we get the graphs of the neighborhoods of three different scales and ,picture , The relationship between resolutions is: , m is the number of genes, For the location.

[0009] Furthermore, the highly variable genes corresponding to the graphs of neighborhoods of different scales are calculated, including: Constructing adjacency matrices based on graphs of neighborhoods of different scales and diagonal matrix ; Based on the adjacency matrix and diagonal matrix , and the Fourier coefficients are calculated and Fourier modes ; Computational Gene , Defined as: , ,in, is the Laplace matrix No. eigenvalues, is the first Quantity, is the first Quantity, is the gene expression level, n is the total number of sites, m is the number of genes; Determine whether each gene is a highly variable gene: When a gene meets two conditions, it is considered a highly variable gene. First, its Greater than all genes The inflection point of the distribution, second, its front Low frequency of Than before High frequency of There is statistical difference.

[0010] Furthermore, based on the adjacency matrix and diagonal matrix , calculate the Fourier coefficients and Fourier modes When combined with the adjacency matrix and diagonal matrix , calculate the Laplacian matrix , n is the number of sites, using spectral decomposition , , and the eigenvalue , eigenvector , corresponding to the Fourier coefficients and Fourier modes .

[0011] Furthermore, the mask map generated based on multi-scale highly variable genes and their corresponding spatial position information can be replaced by a pruned graph, and the spatial embedding features of the corresponding scale are obtained based on the pruned graph and the graph variational autoencoder.

[0012] Furthermore, the spatial embedding feature is expressed as: , in, represents the spatial embedding feature, represents the first scale space embedding feature, Represents the first scale space embedding feature The importance of spatial domain identification, represents the second scale space embedding feature, Represents the second scale space embedding feature The importance of spatial domain identification, represents the third scale space embedding feature, Represents the third scale space embedding feature The importance of spatial domain identification; The gene embedding feature is expressed as: , in, represents the gene embedding feature, represents the first-scale gene embedding feature, Represents the first-scale gene embedding feature The importance of spatial domain identification, represents the second-scale gene embedding feature, Represents the second-scale gene embedding feature The importance of spatial domain identification, represents the third-scale gene embedding feature, Represents the third-scale gene embedding feature Importance of spatial domain identification.

[0013] A second aspect of the present invention provides a spatial transcriptome spatial domain identification system based on multi-scale neighborhoods, comprising: A data acquisition module, used to acquire spatial transcriptome data; A multi-scale highly variable gene identification module is used to construct graphs of neighborhoods of different scales based on spatial transcriptome data and calculate the highly variable genes corresponding to the graphs of neighborhoods of different scales; A spatial embedding feature extraction module is used to generate a mask map based on multi-scale highly variable genes and their corresponding spatial position information, extract multi-scale spatial embedding features based on the mask map, and adaptively weight the multi-scale spatial embedding features to obtain spatial embedding features; Gene embedding feature extraction module, which is used to extract multi-scale gene embedding features based on multi-scale highly variable genes, and adaptively weight the multi-scale gene embedding features to obtain gene embedding features; The spatial domain recognition module is used to concatenate the spatial embedding features and the gene embedding features to obtain the embedding features, and iteratively cluster the embedding features to obtain the spatial domain recognition results.

[0014] Furthermore, the system also includes a data preprocessing module, which is used to preprocess the spatial transcriptome data after obtaining the spatial transcriptome data, including: performing quality control on the gene expression spectrum data, screening out sites and genes that meet the conditions, and then normalizing, logarithmic transforming and standardizing to obtain the preprocessed gene expression spectrum data.

[0015] Furthermore, in the spatial embedding feature extraction module, the mask map generated based on multi-scale highly variable genes and their corresponding spatial position information can be replaced by a pruning graph, and the spatial embedding features of the corresponding scale are obtained based on the pruning graph and the graph variational autoencoder.

[0016] Compared with the prior art, the present invention has the following beneficial effects: The present invention is based on the identification of multi-scale highly variable gene extraction spatial embedding features and gene embedding features, considers the continuity of space, accurately maps the real structure of the tissue, solves the problem of poor spatial continuity caused by insufficient feature capture in the prior art, and realizes accurate and efficient identification in the spatial domain.

[0017] Advantages of additional aspects of the present invention will be given in part in the following description, and in part will become obvious from the following description, or will be learned through practice of the present invention. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] The accompanying drawings in the specification, which constitute a part of the present invention, are used to provide a further understanding of the present invention. The exemplary embodiments of the present invention and their descriptions are used to explain the present invention and do not constitute improper limitations on the present invention.

[0019] Figure 1 It is a flow chart of a spatial domain identification method of a spatial transcriptome based on multi-scale neighborhood provided by an embodiment of the present invention; Figure 2 It is a block diagram of a spatial domain identification method of a spatial transcriptome based on multi-scale neighborhoods provided by an embodiment of the present invention; Figure 3 The figures are effect comparison diagrams using pathological images as negative films provided by an embodiment of the present invention, wherein (a) is a manually annotated recognition effect diagram, (b) is a stLearn method recognition effect diagram, (c) is a Seurat method recognition effect diagram, (d) is a SpaceFlow method recognition effect diagram, (e) is a stAA method recognition effect diagram, and (f) is a recognition effect diagram of the method of the present invention. DETAILED DESCRIPTION

[0020] The present invention will be further described below in conjunction with the accompanying drawings and embodiments.

[0021] It should be noted that the following detailed descriptions are all illustrative and intended to provide further explanation of the present invention. Unless otherwise specified, all technical and scientific terms used herein have the same meanings as those commonly understood by those skilled in the art to which the present invention belongs.

[0022] It should be noted that the terms used herein are only for describing specific embodiments and are not intended to limit exemplary embodiments according to the present invention. As used herein, unless the context clearly indicates otherwise, the singular form is also intended to include the plural form. In addition, it should be understood that when the terms "comprising" and / or "including" are used in this specification, it indicates the presence of features, steps, operations, devices, components and / or combinations thereof.

[0023] The existing methods mentioned in the background technology improve the accuracy of spatial domain recognition to a certain extent by utilizing spatial position information or further integrating histological pathology images, but there are still problems such as poor spatial continuity caused by insufficient feature capture, high algorithm spatial complexity and time complexity. The present invention performs a graph Fourier transform of a multi-scale neighborhood based on spatial transcriptome data and identifies multi-scale highly variable genes; generates a mask map based on multi-scale highly variable genes and their corresponding spatial position information, extracts multi-scale spatial embedding features based on the mask map, and adaptively weights the multi-scale spatial embedding features to obtain spatial embedding features; extracts multi-scale gene embedding features based on multi-scale highly variable genes, and adaptively weights the multi-scale gene embedding features to obtain gene embedding features; splices the spatial embedding features and the gene embedding features to obtain embedding features, and iteratively clusters the embedding features to obtain spatial domain recognition results. The problem of poor spatial continuity caused by insufficient feature capture is solved, and accurate and efficient recognition of the spatial domain (regions that are spatially consistent with respect to gene expression profiles and histological pathology images) is achieved.

[0024] Embodiment 1 like Figure 1 and Figure 2 As shown, this embodiment provides a spatial transcriptome spatial domain identification method based on multi-scale neighborhood, comprising the following steps: Step 1: Obtain spatial transcriptome data and perform preprocessing; The spatial transcriptome data includes gene expression profile data , spatial location information With pathological images, we can perform quality control on gene expression profile data and select sites that meet the conditions. and genes, followed by normalization, logarithmic transformation and standardization to obtain the preprocessed gene expression data ,in n is the number of sites, m is the number of genes; Step 2: Preprocess the gene expression data , spatial location information Convert graph signals into frequency domain information and extract multi-scale highly variable genes; The specific steps include: Step 201: Based on spatial location information , calculate any two The Euclidean distance between them is used for the subsequent nearest Neighbor judgment; Step 202: Constructing graphs of neighborhoods of different scales , ; In this embodiment, a simple graph is assumed ,in yes indivual The set of yes Recently Neighbor or yes Recently Neighboring time ( Thought), and There is an edge between ,Depend on The collection composition .Will Pick , , , we get the graphs of the neighborhoods of three different scales , and .

[0025] Among them, Figure , and The relationship between resolutions is: ; Step 203: Calculate the graphs of neighborhoods of different scales , and The corresponding highly variable genes; The specific steps include: Step 2031: Construct an adjacency matrix based on graphs of neighborhoods of different scales and diagonal matrix ; In this embodiment, , diagonal matrix ,in for degree; Step 2032: Based on the adjacency matrix and diagonal matrix , and the Fourier coefficients are calculated , Fourier mode ; In this embodiment, combined with the adjacency matrix and diagonal matrix , calculate the Laplacian matrix , using spectral decomposition , , and the eigenvalue , eigenvector , corresponding to , ; Step 2033, calculate the gene ; The gene expression Converting to the spectral domain, we get , Defined as ,in ,in, is the Laplace matrix No. eigenvalues, is the first Quantity, is the first A quantity.

[0026] Step 2034: Determine whether each gene is a highly variable gene: When a gene meets two conditions, it is considered a highly variable gene, including: (1) its Greater than all genes The inflection point of the distribution, (2) Low frequency of Than before High frequency of There is a statistical difference; finally , and The corresponding highly variable genes; Step 3: Generate a mask map based on multi-scale highly variable genes and their corresponding spatial position information, encode and decode the mask map, capture the multi-scale spatial embedding features, and adaptively weight the multi-scale spatial embedding features to obtain the spatial embedding features; In this embodiment, based on low-resolution highly variable genes, medium-resolution highly variable genes, high-resolution highly variable genes and their corresponding spatial position information Construct graphs separately and input them into three graph variational autoencoders to obtain low-scale spatial embedding features. , mesoscale spatial embedding features With high-scale spatial embedding features .

[0027] Then, the attention mechanism is used to adaptively learn low-scale spatial embedding features. , mesoscale spatial embedding features With high-scale spatial embedding features The importance of spatial domain identification; Low-scale spatial embedding features The importance of spatial domain recognition is ,in is the activation function, is the weight matrix, is the bias vector; similarly, the mid-scale spatial embedding feature is calculated Importance of spatial domain identification and high-scale spatial embedding features Importance of spatial domain identification , and the final weighted spatial embedding feature is .

[0028] In this embodiment, in the adaptive graph variational autoencoder module, a pruning module may be selectively adopted; The low-resolution highly variable genes, medium-resolution highly variable genes, and high-resolution highly variable genes were analyzed separately. Dimensionality reduction, , , ; right , , conduct Pre-clustering, pruning is performed based on the pre-clustering results.

[0029] At once For example, if in the pre-clustering results, if a of neighbors belong to different clusters, then Prune and get the pruning graph ; Using pruned graphs in the Adaptive Graph Variational Autoencoder module Instead of directly constructing a normal graph .

[0030] Step 4: Encode and decode multi-scale highly variable genes, capture multi-scale gene embedding features, and adaptively weight the multi-scale gene embedding features to obtain gene embedding features; In this embodiment, low-resolution highly variable genes, medium-resolution highly variable genes, and high-resolution highly variable genes are respectively input into corresponding mask autoencoders to obtain low-scale gene embedding features. , mesoscale gene embedding features High-scale gene embedding features .

[0031] Then, use the attention mechanism to adaptively learn , and Importance of spatial domain recognition. Low-scale spatial embedding features The importance of spatial domain recognition is , similarly calculated , , the spatial embedding feature is 。

[0032] The present invention is based on the identification of multi-scale highly variable gene extraction spatial embedding features and gene embedding features, considers the continuity of space, accurately maps the real structure of the tissue, solves the problem of poor spatial continuity caused by insufficient feature capture in the prior art, and realizes accurate and efficient identification in the spatial domain.

[0033] Step 5: Concatenate the spatial embedding features and the gene embedding features to obtain the embedding features, and iteratively cluster the embedding features to obtain the clustering results; In this implementation, and Splicing to get embedded features .right Perform iterative clustering to obtain clustering results .

[0034] like Figure 3 The figure shows the effect comparison between the present invention and four existing methods, among which (a) is the manual annotation recognition effect diagram, (b) is the stLearn method recognition effect diagram, (c) is the Seurat method recognition effect diagram, (d) is the SpaceFlow method recognition effect diagram, (e) is the stAA method recognition effect diagram, and (f) is the recognition effect diagram of the present invention method. The experimental results show that the present invention method performs well in spatial domain recognition, can effectively improve the accuracy of spatial domain recognition, has stronger distinguishing ability and data adaptability than other methods, and has higher accuracy and robustness when processing regional boundaries that are difficult to clearly identify by other methods.

[0035] Embodiment 2 This embodiment provides a spatial transcriptome spatial domain identification system based on multi-scale neighborhoods, including: A data acquisition module, used to acquire spatial transcriptome data; A multi-scale highly variable gene identification module is used to construct graphs of neighborhoods of different scales based on spatial transcriptome data and calculate the highly variable genes corresponding to the graphs of neighborhoods of different scales; A spatial embedding feature extraction module is used to generate a mask map based on multi-scale highly variable genes and their corresponding spatial position information, extract multi-scale spatial embedding features based on the mask map, and adaptively weight the multi-scale spatial embedding features to obtain spatial embedding features; Gene embedding feature extraction module, which is used to extract multi-scale gene embedding features based on multi-scale highly variable genes, and adaptively weight the multi-scale gene embedding features to obtain gene embedding features; The spatial domain recognition module is used to concatenate the spatial embedding features and the gene embedding features to obtain the embedding features, and iteratively cluster the embedding features to obtain the spatial domain recognition results.

[0036] Furthermore, the system also includes a data preprocessing module, which is used to preprocess the spatial transcriptome data after obtaining the spatial transcriptome data, including: performing quality control on the gene expression spectrum data, screening out sites and genes that meet the conditions, and then normalizing, logarithmic transforming and standardizing to obtain the preprocessed gene expression spectrum data.

[0037] In the spatial embedding feature extraction module, the mask map generated based on multi-scale highly variable genes and their corresponding spatial position information can be replaced by a pruned map, and the spatial embedding features of the corresponding scale are obtained based on the pruned map and the graph variational autoencoder.

[0038] Since the embodiments of the system part correspond to the embodiments of the method part, please refer to the description of the embodiments of the method part for the embodiments of the system part, which will not be repeated here. And it has the same beneficial effects as the above-mentioned spatial transcriptome spatial domain identification method based on multi-scale neighborhood. Based on the identification of multi-scale highly variable gene extraction spatial embedding features and gene embedding features, considering the continuity of space, accurately mapping the real structure of the organization, solving the poor spatial continuity caused by insufficient feature capture in the prior art, and realizing accurate and efficient identification of the spatial domain.

[0039] Embodiment 3 This embodiment provides a computer-readable storage medium having a computer program stored thereon. When the program is executed by a processor, the steps in the spatial domain identification method of a spatial transcriptome based on multi-scale neighborhoods as described above are implemented.

[0040] Embodiment 4 This embodiment provides a computer device, including a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the program, the steps in the spatial domain identification method of a spatial transcriptome based on multi-scale neighborhoods as described above are implemented.

[0041] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. For those skilled in the art, the present invention may have various modifications and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.

Claims

1. A spatial transcriptome spatial domain identification method based on multi-scale neighborhood, characterized in that: The steps include: Obtain spatial transcriptome data; Based on the spatial transcriptome data, we construct graphs of neighborhoods at different scales and calculate the highly variable genes corresponding to the graphs of neighborhoods at different scales. A mask map is generated based on multi-scale highly variable genes and their corresponding spatial position information, multi-scale spatial embedding features are extracted based on the mask map, and spatial embedding features are obtained by adaptively weighting the multi-scale spatial embedding features; Based on multi-scale highly variable genes, multi-scale gene embedding features are extracted, and the multi-scale gene embedding features are adaptively weighted to obtain gene embedding features; The spatial embedding features and gene embedding features are concatenated to obtain the embedded features, and the embedded features are iteratively clustered to obtain the spatial domain recognition results.

2. The spatial domain identification method of spatial transcriptome based on multi-scale neighborhood as claimed in claim 1, characterized in that: After obtaining the spatial transcriptome data, the spatial transcriptome data is preprocessed, including: quality control of the gene expression profile data, screening out sites and genes that meet the conditions, and then normalizing, logarithmic transformation and standardization to obtain the preprocessed gene expression profile data.

3. The spatial domain identification method of spatial transcriptome based on multi-scale neighborhood as claimed in claim 1, characterized in that: The process of constructing the graph of neighborhoods of different scales includes: Based on the spatial location information in the spatial transcriptome data, the Euclidean distance between any two sites is calculated for subsequent nearest Neighbor judgment; Set a simple graph ,in yes indivual The set of , if and only if yes Recently Neighbor or yes Recently Neighboring time, and There is an edge between ,Depend on The collection composition ,Will Pick , , , we get the graphs of the neighborhoods of three different scales and ,picture , The relationship between resolutions is: , m is the number of genes, For the location.

4. The spatial domain identification method of spatial transcriptome based on multi-scale neighborhood as claimed in claim 1, characterized in that: Compute highly variable genes corresponding to graphs of neighborhoods of different scales, including: Constructing adjacency matrices based on graphs of neighborhoods of different scales and diagonal matrix ; Based on the adjacency matrix and diagonal matrix , and the Fourier coefficients are calculated and Fourier modes ; Computational Gene , Defined as: , ,in, is the Laplace matrix No. eigenvalues, is the first Quantity, is the first Quantity, is the gene expression level, n is the total number of sites, m is the number of genes; Determine whether each gene is a highly variable gene: When a gene meets two conditions, it is considered a highly variable gene. First, its Greater than all genes The inflection point of the distribution, second, its front Low frequency of Than before High frequency of There is statistical difference.

5. The spatial domain identification method of spatial transcriptome based on multi-scale neighborhood as claimed in claim 4, characterized in that: Based on the adjacency matrix and diagonal matrix , and the Fourier coefficients are calculated and Fourier modes When combined with the adjacency matrix and diagonal matrix , calculate the Laplacian matrix , n is the number of sites, using spectral decomposition , , and the eigenvalue , eigenvector , corresponding to the Fourier coefficients and Fourier modes .

6. The spatial domain identification method of spatial transcriptome based on multi-scale neighborhood as claimed in claim 1, characterized in that: The mask map generated based on multi-scale highly variable genes and their corresponding spatial position information can be replaced by a pruned map, and the spatial embedding features of the corresponding scale are obtained based on the pruned map and the graph variational autoencoder.

7. The spatial domain identification method of spatial transcriptome based on multi-scale neighborhood as claimed in claim 1, characterized in that: The spatial embedding feature is expressed as: ， in, represents the spatial embedding feature, represents the first scale space embedding feature, Represents the first scale space embedding feature The importance of spatial domain identification, represents the second scale space embedding feature, Represents the second scale space embedding feature The importance of spatial domain identification, represents the third scale space embedding feature, Represents the third scale space embedding feature The importance of spatial domain identification; The gene embedding feature is expressed as: ， in, represents the gene embedding feature, represents the first-scale gene embedding feature, Represents the first-scale gene embedding feature The importance of spatial domain identification, represents the second-scale gene embedding feature, Represents the second-scale gene embedding feature The importance of spatial domain identification, represents the third-scale gene embedding feature, Represents the third scale gene embedding feature Importance of spatial domain identification.

8. A spatial transcriptome spatial domain identification system based on multi-scale neighborhoods, characterized in that: include: A data acquisition module, used to acquire spatial transcriptome data; A multi-scale highly variable gene identification module is used to construct graphs of neighborhoods of different scales based on spatial transcriptome data and calculate the highly variable genes corresponding to the graphs of neighborhoods of different scales; A spatial embedding feature extraction module is used to generate a mask map based on multi-scale highly variable genes and their corresponding spatial position information, extract multi-scale spatial embedding features based on the mask map, and adaptively weight the multi-scale spatial embedding features to obtain spatial embedding features; Gene embedding feature extraction module, which is used to extract multi-scale gene embedding features based on multi-scale highly variable genes, and adaptively weight the multi-scale gene embedding features to obtain gene embedding features; The spatial domain recognition module is used to concatenate the spatial embedding features and the gene embedding features to obtain the embedding features, and iteratively cluster the embedding features to obtain the spatial domain recognition results.

9. The spatial transcriptome spatial domain identification system based on multi-scale neighborhood as claimed in claim 8, characterized in that: The system also includes a data preprocessing module, which is used to preprocess the spatial transcriptome data after obtaining the spatial transcriptome data, including: performing quality control on the gene expression profile data, screening out sites and genes that meet the conditions, and then performing normalization, logarithmic transformation and standardization to obtain the preprocessed gene expression profile data.

10. The spatial domain identification system of spatial transcriptome based on multi-scale neighborhood as claimed in claim 8, characterized in that: In the spatial embedding feature extraction module, the mask map generated based on multi-scale highly variable genes and their corresponding spatial position information can be replaced by a pruned map, and the spatial embedding features of the corresponding scale are obtained based on the pruned map and the graph variational autoencoder.

Images