Tissue region division method, system and equipment based on space metabonomics data and medium

Abstract

The invention discloses a tissue region division method, system and device based on space metabonomics data and a medium, and relates to the field of space metabonomics, and the method comprises the steps: converting the space metabonomics data of a tissue slice into an expression matrix and a first adjacent matrix; the method comprises the following steps: performing Hamp; e, cutting the dyed slice image into a plurality of image blocks; a SpaMIC model is constructed and trained; the SpaMIC model comprises a region division module and a feature extraction module; and based on the expression matrix, the first adjacent matrix and the image block, performing region division on the tissue slice by adopting a trained SpaMIC model. According to the invention, efficient and accurate division of the space structure region of the organization can be realized.

Description

Technical Field

[0001] The present application relates to the field of spatial metabolomics, and in particular to a method, system, device and medium for dividing tissue regions based on spatial metabolomics data. Background Art

[0002] As an emerging scientific research technology, spatial metabolomics has shown great potential in analyzing the spatial distribution of metabolites in cells and tissues. This technology can provide the distribution of metabolite information in microscopic space. This information usually shows significant regional specificity and is crucial for revealing the fine regulatory mechanism of cellular metabolic activities. However, the current analysis tools for spatial metabolomics data are not yet mature, especially in terms of regional division and integrated analysis. On the one hand, due to the limitations of the data format of spatial metabolomics data itself, different data preprocessing and feature extraction are required for this type of data; on the other hand, due to the relatively lack of spatial heterogeneity revealing ability of metabolomics data, it is necessary to combine other omics data, such as transcriptomics data and optical image information, for joint analysis to obtain more accurate regional division and more comprehensive biological information. Summary of the invention

[0003] The purpose of this application is to provide a method, system, device and medium for tissue region division based on spatial metabolomics data, which can achieve efficient and accurate division of the spatial structural regions of the tissue.

[0004] To achieve the above objectives, this application provides the following solutions:

[0005] In a first aspect, the present application provides a method for tissue region division based on spatial metabolomics data, comprising:

[0006] A method for dividing tissue regions based on spatial metabolomics data, characterized by comprising:

[0007] The spatial metabolomics data of tissue sections were converted into expression matrices and first adjacency matrices;

[0008] Crop the H&E stained slice image into multiple image blocks;

[0009] Constructing and training a SpaMIC model; the SpaMIC model includes a region division module and a feature extraction module;

[0010] Based on the expression matrix, the first adjacency matrix and the image block, the trained SpaMIC model is used to perform region division on the tissue slice.

[0011] In a second aspect, the present application provides a system for dividing tissue regions based on spatial metabolomics data, comprising:

[0012] a conversion module for converting spatial metabolomics data of tissue sections into expression matrices and first adjacency matrices;

[0013] An image block cropping module is used to crop the H&E stained slice image into multiple image blocks;

[0014] A model building and training module, used to build and train a SpaMIC model; the SpaMIC model includes a region division module and a feature extraction module;

[0015] A region division module is used to perform region division on the tissue slice using the trained SpaMIC model based on the expression matrix, the first adjacency matrix and the image block.

[0016] In a third aspect, the present application provides a computer device, comprising: a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the above-mentioned method for tissue region division based on spatial metabolomics data.

[0017] In a fourth aspect, the present application provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the above-mentioned method for tissue region division based on spatial metabolomics data.

[0018] According to the specific embodiments provided in this application, this application has the following technical effects:

[0019] The present application provides a method, system, device and medium for tissue region division based on spatial metabolomics data. The spatial metabolomics data of tissue sections usually show obvious regional specificity. The SpaMIC model can efficiently process and analyze spatial metabolomics data by integrating deep learning, greatly improving the efficiency of data analysis and realizing efficient and accurate division of spatial structural regions of tissues. BRIEF DESCRIPTION OF THE DRAWINGS

[0020] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings required for use in the embodiments will be briefly introduced below. 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 paying creative work.

[0021] Figure 1 A schematic diagram of a process for dividing tissue regions based on spatial metabolomics data provided in one embodiment of the present application;

[0022] Figure 2Schematic diagram of the results of processing the pig embryo dataset using the SpaMIC model in implementation case 1; wherein (a) is the original H&E stained histological image, (b) is the spatial structure domain analyzed by the SpaMIC model, (c) is the spatial structure domain analyzed by the SpaMIC model using the default Louvain clustering method, (d) is the spatial structure domain analyzed by the SmartGate method, and (e) is the spatial structure domain analyzed by the GraphST method;

[0023] Figure 3 Schematic diagram of the results of processing mouse kidney slices using the SpaMIC model in implementation case 2; wherein (a) is the original H&E stained histological image, (b) is the domain analysis result of spatial metabolomics data, (c) is the domain analysis result of H&E omics image data, and (d) is the result after mapping modification;

[0024] Figure 4 Schematic diagram of the results of processing mouse brain slices using the SpaMIC model in implementation case 3; (a) is the spatial structural domain analyzed using only spatial metabolomics data, (b) is the spatial structural domain analyzed after integrating spatial transcriptomics features, (c) is the spatial structural domain analyzed after integrating H&E image feature omics features, and (d) is the spatial distribution map of dopamine analysis;

[0025] Figure 5 A schematic diagram of the structure of a computer device provided in one embodiment of the present application. DETAILED DESCRIPTION

[0026] The following will be combined with the drawings in the embodiments of the present application to clearly and completely describe the technical solutions in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of this application.

[0027] In the context of existing technologies, this application can obtain rich spatial texture detail information of tissues by analyzing optical microscopic images of tissue sections, which can be used as reference images and combined with the metabolite information of the spatial metabolome to achieve refined division of tissue spatial regions. In addition, within the tissue, combined with transcriptomics sequencing technology, the cell types, functions, and gene expressions of different regions show significant heterogeneity for integrated analysis, which helps to achieve more accurate regional division and more comprehensive interpretation of biological characteristics.

[0028] In order to make the above-mentioned objects, features and advantages of the present application more obvious and easy to understand, the present application is further described in detail below in conjunction with the accompanying drawings and specific implementation methods.

[0029] In an exemplary embodiment, Figure 1 As shown, a method for dividing tissue regions based on spatial metabolomics data is provided, which is executed by a computer device, specifically, it can be executed by a computer device such as a terminal or a server alone, or it can be executed by a terminal and a server together. In the embodiment of the present application, the method is applied to a server as an example for explanation, including the following steps S1 to S4. Wherein:

[0030] S1: Conversion of spatial metabolomics data of tissue sections into expression matrix and first adjacency matrix.

[0031] The spatial metabolomics data were first converted into an expression matrix X and a first adjacency matrix A, where rows in the expression matrix X represent pixels and columns represent selected m / z values. The expression matrix was normalized and logarithmically transformed using the scanpy package, where the sum of expression values ​​within a single pixel after normalization was 1e 4 The K nearest neighbor algorithm is used to transform the spatial position coordinates into an undirected adjacency network, where node i and node j are k nearest neighbors if and only if they are each other, and the default value of k is 5. In addition, a method for calculating the adjacency matrix A based on the Euclidean distance range with a predefined radius r is also provided, which can be flexibly selected in different data sets.

[0032] S2: Crop the H&E stained slice image into multiple image blocks.

[0033] The center point is determined using a fixed spacing, and an image block of pixels within the center point range W×H is cropped, where H and W are the height and width of the image block, respectively. In this embodiment, H and W are set to 224 (in a specific data set, H and W are set to 112), where the fixed spacing is 200.

[0034] Alternatively, an image block of pixels in a range of W×H is cropped around each coordinate point of the corresponding H&E stained slice image according to the spatial coordinates of the spatial omics data, where H and W are the height and width of the image block, respectively. In this embodiment, H and W are set to 256 (in a specific data set, H and W are set to 112).

[0035] S3: construct and train a SpaMIC model; the SpaMIC model includes a region division module and a feature extraction module.

[0036] (1) Area division module

[0037] The processing of the regional division module is the initial processing step of the SpaMIC model. In order to effectively reduce the dimension of the original data and the interference of noise, this application introduces an iterative clustering method to extract metabolites with regional representativeness.

[0038] The region segmentation module adopts a graph attention network autoencoder. As a deep learning network, the graph attention network autoencoder can learn node relationships from the graph structure and reconstruct node features. It mainly consists of two parts: an encoder and a decoder. The encoder adopts the basic structure of the GAT_module module of the MSpti model. The structure of the decoder is symmetrical to the structure of the encoder. This application calls such a complete graph attention network autoencoder the SpaMIC_AE model. The training process minimizes the reconstruction loss of the normalized expression as follows:

[0039]

[0040] Among them, x i is the expression value of the original matrix, x′ i The expression values ​​of the matrix output by the SpaMIC_AE model.

[0041] The region division module includes two iterative training processes. In the first iterative training, the sample expression matrix and the first sample adjacency matrix are used as input to train the SpaMIC_AE model (i.e., the region division module), and the sample expression matrix and the first sample adjacency matrix are converted from the spatial metabolomics data of the sample tissue section. Its input layer and output layer are equal to the feature number of the sample expression matrix, the number of hidden layers is 512-30, and the training uses the Adam optimizer, the initial learning rate, and the training cycle. The low-dimensional vector output by the encoder is tacitly clustered using the scanpy package for louvain, and the marker molecules in different cell groups are obtained by differential analysis. For each cell group, the top 1000 feature molecules are extracted as columns to form a new sample expression matrix, and the number of cells of the updated sample expression matrix is ​​the same as the original expression matrix. In addition, the feature molecules of interest can be added to the updated sample expression matrix according to the molecules of the known m / z mass-to-charge ratio to avoid the omission of metabolite molecules with regional representativeness in the automatic feature screening process. In the second iteration training, the updated sample expression matrix and the first sample adjacency matrix are used as input to train the SpaMIC_AE model, whose input layer and output layer are equal to the number of features of the updated sample expression matrix Y, and the rest of the training parameters are the same as above. It is particularly noted that in the processing of this module, it is not limited to the processing of spatial metabolomics data, and the expression matrix processed by other spatial omics data such as spatial transcriptomics data also has similar effects.

[0042] (2) Feature extraction module

[0043] The feature extraction module combines image information to perform fine processing and boundary optimization on the results of regional division. The feature extraction module adopts the ResNet-50 network structure. This embodiment uses the ResNet-50 network structure pre-trained by ImageNet21k to extract 2048-dimensional feature vectors from image blocks, and the composed image feature expression matrix and the second adjacency matrix are used as the input of the regional division module.

[0044] S4: Based on the expression matrix, the first adjacency matrix and the image block, the trained SpaMIC model is used to perform regional division on the tissue slice. Specifically, it includes: inputting the expression matrix and the first adjacency matrix into the regional division module to obtain a first regional division result; inputting the image block into the feature extraction module to extract feature vectors to form an image feature expression matrix and a second adjacency matrix; inputting the image feature expression matrix and the second adjacency matrix into the regional division module to obtain a second regional division result; based on the first regional division result and the second regional division result, mutual information is used to perform image registration to obtain the final regional division result of the tissue slice.

[0045] Since the stained image slices and sequenced tissue slices are usually from adjacent rather than the same tissue slice, before adjusting and correcting the regional division boundaries of sequencing genomics based on the image information, the image data space division result S i (i.e., the second region division result, the moving image) is registered to the spatial division result S of the spatial transcriptomics data using the mutual information image m (i.e. the first region division result, fixed image) in the same reference coordinate system. Specifically, the mutual information between two images can be reflected by entropy and joint entropy, and the mutual information between images can be maximized through rigid transformations such as translation, rotation, and scaling as follows:

[0046] MI(s i ,s m )=H(s i )+H(s m )-H(s i ,s m )

[0047] Among them, MI(s i ,s m ) is the mutual information, H(s i ) and H(s m ) are the entropy of the moving image and the fixed image respectively, H(s i ,s m ) is the joint entropy of the two images. In the overlapped images after registration, the spatial morphological information of cells can be used to adjust the spatial metabolomics domain to make it more consistent with the actual distribution.

[0048] In an exemplary embodiment, the SpaMIC model constructed in the present application further includes a feature fitting module; the feature fitting module is used to fit multimodal data. The SpaMIC model is also used to perform regional division of the tissue slice based on the expression matrix, the first adjacency matrix and other omics feature matrices. Specifically:

[0049] (1) Inputting the expression matrix, the first adjacency matrix and other omics feature matrices into the feature fitting module to obtain an integrated feature expression matrix.

[0050] The feature fitting module can fit the spatial metabolomics data and other omics data, and integrate the features x from the two modalities. i (expression matrix and first adjacency matrix) and y i (Other omics feature matrices) are input into the multimodal autoencoder. The structure of the encoder and decoder is the same as that in the SpaMIC_AE model. The hidden layer dimension is set to 1000-1000. The low-dimensional feature z output by the encoder x and z y The initial joint expression z is learned through a nonlinear neural network as the input feature of the decoder. In addition, the low-dimensional feature z x and z y The feature representation h is calculated separately through an additional neural network x and h y , the feature dimension is 128, and the data of the two modalities are adapted through the strategy of contrastive learning. The overall loss function includes the standard reconstruction loss based on the mean square error of the two modal data and the contrastive learning loss function based on the normalized temperature scaled cross entropy loss function (NT-Xent loss). The multimodal autoencoder uses the Adam optimizer, with an initial learning rate lr = 0.0005 and a training cycle n = 500. The specific method is as follows:

[0051]

[0052]

[0053] z=f projector (concat(z x ,z y ))

[0054] h x =f projector (z x )

[0055] h y =f projector (z y )

[0056]

[0057] Among them, f encoderx (·) and f encodery (·) are the decoders corresponding to the two modes in the multimodal autoencoder, and f projector (·) is a nonlinear neural network consisting of two linear layers and a ReLU activation layer, τ is the temperature parameter, sim(·) is the dot product of the vector, α is the weight of the reconstruction loss part and defaults to 1, and β is the weight of the contrast loss part and defaults to 1.

[0058] (2) Inputting the integrated feature expression matrix and the first adjacency matrix into the region division module to obtain the region division result.

[0059] In order to further test the analytical ability of the SpaMIC model constructed in this application in the spatial structure domain, and to verify its performance in boundary recognition and multi-omics feature fitting, this application selected three sets of spatial metabolomics data as applications, including a pig embryo dataset, a mouse kidney dataset, and a mouse Parkinson's injury spatial multimodal analysis (SMA) dataset.

[0060] (I) Implementation Case 1

[0061] (1) Data Collection

[0062] The porcine embryo dataset was captured using a Thermo LTQ instrument with desorption electrospray ionization (DESI) to capture cross-sectional slices of porcine fetuses, and contains 4959 pixels and 10200 mass spectral peaks with a mass range of m / z 150-1000. The normalized dataset can be accessed through the R package Cardinal.

[0063] (2) Data processing

[0064] The original spatial metabolomics data of the slices were converted into the expression matrix X and the first adjacency matrix A. Then, the regional chemical characteristics of the tissue were divided using the regional partitioning module in the SpaMIC model, and its parameters were set to the default parameters. Similarly, the collected data sets were also benchmarked using Louvain clustering, SmartGate and GraphST (version 1.1.1) in Scanpy, and the parameters were set to the default parameters, and a comparison was made with the SpaMIC method.

[0065] (3) Result analysis:

[0066] By H&E staining of sections (such as Figure 2The main morphology of the pig embryo can be clearly divided, but the identification of some regional edges and unclear staining areas is still very vague. The collected spatial metabolomics data was processed by the regional division module of SpaMIC, and 3994 regional representative molecular features were extracted from the original 10200 characteristic molecules, reducing the data size by about 61% and obtaining a better regional characteristic contour map (as shown in Figure 1). Figure 2 The results of SpaMIC regional segmentation were also compared with other methods, among which the Louvain clustering in Scanpy restored the structure of the pig embryo to a certain extent (as shown in (b) in Figure 2). Figure 2 However, it is greatly affected by data noise, and there are many errors at the boundaries and inside of different organs. Although SmartGate and GraphST methods can identify several major organs such as brain, heart and liver in pig embryos (as shown in (c) in Figure 2), Figure 2 However, the divided tissue area is larger than the actual tissue and the number of identified types is insufficient.

[0067] (II) Implementation Case 2

[0068] (1) Data Collection

[0069] Mouse kidneys (7-month-old female C57) were cryopreserved at −80°C and then cut into two consecutive sections with a cryostat, 20 μm and 10 μm thick, respectively. Mass spectrometry imaging experiments were performed on thicker sections using a desorption spray ionization (DESI) ion source (Prosolia, USA) coupled to a Synapt G2-Si mass spectrometer (Waters, UK) with a step size of 100 × 100 μm (i.e., pixel resolution). Chromatographic grade methanol (Merck, Germany) was used in DESI experiments at a flow rate of 1.5 μL / min. The positive ion mode was selected, with a capillary voltage of 4.2 kV and a sampling cone voltage of 80 V for DESI. The nebulizing nitrogen pressure of DESI was 0.5 MPa, and the ion transfer capillary temperature was 150°C. The mass range was 200–1000 m / z, with a mass resolution of 10,000. Thinner sections were subjected to standard hematoxylin and eosin (H&E) staining protocols and then digital pathology slides were scanned at 20x magnification (i.e., 0.5 × 0.5 μm pixel resolution). The collected mouse kidney dataset is available through the ProteomeXchangeConsortium database (accession number PXD038876) (data in public databases should be used for tool testing).

[0070] (2) Data processing

[0071] The original spatial metabolomics data of the slices are converted into the expression matrix X and the first adjacency matrix A. Then, the regional division module in the SpaMIC model is used to divide the chemical characteristics of the tissue regions. A rough regional division outline map can be obtained. Then, the feature extraction module in the SpaMIC model is used to fuse the H&E stained tissue regional division information into the above regional division outline map to refine the division of tissue structure regions.

[0072] (3) Results analysis

[0073] Through the H&E tissue staining of mouse kidney sections, the kidney tissue can be roughly divided into several main areas: cortex, medulla, and pelvis. In the spatial structure analysis diagram (MSI) using metabolomics data alone, these three areas can be presented more clearly. However, there are a certain number of metabolite information distribution results that are not ideal. For example, a part of the metabolites on the left side of the pelvic region (MSI metabolite cluster 2) have a jagged protrusion, while this part appears blank in the H&E stained tissue area. This is because the environmental noise signal has caused some interference to the division of the spatial clustering results. Subsequently, the tissue area division information of the optical image H&E staining (such as Figure 3 (a) in the figure) is applied to boundary recognition and adjustment, so that the spatial region structure division result (such as Figure 3 (b) in the figure can be obviously corrected, and its spatial clustering result is closer to the actual division (as shown in Figure 3 This further illustrates the necessity of using tissue imaging image omics data for regional division, and also reflects the superiority of the method and strategy of this application. The SpaMIC model combines the two modal features to obtain a tissue structure distribution map with refined division results (such as Figure 3 As shown in (d) in the figure, the boundary division between the pelvic region and the medulla region is more consistent with the actual situation.

[0074] (III) Implementation Case 3

[0075] (1) Data Collection

[0076] H&E images from the mouse Parkinson's lesion spatial multimodal analysis (SMA) dataset are available at Mendeley Data, and raw MALDI-MSI data in imzML format and MS / MS spectra from tissue collections in mzML and vendor-specific baf formats are available at Figshare. In situ sequencing data are available at Zenodo. Spatial transcriptomics raw sequencing data, microscopy images, and Spaceranger output files for mouse samples are available at Figshare.

[0077] (2) Data processing

[0078] The original slice mass spectrometry imaging data is converted into the expression matrix X and the first adjacency matrix A. Then, the regional division module in the SpaMIC model is used to divide the chemical characteristics of the tissue region. A rough regional division outline map can be obtained. Then, the feature extraction module in the SpaMIC model is used to fuse the H&E stained tissue regional division information into the above regional division outline map to refine the division of tissue structure regions. Finally, the integrated analysis module in the SpaMIC model is used to fit and analyze the spatial metabolomics data and spatial transcriptome data.

[0079] (3) Results analysis

[0080] To demonstrate the advantages of SpaMIC in multimodal joint analysis, the SNc slice data (SMA dataset number V11T16-085_C1) were integrated and analyzed. Figure 4 As shown, the regional division of spatial metabolomics data alone (such as Figure 4 (a) in Figure 2) and regional division integrating spatial metabolome and spatial transcriptome data (as shown in Figure 2 Figure 4 Compared with (b) in Figure 2, it is found that some small areas of the spatial metabolome division are not identified, and important information is lost. The integrated data is then integrated into the information of the H&E staining image to make the regional division clearer (as shown in Figure 2). Figure 4 Comparing the division results of these three types of data, it is found that the addition of transcriptome data allows SpaMIC to divide an additional specific region, and this specific region is consistent with the dopamine-enriched site (such as Figure 4 This also shows that SpaMIC can obtain more refined spatial structural domains by integrating and analyzing multi-omics data and then performing regional division, reflecting the advantages of the feature fitting module in SpaMIC.

[0081] Based on the same inventive concept, the embodiment of the present application also provides a system for implementing the above-mentioned method for tissue region division based on spatial metabolomics data. The implementation scheme for solving the problem provided by the system is similar to the implementation scheme recorded in the above method, so the specific limitations in one or more embodiments of the tissue region division system based on spatial metabolomics data provided below can be referred to the limitations of the tissue region division method based on spatial metabolomics data above, and will not be repeated here.

[0082] In an exemplary embodiment, a system for dividing tissue regions based on spatial metabolomics data is provided, comprising:

[0083] A conversion module is used to convert the spatial metabolomics data of tissue sections into expression matrices and first adjacency matrices.

[0084] The image block cropping module is used to crop the H&E stained slice image into multiple image blocks.

[0085] The model building and training module is used to build and train the SpaMIC model; the SpaMIC model includes a region division module and a feature extraction module.

[0086] A region division module is used to perform region division on the tissue slice using the trained SpaMIC model based on the expression matrix, the first adjacency matrix and the image block.

[0087] In an exemplary embodiment, a computer device is provided, including a memory and a processor, wherein a computer program is stored in the memory, and the processor implements the steps in the above-mentioned method embodiments when executing the computer program. The computer device may be a server or a terminal, and its internal structure diagram may be as shown in FIG. Figure 5 As shown. The computer device includes a processor, a memory, an input / output interface (Input / Output, referred to as I / O) and a communication interface. Wherein, the processor, the memory and the input / output interface are connected through a system bus, and the communication interface is connected to the system bus through the input / output interface. Wherein, the processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, a computer program and a database. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The database of the computer device is used to store data to be processed. The input / output interface of the computer device is used to exchange information between the processor and an external device. The communication interface of the computer device is used to communicate with an external terminal through a network connection. When the computer program is executed by the processor, a method for dividing tissue regions based on spatial metabolomics data is implemented.

[0088] Those skilled in the art will understand that Figure 5 The structure shown in the figure is only a block diagram of a part of the structure related to the solution of the present application, and does not constitute a limitation on the computer device to which the solution of the present application is applied. The specific computer device may include more or fewer components than those shown in the figure, or combine certain components, or have a different arrangement of components. In an exemplary embodiment, a computer device is provided, including a memory and a processor, wherein a computer program is stored in the memory, and the processor implements the steps in the above-mentioned method embodiments when executing the computer program.

[0089] In an exemplary embodiment, a computer-readable storage medium is provided, storing a computer program, and when the computer program is executed by a processor, the steps in the above method embodiments are implemented.

[0090] In an exemplary embodiment, a computer program product is provided, including a computer program, and when the computer program is executed by a processor, the steps in the above method embodiments are implemented.

[0091] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, stored data, displayed data, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties, and the collection, use and processing of relevant data must comply with relevant regulations.

[0092] Those of ordinary skill in the art can understand that all or part of the processes in the above-mentioned embodiment methods can be completed by instructing the relevant hardware through a computer program, and the computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to the memory, database or other medium used in the embodiments provided in the present application can include at least one of non-volatile and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetic random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. By way of illustration and not limitation, RAM may be in various forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM).

[0093] The database involved in each embodiment provided in this application may include at least one of a relational database and a non-relational database. The non-relational database may include a distributed database based on blockchain, etc., but is not limited thereto. The processor involved in each embodiment provided in this application may be a general-purpose processor, a central processing unit, a graphics processor, a digital signal processor, a programmable logic device, a data processing logic device based on quantum computing, etc., but is not limited thereto.

[0094] The technical features of the above embodiments may be combined arbitrarily. To make the description concise, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0095] This article uses specific examples to illustrate the principles and implementation methods of this application. The description of the above embodiments is only used to help understand the method and core ideas of this application. At the same time, for those skilled in the art, according to the ideas of this application, there will be changes in the specific implementation methods and application scope. In summary, the content of this specification should not be understood as limiting this application.

Claims

1. A method for tissue region division based on spatial metabolomics data, characterized in that: include: The spatial metabolomics data of tissue sections were converted into expression matrices and first adjacency matrices; Crop the H&E stained slice image into multiple image blocks; Constructing and training a SpaMIC model; the SpaMIC model includes a region division module and a feature extraction module; Based on the expression matrix, the first adjacency matrix and the image block, the trained SpaMIC model is used to perform region division on the tissue slice.

2. The method for tissue region division based on spatial metabolomics data according to claim 1, characterized in that: Based on the expression matrix, the first adjacency matrix and the image block, the trained SpaMIC model is used to perform region division on the tissue slice, specifically including: Inputting the expression matrix and the first adjacency matrix into the region division module to obtain a first region division result; Inputting the image block into the feature extraction module to extract feature vectors, and obtaining an image feature expression matrix and a second adjacency matrix; Inputting the image feature expression matrix and the second adjacency matrix into the region division module to obtain a second region division result; Based on the first region division result and the second region division result, the image registration is performed using mutual information to obtain the final region division result of the tissue slice.

3. The method for tissue region division based on spatial metabolomics data according to claim 1, characterized in that: The region division module includes two iterative training processes; In the first iteration training process, the sample expression matrix and the first sample adjacency matrix are used as inputs of the region partition module, and the output is the feature numerator of the sample expression matrix; The sample expression matrix and the first sample adjacency matrix are obtained by transforming the spatial metabolomics data of the sample tissue slices; Updating the sample expression matrix based on the characteristic molecules; In the second iterative training, the updated sample expression matrix and the first sample adjacency matrix are used as inputs of the region partition module.

4. The method for tissue region division based on spatial metabolomics data according to claim 1, characterized in that: The region segmentation module adopts a graph attention network autoencoder.

5. The method for tissue region division based on spatial metabolomics data according to claim 1, characterized in that: The feature extraction module adopts the ResNet-50 network structure.

6. The method for tissue region division based on spatial metabolomics data according to claim 1, characterized in that: The SpaMIC model also includes a feature fitting module; The SpaMIC model is also used to perform regional segmentation on the tissue section based on the expression matrix, the first adjacency matrix and other omics feature matrices.

7. The method for tissue region division based on spatial metabolomics data according to claim 6, characterized in that: The tissue section is divided into regions based on the expression matrix, the first adjacency matrix and other omics feature matrices, specifically comprising: The expression matrix, the first adjacency matrix and other omics feature matrices are input into the feature fitting module to obtain an integrated feature expression matrix. The integrated feature expression matrix and the first adjacency matrix are input into the region division module to obtain a region division result.

8. A tissue region partitioning system based on spatial metabolomics data, characterized in that: include: a conversion module for converting spatial metabolomics data of tissue sections into expression matrices and first adjacency matrices; An image block cropping module is used to crop the H&E stained slice image into multiple image blocks; A model building and training module, used to build and train a SpaMIC model; the SpaMIC model includes a region division module and a feature extraction module; A region division module is used to perform region division on the tissue slice using the trained SpaMIC model based on the expression matrix, the first adjacency matrix and the image block.

9. A computer device comprising: A memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the method for tissue region segmentation based on spatial metabolomics data as described in any one of claims 1 to 7.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the method for dividing tissue regions based on spatial metabolomics data described in any one of claims 1 to 7 is implemented.

Images