Method, system and terminal for three-dimensional construction of single-cell spatial transcriptome

Abstract

The three-dimensional construction method, system and terminal of the single-cell spatial transcriptome of the application capture the position-specific spatial transcriptome data of the mouse early embryo development primitive gut movement period by using the spatial transcriptome technology, and then the spatial transcriptome atlas of the outer, middle and inner germ layers is three-dimensionally reconstructed by establishing a mathematical model, a full-genome space-time expression database is established, the full-genome high-resolution digital in-situ hybridization atlas of the single-cell precision of the mouse embryo development primitive gut movement period is realized, it is the most comprehensive, complete and highest-precision interactive space-time transcriptome database in the world at present. Moreover, the application has the functions of identifying spatial marker genes and spatial positioning of single cells (or cell groups). Meanwhile, the spatial single-cell algorithm contained in the application simulates the spatial distribution of single cells in different germ layers, and further three-dimensionally reconstructs the spatial transcriptome atlas of the single-cell precision, the work provides comprehensive data and brand-new ideas for understanding the germ layer lineage establishment and the fate regulation mechanism of pluripotent stem cells, and promotes the development of the early embryo development and stem cell regenerative medicine related fields.

Description

Technical Field

[0001] This invention relates to the field of spatial transcriptomics technology, and in particular to a method, system, and terminal for constructing a three-dimensional spatial transcriptome in a single cell. Background Technology

[0002] In recent years, single-cell transcriptomics technology has greatly advanced molecular and cell biology research, allowing the study of interactions between individual cells at the whole-cell transcriptome level. However, during single-cell sequencing, cell location information is also lost. For developmental biology research, especially early embryonic development, cell location information largely determines cell developmental fate.

[0003] However, existing spatial transcriptomics technology has certain limitations. It does not provide a comprehensive, complete, and high-precision method for studying the spatiotemporal dynamic molecular expression patterns of early embryonic development at the whole genome level. It cannot "restore" embryonic tissue cells or exogenous pluripotent stem cells to specific locations in the in vivo embryo, nor can it locate cells of unknown origin in corresponding parts of the in vivo embryo. Therefore, it is urgent to combine spatial transcriptomics technology, single-cell transcriptomics technology, and mathematical modeling to establish a high-resolution digital in situ hybridization map of the entire genome with single-cell precision during the gastrulation period of mouse embryonic development, as well as an interactive spatiotemporal transcriptomics database. Summary of the Invention

[0004] In view of the shortcomings of the prior art described above, the purpose of this invention is to provide a method, system and terminal for constructing a three-dimensional single-cell spatial transcriptome, so as to solve the above technical problems in the prior art.

[0005] To achieve the above and other related objectives, this invention provides a method for constructing a three-dimensional spatial transcriptome of a single cell. The method includes: segmenting embryos at a target developmental stage using spatial transcriptomics technology, capturing location-specific spatial transcriptomics data, and accumulating a target number of embryonic spatial transcriptomics samples; obtaining two-dimensional spatial expression patterns of each gene based on a corn map model and the spatial transcriptomics data of each embryonic sample; obtaining three-dimensional spatial expression patterns of each gene based on a three-dimensional corn map model and the spatial transcriptomics data of each embryonic sample; mapping each single cell onto the embryo at the target developmental stage using a multidimensional single-cell mapping model based on single-cell transcriptomics data and the spatial transcriptomics data of each embryonic sample to determine the optimal mapping position of the single cell; simulating the shape of different germ layers of an early embryo using a spatial simulation model, and further simulating and obtaining a single-cell-precision embryonic spatial map based on the optimal mapping position of the single cell; and calculating the Euclidean distance between single cells using a spatial optimization model to simulate the three-dimensional spatial distribution map of the single cells.

[0006] In one embodiment of the present invention,

[0007] Using spatial transcriptomics technology, embryos at the target developmental stage are segmented to capture location-specific spatial transcriptomics data, accumulating a target number of embryonic spatial transcriptomics samples. This includes: horizontally cutting the embryos at the target developmental stage along their proximal and distal axes to obtain m sequentially arranged slices; performing laser microdissection on each slice based on the germ layer developmental state to obtain n embryo samples; and performing transcriptome sequencing on each embryo sample to obtain corresponding spatial transcriptomics data. Specifically, the laser microdissection on each slice based on the germ layer developmental state includes: based on the segmentation rules of the first slice located at the distal and distal axes, laser microdissection of the ectoderm and endoderm is performed according to the developmental state of the first slice, with each segmented region representing one embryo sample; the segmentation rules of the first slice include: segmenting the anterior and posterior ends of the ectoderm as segmentation regions; and segmenting the entire endoderm as a segmentation region; based on the segmentation rules of the second slice, laser microdissection of the second slice... During development, the ectoderm, mesoderm, and endoderm are laser-microscopically dissected, with each dissected area serving as an embryo sample. The segmentation rules for the second layer slice include: segmenting the ectoderm by its anterior, posterior, left, and right sides; segmenting the mesoderm by its anterior and posterior sides; and segmenting the endoderm by its anterior and posterior sides. Based on standard segmentation rules, and according to the developmental state of layers other than the first and second layers, the ectoderm, mesoderm, and endoderm are laser-microscopically dissected, with each dissected area serving as an embryo sample. The standard segmentation rules include: segmenting the ectoderm by its anterior, posterior, left anterior, left posterior, right anterior, and right posterior sides; segmenting the mesoderm and endoderm of layers other than the first and second layers; and segmenting the endoderm and endoderm of layers other than the first and second layers.

[0008] In one embodiment of the present invention, obtaining the two-dimensional spatial expression pattern of each gene based on the spatial transcriptome data of each embryo sample includes: for each layer of laser microdissection sample, arranging each embryo sample from left to right on a plane according to the anteroposterior axis of the embryo; adjusting the spacing between each embryo sample to assign a planar coordinate to each embryo sample; and drawing a corresponding color for each embryo sample based on the expression value of each gene in each sample to obtain the two-dimensional spatial expression pattern of each gene.

[0009] In one embodiment of the present invention, obtaining the three-dimensional spatial expression pattern of each gene based on the spatial transcriptome data of each embryo sample specifically includes: mapping each embryo sample onto a three-dimensional model used to simulate the shape of the embryo, and assigning three-dimensional spatial coordinates to each embryo sample; drawing corresponding colors for each embryo sample based on the expression values ​​of each gene in each sample, so as to obtain the three-dimensional spatial expression pattern of each gene.

[0010] In one embodiment of the present invention, the step of mapping each single cell to an embryo at a target developmental stage using a multidimensional single-cell mapping model based on single-cell transcriptome data and spatial transcriptome data of each embryo sample to determine the optimal mapping position of the single cell includes: extracting spatially specific marker genes based on single-cell transcriptome data, wherein the marker genes divide the target embryo into multiple spatial domains; calculating the Spearman correlation coefficient between each single cell and the embryo sample; calculating the three-dimensional spatial coordinates corresponding to each single cell based on the Spearman correlation coefficient corresponding to each single cell sample and in combination with the three-dimensional spatial coordinates of the embryo sample; calculating the distance between the corresponding three-dimensional spatial coordinates and the three-dimensional spatial coordinates of each embryo sample, and taking the three-dimensional spatial coordinates of the embryo sample with the smallest corresponding distance as the optimal mapping position of the corresponding single cell.

[0011] In one embodiment of the present invention, the step of calculating the corresponding three-dimensional spatial coordinates based on the Spearman correlation coefficient corresponding to each single-cell sample and the spatial coordinates of the embryo sample includes: selecting the three largest Spearman correlation coefficients among the Spearman correlation coefficients corresponding to each single-cell sample; and, in combination with the three-dimensional spatial coordinates of the embryo sample, applying a spatial smoothing algorithm to calculate the three-dimensional spatial coordinates of each single cell based on the three largest Spearman correlation coefficients corresponding to each single cell and the three-dimensional spatial coordinates of the corresponding embryo sample.

[0012] In one embodiment of the present invention, the step of simulating the shape of different germ layers of an early embryo based on a spatial simulation model, and further simulating and obtaining a single-cell-precision embryonic spatial atlas based on the optimal mapping position of a single cell includes: simulating the germ layer shape of an embryo at a target developmental stage to obtain a target embryo shape model; establishing a circular model and segmenting the circular space to simulate the embryo samples segmented by spatial transcriptomics technology; locating single cells to their corresponding embryo sample positions based on the optimal mapping position of a single cell, drawing corresponding colors for single cells according to cell type, and drawing corresponding colors for each single cell according to the expression level of each gene in the single cell; and using a bubble algorithm based on the gene expression of single cells to spatially rearrange the single cells in the circular space region where each embryo sample is located to obtain a single-cell-precision embryonic spatial atlas.

[0013] In one embodiment of the present invention, the step of calculating the Euclidean distance between single cells based on a spatial optimization model to simulate the three-dimensional spatial distribution of single cells includes: extracting the single-cell gene expression profiles of each annular spatial region in the embryonic spatial atlas with single-cell precision and performing a logarithmic transformation; calculating the Euclidean distance between every two cells in each annular spatial region based on the logarithmically transformed single-cell gene expression profiles in each annular spatial region, and establishing a corresponding Euclidean distance matrix; extracting the farthest transcriptome distance between single cells and the maximum distance within each annular spatial region in each Euclidean distance matrix, and standardizing the Euclidean distance matrix to obtain a standardized Euclidean distance matrix; using the functions of each annular spatial region as constraints, applying the least squares method, and simulating the optimal coordinates of each cell according to the corresponding standardized Euclidean distance matrix to obtain a three-dimensional spatial distribution atlas of single cells.

[0014] To achieve the above and other related objectives, this invention provides a three-dimensional construction system for single-cell spatial transcriptomes. The system includes: a segmentation module for segmenting embryos at a target developmental stage using spatial transcriptomics technology, capturing location-specific spatial transcriptomics data, and accumulating a target number of embryonic spatial transcriptomics samples; a two-dimensional spatial expression module connected to the segmentation module for obtaining the two-dimensional spatial expression pattern of each gene based on the spatial transcriptomics data of each embryonic sample; a three-dimensional spatial expression module connected to the segmentation module for obtaining the three-dimensional spatial expression pattern of each gene based on a three-dimensional corn map model and the spatial transcriptomics data of each embryonic sample; and a multidimensional single-cell mapping module connected to... The segmentation module and the three-dimensional spatial expression module are used to map each single cell onto the target developmental stage embryo based on single-cell transcriptome data and spatial transcriptome data of each embryo sample using a multidimensional single-cell mapping model to determine the optimal mapping position of the single cell. The spatial simulation module, connected to the multidimensional single-cell mapping module and the three-dimensional spatial expression module, is used to simulate the shape of different germ layers of the early embryo based on the spatial simulation model, and further simulate to obtain a single-cell accurate embryonic spatial map based on the optimal mapping position of the single cell. The spatial optimization module, connected to the spatial simulation module, is used to calculate the Euclidean distance between single cells based on the spatial optimization model to simulate the three-dimensional spatial distribution map of the single cells.

[0015] To achieve the above and other related objectives, the present invention provides a three-dimensional construction terminal for a single-cell spatial transcriptome, comprising: one or more memory devices and one or more processors; the one or more memory devices are used to store a computer program; the one or more processors are connected to the memory devices and are used to run the computer program to execute the three-dimensional construction method of the single-cell spatial transcriptome.

[0016] As described above, this invention provides a method, system, and terminal for constructing a three-dimensional spatial transcriptome of a single cell, offering the following advantages: This invention captures location-specific spatial transcriptome data using spatial transcriptome technology and, based on an established mathematical model, reconstructs three-dimensional spatial transcriptome maps of the ectoepithelioblastoma (outer, mesoepithelioblastoma, and endoderm), establishing an encyclopedic spatiotemporal expression database of the entire genome. The spatial transcriptome map constructed by this invention achieves high-resolution digital in situ hybridization maps of the entire genome with single-cell precision, making it the most comprehensive, complete, and accurate interactive spatiotemporal transcriptome database currently available. Attached Figure Description

[0017] Figure 1 The diagram shows a flowchart of a method for constructing a three-dimensional spatial transcriptome in a single cell according to an embodiment of the present invention.

[0018] Figure 2 The diagram shown is a schematic diagram of E7.5 embryo spatial segmentation and the establishment of a two-dimensional spatial expression mode in an embodiment of the present invention.

[0019] Figure 3 The diagram shows the steps for establishing a three-dimensional spatial representation pattern of the ectoderm in one embodiment of the present invention.

[0020] Figure 4 The diagram shown is a schematic representation of the three germ layers (ectoderm, mesoderm, and endoderm) of an E7.5 embryo based on a circular model in one embodiment of the present invention.

[0021] Figure 5 The diagram shown is a schematic diagram of cutting and dividing a circular model of three germ layers in one embodiment of the present invention.

[0022] Figure 6 The diagram shown is a schematic representation of a model space in which single cells are uniformly distributed at each embryo sample location, as described in one embodiment of the present invention.

[0023] Figure 7 The diagram shown is a schematic representation of the bubble sort algorithm formula in one embodiment of the present invention.

[0024] Figure 8 The diagram shows a single-cell rearrangement within the 8P space of an E7.5 embryo in one embodiment of the present invention.

[0025] Figure 9 The diagram shows the single-cell spatial expression pattern of the gene Bmp4 in the ectoderm of an E7.5 embryo, according to an embodiment of the present invention.

[0026] Figure 10 The diagram shown is a spatial atlas of E7.5 embryonic single-cell precision in one embodiment of the present invention.

[0027] Figure 11The diagram shows the operation steps of the spatial optimization model in one embodiment of the present invention.

[0028] Figure 12 The diagram shown is a schematic representation of the spatial expression pattern of gene T in one embodiment of the present invention.

[0029] Figure 13 The diagram shows the spatial expression pattern of gene T in the ectoderm, mesoderm, and endoderm in one embodiment of the present invention.

[0030] Figure 14 The diagram shows a heatmap illustrating the expression patterns of 10 marker genes in an E7.5 embryo according to an embodiment of the present invention.

[0031] Figure 15 The diagram shows the spatial expression patterns of 10 marker genes and 9 spatial domains of an E7.5 embryo in one embodiment of the present invention.

[0032] Figure 16 The diagram shown is a schematic of the operation process of a multidimensional single-cell mapping model in one embodiment of the present invention.

[0033] Figure 17 This diagram illustrates the verification of spatial mapping accuracy of a single cell separated from a known location in an E7.5 embryo, according to an embodiment of the present invention.

[0034] Figure 18 The diagram shows a single-cell cluster distribution of an E7.5 embryo in one embodiment of the present invention.

[0035] Figure 19 The diagram shows the spatial distribution of various cell types in an E7.5 embryo according to one embodiment of the present invention.

[0036] Figure 20 The diagram shown is a schematic representation of the clustering results of single cells at position 6P of embryo E6.75 in an embodiment of the present invention, based on t-SNE analysis.

[0037] Figure 21 The diagram shown is a simulation of the spatial distribution of single cells within position 6P of embryo E6.75 in one embodiment of the present invention.

[0038] Figure 22 The diagram shown is a schematic representation of a three-dimensional construction system for a single-cell spatial transcriptome according to an embodiment of the present invention.

[0039] Figure 23 The diagram shown is a schematic representation of the three-dimensional construction terminal of a single-cell spatial transcriptome in one embodiment of the present invention. Detailed Implementation

[0040] The following specific examples illustrate the implementation of the present invention. Those skilled in the art can easily understand other advantages and effects of the present invention from the content disclosed in this specification. The present invention can also be implemented or applied through other different specific embodiments, and various details in this specification can also be modified or changed based on different viewpoints and applications without departing from the spirit of the present invention. It should be noted that, unless otherwise specified, the following embodiments and features described therein can be combined with each other.

[0041] It should be noted that in the following description, reference is made to the accompanying drawings, which illustrate several embodiments of the invention. It should be understood that other embodiments may also be used, and changes in mechanical composition, structure, electrical system, and operation may be made without departing from the spirit and scope of the invention. The following detailed description should not be considered limiting, and the scope of the embodiments of the invention is defined only by the claims of the published patents. The terminology used herein is for the purpose of describing particular embodiments only and is not intended to limit the invention. Spatially related terms, such as “upper,” “lower,” “left,” “right,” “below,” “below,” “lower part,” “above,” “upper part,” etc., may be used herein to illustrate the relationship between one element or feature shown in the figures and another element or feature.

[0042] Throughout this specification, when it is said that a part is "connected" to another part, this includes not only "direct connection" but also "indirect connection" by placing other elements in between. Furthermore, when it is said that a part "includes" a certain constituent element, unless otherwise stated otherwise, this does not exclude other constituent elements, but rather means that other constituent elements may also be included.

[0043] The terms "first," "second," and "third," etc., used herein are for the purpose of describing various parts, components, regions, layers, and / or segments, but are not limiting. These terms are used only to distinguish one part, component, region, layer, or segment from others. Therefore, the "first part," "component," "region," "layer," or "segment" described below may refer to a "second part," "component," "region," "layer," or "segment" without departing from the scope of this invention.

[0044] Furthermore, as used herein, the singular forms “a,” “an,” and “the” are intended to include the plural forms as well, unless the context indicates otherwise. It should be further understood that the terms “comprising,” “including,” indicate the presence of the stated feature, operation, element, component, item, kind, and / or group, but do not preclude the presence, occurrence, or addition of one or more other features, operations, elements, components, items, kinds, and / or groups. The terms “or” and “and / or” as used herein are interpreted as inclusive, or mean any one or any combination thereof. Thus, “A, B, or C” or “A, B, and / or C” means “any one of: A; B; C; A and B; A and C; B and C; A, B, and C.” Exceptions to this definition arise only when combinations of elements, functions, or operations are inherently mutually exclusive in some manner.

[0045] This invention provides a method, system, and terminal for constructing a three-dimensional spatial transcriptome of a single cell. By capturing location-specific spatial transcriptome data based on spatial transcriptome technology and using an established mathematical model, it reconstructs three-dimensional spatial transcriptome maps of the ectoepithelial, mesoderm, and endoderm, establishing an encyclopedic spatiotemporal expression database of the entire genome. The spatial transcriptome map constructed by this invention achieves high-resolution digital in situ hybridization maps of the entire genome at single-cell precision, and is currently the most comprehensive, complete, and accurate interactive spatiotemporal transcriptome database.

[0046] The present invention will now be described in detail with reference to the accompanying drawings, so that those skilled in the art can readily implement it. The present invention can be embodied in many different forms and is not limited to the embodiments described herein.

[0047] like Figure 1 A flowchart illustrating a method for constructing a three-dimensional spatial transcriptome in a single cell according to an embodiment of the present invention is shown.

[0048] Using our established geo-seq technique, we captured position-specific geo-seq data for early mouse embryonic gastrulation (E6.5, E6.75, E7.0, E7.25, E7.5). We then constructed three-dimensional cellular geo-seq models for each developmental stage, using embryos at each developmental stage as target developmental periods.

[0049] The method includes:

[0050] Step S1: Using spatial transcriptomics technology, embryos at the target developmental stage are segmented to capture location-specific spatial transcriptomics data, accumulating a target number of embryonic spatial transcriptomics samples.

[0051] In one embodiment, step S1 includes:

[0052] Embryos at the target developmental stage were horizontally cut along their proximal and distal axes to obtain m-layer slices arranged in sequence.

[0053] Based on the germ layer developmental stage, each layer slice was laser-microscopically dissected to obtain n embryo samples. In one specific embodiment, spatial transcriptomics technology was used to horizontally cut the embryos at the target developmental stage into 9 sequentially arranged slices S1-S9 along the distal-proximal axis. Figure 2 As shown; based on the germ layer shape, each germ layer was cut separately to obtain 81 embryo samples;

[0054] The rules for cutting the germ layers include:

[0055] Based on the slicing rules of the first layer slice located at the farthest end of the proximal and distal axes, the first layer S1 is divided into ectoderm and endoderm segments, and each segmented segment is used as an embryo sample; wherein, the slicing rules of the first layer slice include: the anterior (Anterior, A) and posterior (Posterior, P) of the ectoderm are segmented as segments; and the entire endoderm is segmented as a segment.

[0056] Based on the segmentation rules of the second layer slice, the ectoderm, mesoderm, and endoderm regions are segmented in addition to the second layer S2, and each segmented region is treated as an embryo sample. The segmentation rules of the second layer slice include: segmenting the anterior (Posterior, P), posterior (Posterior, P), left (Left, L), and right (Right, R) regions of the ectoderm in the second layer S2; segmenting the anterior (Anterior Mesoderm, MA) and posterior (Posterior Mesoderm, MP) regions of the mesoderm in the second layer S2; and segmenting the anterior (Anterior Endoderm, EA) and posterior (Posterior Endoderm, EP) regions of the endoderm in the second layer S2.

[0057] The standard segmentation rules include: dividing the ectoderm (Posterior, P), posterior, left anterior (Left 1, L1), left posterior (Left 2, L2), right anterior (R1, R1), and right posterior (R2, R2) of layers S3-S9 (excluding the first layer S1 and the second layer S2) as segmentation regions; dividing the mesoderm (Anterior Mesoderm, MA) and posterior Mesoderm (Posterior Mesoderm, MP) of layers S3-S9 (excluding the first and second layers) as segmentation regions; and dividing the endoderm (Anterior Endoderm, EA) and posterior Endoderm (Posterior Endoderm, EP) of layers S3-S9 (excluding the first and second layers) as segmentation regions, such as... Figure 2 As shown. Following the above protocol, a total of 81 samples were obtained, and transcriptome sequencing was performed to obtain spatial transcriptome data.

[0058] Step S2: Based on the maize plot model, obtain the two-dimensional spatial expression pattern of each gene according to the spatial transcriptome data of each embryo sample;

[0059] In one embodiment, obtaining a two-dimensional spatial representation pattern using a corn plot model includes:

[0060] For each layer of laser microdissection samples, a model is built to arrange the spatial transcriptome data of each embryo sample from left to right on a plane, according to the anterior-posterior axis of the embryo. Specifically, each dot represents a dissection sample, and we arrange them from left to right on the plane according to the anterior-posterior axis of the embryo. The specific arrangement rules are as follows: the endoderm is the outermost layer in the generated corn image, corresponding to the outermost two columns on the plane, with the anterior endoderm on the far left and the posterior endoderm on the far right; the ectoderm is inside the model, and the mesoderm is distributed between the ectoderm and endoderm.

[0061] Adjust the spacing between each embryo sample to assign planar coordinates to each embryo sample; specifically, by adjusting the spacing between samples, assign planar coordinates to each cut sample so that the entire image shows the shape of an embryo.

[0062] Each embryo sample is colored according to the expression level of each gene in each sample to obtain the two-dimensional spatial expression pattern of each gene. Specifically, different colors are used to assign different expression levels to each sample, reflecting the gene expression pattern throughout the embryo. The resulting corn image is shown below. Figure 2 As shown.

[0063] Step S3: Based on the three-dimensional maize map model, obtain the three-dimensional spatial expression pattern of each gene according to the spatial transcriptome data of each embryo sample.

[0064] In one embodiment, obtaining a three-dimensional spatial representation mode using a 3D corn plot model includes:

[0065] Each embryo sample input to the model is mapped onto a target simulation model used to simulate the shape of the embryo, and each embryo sample is assigned three-dimensional spatial coordinates. Specifically, a target simulation model is first established to simulate the shape of the embryo, with the ectoderm, mesoderm, and endoderm simulated from the inside out. According to the Geo-seq strategy, the cut samples are positioned on the surface of the simulation model, and each cut sample is assigned three-dimensional spatial coordinates. Preferably, a semi-ellipsoidal model is established to simulate the shape of the embryo, with the ectoderm, mesoderm, and endoderm simulated by three concentric semi-ellipsoids from the inside out. Figure 3 The steps for establishing a three-dimensional spatial representation pattern of the ectoderm are demonstrated.

[0066] Based on the expression levels of each gene in each sample, a corresponding color is drawn for each embryo sample to obtain the three-dimensional spatial expression pattern of each gene. In the 3D Corn Plot model, the ectoderm, mesoderm, and endoderm are displayed separately.

[0067] Step S4: Based on single-cell transcriptome data and spatial transcriptome data of each embryo sample, a multidimensional single-cell mapping model is used to map each single cell onto the target developmental stage of the embryo to determine the optimal mapping position of the single cell.

[0068] In one embodiment, the method for determining the optimal mapping location of each single cell using the Multidimensional Single-Cell Mapping Model (MDSC Mapping Model) includes:

[0069] Based on the spatial expression specificity of embryonic marker genes, embryonic samples are divided into multiple spatial domains; the preferred division is nine spatial domains. Specifically, it is necessary to identify the spatial domains and marker genes (zipcode genes, zipcodes for short) of the embryos; highly variable genes (HVGs) are selected from all Geo-seq samples of embryos at the target developmental stage; secondly, principal component analysis (PCA) is performed on the selected genes; finally, the genes with the highest correlation to the principal components are selected to obtain the final set of zipcodes. Based on gene expression specificity, the zipcodes divide each embryonic sample into nine spatial domains.

[0070] The Spearman correlation coefficient between each single-cell sample and each embryo sample was calculated. Specifically, the expression profiles of zipcodes in single cells of embryos at the target developmental stage were extracted from the Gastrulation Atlas database. Then, based on this expression profile matrix, the Spearman's rank correlation coefficient (SRCC) between single cells and all embryo samples was calculated, with each single cell corresponding to 81 SRCC values ​​of embryo samples at the target developmental stage.

[0071] Based on the Spearman correlation coefficients corresponding to each single cell sample, and combined with the three-dimensional spatial coordinates of the embryo sample, the three-dimensional spatial coordinates corresponding to each single cell are calculated.

[0072] Calculate the distance between the three-dimensional spatial coordinates of the single cell and the spatial coordinates of each embryo sample, and take the three-dimensional spatial coordinates of the embryo sample with the smallest distance as the optimal mapping position of the corresponding single cell.

[0073] The three-dimensional spatial coordinates corresponding to each embryo sample are obtained in step S3 based on the three-dimensional corn map model.

[0074] In one specific embodiment, calculating the corresponding three-dimensional spatial coordinates based on the Spearman correlation coefficients for each single-cell sample includes:

[0075] Select the three Spearman correlation coefficients with the largest values ​​among all Spearman correlation coefficients for each single-cell sample;

[0076] Based on the spatial smoothing algorithm, the three-dimensional spatial coordinates of each single-cell sample are calculated based on the three largest Spearman correlation coefficients corresponding to each single-cell sample and the three-dimensional spatial coordinates of the corresponding embryo sample.

[0077] Step S5: Based on the spatial simulation model, the germ layer shape is simulated according to the optimal mapping position of each single cell to obtain a single-cell precision embryo spatial map.

[0078] In one embodiment, the method of obtaining embryonic spatial atlases with single-cell precision using a spatial simulation model includes:

[0079] The germ layer shape of the embryo at the target developmental stage is simulated to obtain a target embryo shape model;

[0080] Based on the germ layer shape and segmentation rules, a circular model was established to simulate the germ layer shape. The ectoderm, mesoderm, and endoderm were each represented by three concentric rings from the inside out, simulating embryo samples segmented by spatial transcriptomics technology. Figure 4Following the Geo-seq strategy, the circular model of the three germ layers was divided into six spaces: ectoderm was divided into A, P, L1, L2, R1, and R2; mesoderm was divided into two spaces: MA and MP; and endoderm was divided into two spaces: EA and EP. Figure 5 ).

[0081] Based on the optimal mapping position of each single cell, each single cell is located at its corresponding embryo sample location, and the cell type color is drawn for each single cell. Specifically, based on the optimal mapping position of each single cell, the MDSC Mapping model is used to locate each single cell at its respective embryo sample location, evenly distributing them within the corresponding annular spatial region. Figure 6 The color codes indicate different cell types.

[0082] Based on the bubble sort algorithm, single cells within the annular spatial region of each embryo sample are rearranged to obtain a single-cell-precision embryonic spatial map. Specifically, the bubble sort algorithm is used to spatially rearrange the cells. During the gastrulation stage of early embryonic development, different regions of the embryo exhibit different signal gradient distributions, such as the BMP signal gradient along the proximal-distal axis in the posterior region and the WNT signal gradient along the anterior-posterior axis in the proximal region. Taking the spatial location 8P of the E7.5 embryo as an example, we use the bubble sort algorithm formula (e.g., ...) according to the gradient expression of the gene Bmp4. Figure 7 As shown), rearrange the single cells in the 8P space ( Figure 8 (As shown). For the annular space corresponding to the location of each embryo sample, the cells are rearranged using the bubble sort algorithm to obtain a single-cell precision embryo spatial map. Figure 9 The spatial expression pattern of the gene Bmp4 in the ectoderm of E7.5 embryos was shown. Figure 10 Spatial maps of various cell types in the E7.5 embryo were presented at the overall level and at the distribution of the ectoderm, mesoderm, and endoderm, respectively.

[0083] Step S6: Based on the spatial optimization model, calculate the Euclidean distance to obtain the three-dimensional spatial distribution of single cells of different cell types.

[0084] In one embodiment, in developmental biology research, especially early embryonic development, cells located in different regions of the embryo respond to different signal stimuli, resulting in specific cell developmental fates. To more clearly illustrate the cell type distribution in each annular space, we designed a spatial optimization model based on Euclidean distance;

[0085] Methods for obtaining the three-dimensional spatial distribution of single cells of different cell types using spatial optimization models include:

[0086] Single-cell gene expression profiles were extracted from each annular spatial region of the embryonic spatial map at single-cell precision and logarithmically transformed; taking the E6.75 embryo at position 6P as an example, such as Figure 11 As shown in part i in the diagram.

[0087] Based on the single-cell gene expression profiles within each annular spatial region after logarithmic transformation, the Euclidean distance between every two cells within each annular spatial region is calculated. This distance is defined as the "transcriptome distance" between cells, and an Euclidean distance matrix (EDM) is constructed. Taking the E6.75 embryo at position 6P as an example, ... Figure 11 As shown in section ii.

[0088] The farthest transcriptome distance between single cells and the maximum distance within the annular spatial region were extracted within each Euclidean distance matrix. The Euclidean distance matrix was then standardized to obtain a standardized Euclidean distance matrix. Taking the E6.75 embryo at position 6P as an example, the standardization process was performed as follows: Figure 10 The formula shown in part iii in the figure yields the standardized Euclidean distance matrix.

[0089] Using the spatial region functions of each annulus as constraints, the least squares method is applied to simulate the optimal coordinates of each cell based on the corresponding standardized Euclidean distance matrix, thus obtaining the three-dimensional spatial distribution of single cells of different cell types. Taking the E6.75 embryo at position 6P as an example... Figure 11 As shown in section iv.

[0090] To better describe the three-dimensional construction method of single-cell spatial transcriptome, the following specific embodiments are provided for illustration;

[0091] Example 1: A method for constructing a three-dimensional spatial transcriptome in a single cell.

[0092] Taking the three-dimensional construction of a single-cell spatial transcriptome of an E7.5 mouse embryo during the gastrulation stage of early embryonic development as an example.

[0093] The method includes:

[0094] 1. Using Geo-seq, E7.5 embryos were spatially segmented to obtain 81 Geo-seq samples;

[0095] Among them, the E7.5 embryos were horizontally cut into 9 layers along the distal–proximal axis. For each slice containing the ectoderm, mesoderm, and endoderm, the specific cutting scheme is as follows: The ectoderm is divided into anterior (A), posterior (P), left anterior (L1), right anterior (R1), left posterior (L2), and right posterior (R2), for a total of 6 samples (due to cell number limitations, only the anterior and posterior samples were cut in the first slice, and the left and right sides of the second slice were not divided into anterior and posterior samples, and were cut into 4 samples); The mesoderm is divided into anterior (MA) and posterior (MP) samples (the first slice contains only ectoderm and endoderm cells, and mesoderm samples are collected starting from the second slice); The endoderm is divided into anterior (EA) and posterior (EP) samples (due to cell number limitations, the first slice is not divided into anterior and posterior, and the entire endoderm is collected). Following the Geo-seq protocol described above, a total of 81 samples were obtained and their transcriptomes were sequenced to obtain spatial transcriptome data.

[0096] 2. Based on the corn plot model, the two-dimensional spatial expression patterns of each gene are obtained from the spatial transcriptome data of each embryo sample.

[0097] In this model, each dot represents a cut sample. They are arranged from left to right on a plane according to the anterior-posterior axis of the embryo, as follows: the endoderm is the outermost layer of the Corn Plot model, corresponding to the outermost two columns on the plane, with the anterior endoderm on the far left and the posterior endoderm on the far right; the ectoderm is inside the model, and the mesoderm is distributed between the ectoderm and endoderm. Then, by adjusting the spacing between the samples, each cut sample is assigned a planar coordinate, making the entire Corn Plot model resemble the shape of an embryo. In the Corn Plot model, each sample is colored differently based on the gene expression levels in different samples, reflecting the gene expression pattern throughout the embryo. Figure 12 The results of in situ hybridization experiments of the primitive streak marker gene T in E7.5 embryos are presented, as well as the spatial expression pattern of gene T shown by Corn Plot.

[0098] 3. Based on the 3D Corn Plot Model, the three-dimensional spatial expression patterns of each gene are obtained according to the spatial transcriptome data of each embryo sample.

[0099] First, a semi-ellipsoidal model was established to simulate the shape of the embryo. The ectoderm, mesoderm, and endoderm were each simulated by three concentric semi-ellipsoids from the inside out. Following a Geo-seq strategy, the cut samples were positioned on the surface of the semi-ellipsoidal model, and each cut sample was assigned three-dimensional spatial coordinates. In the 3D Corn Plot model, the ectoderm, mesoderm, and endoderm were displayed separately. Different colors were used to represent the gene expression patterns throughout the embryo, reflecting the gene expression patterns across different samples. Figure 13 The expression pattern of the posterior original streak marker gene T in the embryo is shown at E7.5.

[0100] 4. Based on each single-cell sample and each embryo sample, the Multi-Dimensional Single-Cell Mapping Model (MDSC Mapping Model) is used to map each single cell onto the embryo at the target developmental stage in order to determine the optimal mapping position of each single cell.

[0101] By combining single-cell transcriptomics technology and a 3D Corn Plot model, precise spatial localization of single cells can be achieved.

[0102] (1) Identification of spatial domains and marker genes (zipcodes for short) in embryos: First, highly variable genes (HVGs) were selected from all Geo-seq samples of E7.5 embryos, with the top 6,000 HVGs chosen. Second, principal component analysis (PCA) was performed on these 6,000 HVGs. Finally, the 50 genes with the highest correlation were selected from the five principal components PC1–PC5 (considering both positive and negative correlations, the top positive and negative PCloading genes), resulting in 10 sets of zipcodes containing a total of 490 genes (some genes had overlapping dimensions, so the total number of genes in the zipcodes set was not equal to 500) (e.g. Figure 14 (As shown). Based on gene expression specificity, zipcodes divide E7.5 embryos into 9 spatial domains (e.g., ...). Figure 15 (As shown).

[0103] (2) Calculate the correlation between each single-cell sample and the Geo-seq embryo reference sample: First, extract the expression profile of zipcodes for single cells of E7.5 embryos in the Gastrulation Atlas database; second, calculate the Spearman's rank correlation coefficient (SRCC) between single cells and all Geo-seq embryo samples based on this expression profile matrix, with each single cell corresponding to 81 SRCC values ​​of E7.5 embryos.

[0104] (3) Applying a spatial smoothing algorithm to determine the high-confidence locations of each cell: First, extracting the top three matching locations with the highest SRCC values; second, applying... Figure 16 The formula calculates the three-dimensional spatial coordinates and the distance between these coordinates and the location of each Geo-seq sample. Finally, the Geo-seq sample with the smallest distance is determined as the optimal mapping location for the cell.

[0105] (4) Evaluation of the accuracy of the MDSC Mapping model: We took a certain number of single cells from known locations in E7.5 embryos and performed transcriptome sequencing. Then, we applied the MDSC Mapping model to map these single cells onto the embryos and performed correlation analysis between the mapped locations and their actual locations (calculating the Pearson Correlation Coefficient, PCC) (e.g.) Figure 17 The results show that the MDSC Mapping model has very high spatial mapping accuracy (PCC = 0.9738).

[0106] 5. Establish an Annulus Model to simulate the shape of the germ layers, and obtain a single-cell precision embryonic spatial map based on the optimal mapping position of each single cell.

[0107] After obtaining the spatial location of single cells using the MDSC Mapping model, we designed a circular model to represent the spatial distribution of single cells. Taking E7.5 embryo structures and single cells from E7.5 embryos in the Gastrulation Atlas database as examples, the model operation process is as follows:

[0108] (1) A circular model was established to simulate the shape of the germ layers. The ectoderm, mesoderm, and endoderm were each simulated by three concentric circular rings from the inside out. Following the Geo-seq strategy, the circular models of the three germ layers were divided into six spaces: A, P, L1, L2, R1, and R2 for the ectoderm; MA and MP for the mesoderm; and EA and EP for the endoderm. Cells at each Geo-seq location were located using the MDSC Mapping model and were evenly distributed within their corresponding circular spaces.

[0109] (2) Applying the MDSC Mapping model, single cells of E7.5 embryos (e.g., from the Gastrulation Atlas database) were mapped. Figure 18 Locate the Geo-seq position. Figure 19 The spatial distribution of representative cell types (Nascent mesoderm, Notochord, Visceral endoderm, Haematoendothelial progenitors) of E7.5 embryos is shown, with single cells evenly distributed within a circular space.

[0110] (3) Spatial rearrangement of cells using the bubble sort algorithm. During the gastrulation period of early embryonic development, different regions of the embryo exhibit different signal gradient distributions, such as the BMP signal gradient along the proximal-distal axis of the posterior region and the WNT signal gradient along the anterior-posterior axis of the proximal region. Taking the spatial location 8P of the E7.5 embryo as an example, we rearrange the single cells in the 8P space according to the gradient expression of the gene Bmp4 using the bubble sort algorithm.

[0111] (4) For the annular space corresponding to each Geo-seq position, the cells are rearranged using the bubble sort algorithm to obtain a single-cell precision embryonic spatial map. The spatial expression pattern of gene Bmp4 in the ectoderm of E7.5 embryos is shown, and the spatial maps of various cell types in E7.5 embryos are presented from the overall level and the distribution of the three germ layers (ectoderm, mesoderm, and endoderm).

[0112] 6. Based on the Spatial Optimization Model, the Euclidean distance is calculated to obtain the three-dimensional spatial distribution of single cells of different cell types by refining the embryonic spatial map to each annular space with single-cell precision.

[0113] (1) Extract the gene-expression matrix (GEM) of a single cell within the circular segmentation space and perform log2 normalization;

[0114] (2) Based on the logarithmically transformed gene expression profile, calculate the Euclidean distance between every two cells, define this distance as the “transcriptome distance” between cells, and establish the Euclidean distance matrix (EDM);

[0115] (3) Extract the farthest transcriptome distance (d) between cells within the Euclidean distance matrix. max The maximum distance within the annular space (length) max The Euclidean distance matrix is ​​standardized according to the formula.

[0116] (4) Using the circular space function as a constraint, the least squares method is applied to simulate and obtain the optimal coordinates of each cell.

[0117] Taking a single cell at position 6P of embryo E6.75 as an example, t-SNE (t-distributed stochastic neighbor embedding) analysis was first used to show the clustering results of various cell types at this position (e.g., Figure 20 Secondly, a spatial optimization model is applied to simulate the spatial coordinates of a single cell within position 6P of the E6.75 embryo, and the spatial distribution of the single cell is displayed in three dimensions (e.g., Figure 21 ).

[0118] This embodiment utilizes the spatial positionsequation (Geo-seq) technology established in our laboratory, which enables the study of spatiotemporal dynamic molecular expression patterns of early embryonic development at the whole-genome level, based on the establishment of different germ layer patterns. 1-3 Based on Geo-seq technology, we captured location-specific spatial transcriptome data at five time points (E6.5, E6.75, E7.0, E7.25, and E7.5 days) during the gastrulation phase of early mouse embryonic development, at six-hour intervals. By establishing a mathematical model, we reconstructed three-dimensional spatial transcriptome maps of the ectoderm, mesoderm, and endoderm, creating an encyclopedic spatiotemporal expression database of the entire genome (“scGastrulation” website, not yet publicly available). This database achieves high-resolution digital in situ hybridization maps of the entire genome at single-cell precision during the gastrulation phase of mouse embryonic development, and is currently the most comprehensive, complete, and accurate interactive spatiotemporal transcriptome database internationally.

[0119] Our mathematical model is powerful. First, it identifies a set of marker genes that represent the developmental state and specific spatial domains of different stages of gastrulation. These genes have a "zipcode"-like marking function, which can "restore" embryonic tissue cells or exogenous pluripotent stem cells to specific locations in the embryo, thereby locating cells of unknown origin to the corresponding parts of the embryo.

[0120] Secondly, the model incorporates a spatial single-cell algorithm that simulates the spatial distribution of individual cells across different germ layers, further reconstructing a spatial transcriptome atlas with single-cell precision in three dimensions. This work provides comprehensive data and novel insights into understanding germ layer lineage establishment and the fate regulation mechanisms of pluripotent stem cells. It will become an important reference model for classical developmental biology hierarchical lineage theory, while also promoting the development of early embryonic development and stem cell regenerative medicine.

[0121] Similar to the principles of the above embodiments, the present invention provides a three-dimensional construction system for a single-cell spatial transcriptome.

[0122] The following specific embodiments are provided in conjunction with the accompanying drawings:

[0123] like Figure 22 This diagram illustrates the structure of a three-dimensional construction system for a single-cell spatial transcriptome according to an embodiment of the present invention.

[0124] The system includes:

[0125] The segmentation module 221 is used to segment embryos at a target developmental stage using spatial transcriptomics technology, capture location-specific spatial transcriptomics data, and accumulate a target number of embryonic spatial transcriptomics samples.

[0126] The two-dimensional spatial expression module 222, connected to the segmentation module 221, is used to obtain the two-dimensional spatial expression pattern of each gene based on the maize diagram model and the spatial transcriptome data of each embryo sample.

[0127] The three-dimensional spatial expression module 223, connected to the segmentation module 221, is used to obtain the three-dimensional spatial expression pattern of each gene based on the spatial transcriptome data of each embryo sample and the three-dimensional maize map model.

[0128] The multidimensional single-cell mapping module 224 connects the segmentation module 221 and the three-dimensional spatial expression module 223. It is used to map each single cell onto the target developmental stage embryo based on single-cell transcriptome data and spatial transcriptome data of each embryo sample, using a multidimensional single-cell mapping model to determine the optimal mapping position of the single cell.

[0129] The spatial simulation module 225 connects the multidimensional single-cell mapping module 224 and the three-dimensional spatial expression module 223. It is used to simulate the shape of different germ layers of early embryos based on the spatial simulation model, and further simulate and obtain a single-cell precision embryonic spatial map according to the optimal mapping position of the single cell.

[0130] The spatial optimization module 226 is connected to the spatial simulation module 225 and is used to calculate the Euclidean distance between single cells based on the spatial optimization model to simulate the three-dimensional spatial distribution of single cells.

[0131] It should be noted that, as should be understood Figure 22 The division of modules in the system embodiment is merely a logical functional division. In actual implementation, they can be fully or partially integrated into a single physical entity, or they can be physically separated. Furthermore, these units can be implemented entirely in software through processing element calls; they can be implemented entirely in hardware; or some units can be implemented by processing element calls to software, while others are implemented in hardware.

[0132] Since the implementation principle of the three-dimensional construction system of the single-cell spatial transcriptome has been described in the previous embodiments, it will not be repeated here.

[0133] In one embodiment, the segmentation module 221 is used to horizontally cut the embryo at the target developmental stage according to the embryo's proximal and distal axes to obtain m layers of slices arranged in sequence; based on the developmental state of the germ layers, each slice is laser-microscopically cut to obtain n embryo samples; and transcriptome sequencing is performed on each embryo sample to obtain spatial transcriptome data corresponding to each embryo sample; wherein, laser-microscopically cutting each slice based on the developmental state of the germ layers includes: based on the segmentation rules of the first slice located at the farthest end of the proximal and distal axes, laser-microscopically cutting the ectoderm and endoderm according to the developmental state of the first slice, with each cut area as an embryo sample; The segmentation rules for the first layer of slicing include: cutting the ectoderm and posterior end as segmentation regions; cutting the entire endoderm as segmentation regions; and, based on the segmentation rules for the second layer of slicing, performing laser microsurgery on the ectoderm, mesoderm, and endoderm according to the developmental state of the second layer of slicing, with each segmented region serving as an embryo sample. The segmentation rules for the second layer of slicing include: segmenting the ectoderm and posterior end, left and right sides as segmentation regions; segmenting the mesoderm and posterior end as segmentation regions; and segmenting the endoderm and posterior end as segmentation regions.

[0134] Based on standard segmentation rules, laser microscopy is used to cut the ectoderm, mesoderm, and endoderm according to the developmental state of the layers other than the first and second layers, with each cut region serving as an embryo sample. The standard segmentation rules include: cutting the ectoderm into segments at the anterior, posterior, left anterior, left posterior, right anterior, and right posterior ends; cutting the mesoderm into segments at the anterior and posterior ends; and cutting the endoderm into segments at the anterior and posterior ends.

[0135] In one embodiment, the two-dimensional spatial expression module 222 is used to arrange each embryo sample from left to right on a plane according to the anterior-posterior axis of the embryo for each layer of laser micro-cutting; adjust the spacing between each embryo sample to assign planar coordinates to each embryo sample; and draw the corresponding color for each embryo sample based on the expression value of each gene in each sample to obtain the two-dimensional spatial expression pattern of each gene.

[0136] In one embodiment, the three-dimensional spatial expression module 223 is used to map each embryo sample onto a three-dimensional model for simulating the shape of the embryo, and assign three-dimensional spatial coordinates to each embryo sample; based on the expression value of each gene in each sample, the module draws a corresponding color for each embryo sample to obtain the three-dimensional spatial expression pattern of each gene.

[0137] In one embodiment, the multidimensional single-cell mapping module 224 is used to extract spatially specific marker genes based on single-cell transcriptome data, wherein the marker genes divide the target embryo into multiple spatial domains; calculate the Spearman correlation coefficient between each single cell and the embryo sample; calculate the three-dimensional spatial coordinates corresponding to each single cell based on the Spearman correlation coefficient corresponding to each single cell sample and the three-dimensional spatial coordinates of the embryo sample; calculate the distance between the corresponding three-dimensional spatial coordinates and the three-dimensional spatial coordinates of each embryo sample, and take the three-dimensional spatial coordinates of the embryo sample with the smallest corresponding distance as the optimal mapping position of the corresponding single cell.

[0138] In one embodiment, the step of calculating the corresponding three-dimensional spatial coordinates based on the Spearman correlation coefficients corresponding to each single-cell sample includes: selecting the three largest Spearman correlation coefficients among the Spearman correlation coefficients corresponding to each single-cell sample; combining the three-dimensional spatial coordinates of the embryo sample, applying a spatial smoothing algorithm, and calculating the three-dimensional spatial coordinates of each single cell based on the three largest Spearman correlation coefficients corresponding to each single cell and the three-dimensional spatial coordinates of the corresponding embryo sample.

[0139] In one embodiment, the spatial simulation module 225 is used to simulate the germ layer shape of an embryo at a target developmental stage to obtain a target embryo shape model; to establish a circular model and segment the circular space to simulate the embryo samples segmented by spatial transcriptomics technology; based on the optimal mapping position of a single cell, the single cell is located at the corresponding embryo sample position, and the single cell is drawn with a corresponding color according to the cell type. At the same time, according to the expression level of each gene in the single cell, a corresponding color is drawn for each single cell; based on the gene expression of the single cell, the single cells in the circular space region where each embryo sample is located are spatially rearranged using a bubble algorithm to obtain a single-cell precision embryo spatial map.

[0140] In one embodiment, the spatial optimization module 226 is used to extract single-cell gene expression profiles from each annular spatial region in the embryo spatial map with single-cell precision and perform logarithmic transformation; based on the logarithmically transformed single-cell gene expression profiles from each annular spatial region, the Euclidean distance between every two cells in each annular spatial region is calculated, and a corresponding Euclidean distance matrix is ​​established; the farthest transcriptome distance between single cells and the maximum distance within each annular spatial region are extracted from each Euclidean distance matrix, and the Euclidean distance matrix is ​​standardized to obtain a standardized Euclidean distance matrix; using the functions of each annular spatial region as constraints, the least squares method is applied to simulate the optimal coordinates of each cell based on the corresponding standardized Euclidean distance matrix to obtain the three-dimensional spatial distribution of single cells.

[0141] like Figure 23 A schematic diagram of the structure of the three-dimensional construction terminal 10 of the single-cell spatial transcriptome in an embodiment of the present invention is shown.

[0142] The three-dimensional construction terminal 100 for the single-cell spatial transcriptome includes: a memory 101 and a processor 102. The memory 101 is used to store computer programs; the processor 102 runs the computer programs to implement, for example... Figure 1 The method for constructing a three-dimensional single-cell spatial transcriptome.

[0143] Optionally, the number of memories 101 can be one or more, and the number of processors 102 can be one or more. Figure 10 Each example is taken as an instance.

[0144] Optionally, the processor 102 in the three-dimensional construction terminal 100 of the single-cell spatial transcriptome will proceed as follows: Figure 1 The steps described involve loading one or more instructions corresponding to the process of an application into memory 101, and having the processor 102 run the application stored in the first memory 101, thereby achieving the following: Figure 1 Various functions in the three-dimensional construction method of the single-cell spatial transcriptome.

[0145] Optionally, the memory 101 may include, but is not limited to, high-speed random access memory and non-volatile memory. For example, one or more disk storage devices, flash memory devices, or other non-volatile solid-state storage devices; the processor 102 may include, but is not limited to, a central processing unit (CPU), a network processor (NP), etc.; it may also be a digital signal processor (DSP), an application-specific integrated circuit (ASIC), a field-programmable gate array (FPGA), or other programmable logic devices, discrete gate or transistor logic devices, or discrete hardware components.

[0146] Optionally, the processor 102 may be a general-purpose processor, including a central processing unit (CPU), a network processor (NP), etc.; it may also be a digital signal processor (DSP), an application-specific integrated circuit (ASIC), a field-programmable gate array (FPGA), or other programmable logic devices, discrete gate or transistor logic devices, or discrete hardware components.

[0147] The present invention also provides a computer-readable storage medium storing a computer program, wherein the computer program, when executed, implements as follows: Figure 1 The method for constructing a three-dimensional spatial transcriptome of a single cell is shown. The computer-readable storage medium may include, but is not limited to, floppy disks, optical disks, CD-ROMs (Read-Only Optical Disk Memory), magneto-optical disks, ROMs (Read-Only Memory), RAMs (Random Access Memory), EPROMs (Erasable Programmable Read-Only Memory), EEPROMs (Electrically Erasable Programmable Read-Only Memory), magnetic cards or optical cards, flash memory, or other types of media / machine-readable media suitable for storing machine-executable instructions. The computer-readable storage medium may be a product not connected to a computer device or a component used in a computer device.

[0148] In summary, the single-cell spatial transcriptome 3D construction method, system, and terminal of this invention captures location-specific spatial transcriptome data based on spatial transcriptome technology and, based on an established mathematical model, reconstructs 3D spatial transcriptome maps of the ecto, mes, and endoderm layers, establishing an encyclopedic whole-genome spatiotemporal expression database. The spatial transcriptome map constructed by this invention achieves high-resolution digital in situ hybridization maps of the entire genome with single-cell precision, making it the most comprehensive, complete, and accurate interactive spatiotemporal transcriptome database currently available. Therefore, this invention effectively overcomes the various shortcomings of existing technologies and has high industrial application value.

[0149] The above embodiments are merely illustrative of the principles and effects of the present invention and are not intended to limit the invention. Any person skilled in the art can modify or alter the above embodiments without departing from the spirit and scope of the present invention. Therefore, all equivalent modifications or alterations made by those skilled in the art without departing from the spirit and technical concept disclosed in the present invention should still be covered by the claims of the present invention.

Claims

1. A method for constructing a three-dimensional spatial transcriptome in a single cell, characterized in that, The method includes: Using spatial transcriptomics technology, embryos at the target developmental stage are segmented to capture location-specific spatial transcriptomics data, and the target number of embryonic spatial transcriptomics samples are accumulated. Based on the maize plot model, the two-dimensional spatial expression patterns of each gene were obtained from the spatial transcriptome data of each embryo sample. Based on the three-dimensional maize map model, the three-dimensional spatial expression pattern of each gene was obtained according to the spatial transcriptome data of each embryo sample. Based on single-cell transcriptome data and spatial transcriptome data of each embryo sample, a multidimensional single-cell mapping model is used to map each single cell onto the embryo at the target developmental stage to determine the optimal mapping position of the single cell. Based on the spatial simulation model, the shape of different germ layers of early embryos is simulated, and according to the optimal mapping position of a single cell, further simulation is performed to obtain a single-cell precision embryonic spatial map. Based on the spatial optimization model, the Euclidean distance between single cells is calculated to simulate the three-dimensional spatial distribution of single cells; The process of obtaining the three-dimensional spatial expression pattern of each gene based on the spatial transcriptome data of each embryo sample specifically includes: Each embryo sample is mapped onto a 3D model used to simulate the shape of the embryo, and each embryo sample is assigned 3D spatial coordinates; based on the expression values ​​of each gene in each sample, the corresponding color is drawn for each embryo sample to obtain the 3D spatial expression pattern of each gene. The method of mapping each single cell onto a target developmental stage embryo using a multidimensional single-cell mapping model, based on single-cell transcriptome data and spatial transcriptome data of each embryo sample, to determine the optimal mapping position for a single cell includes: extracting spatially specific marker genes based on single-cell transcriptome data, wherein the marker genes divide the target embryo into multiple spatial domains; calculating the Spearman correlation coefficient between each single cell and the embryo sample; calculating the three-dimensional spatial coordinates of each single cell based on the Spearman correlation coefficient corresponding to each single cell sample and the three-dimensional spatial coordinates of the embryo sample; calculating the distance between the corresponding three-dimensional spatial coordinates and the three-dimensional spatial coordinates of each embryo sample, and taking the three-dimensional spatial coordinates of the embryo sample with the smallest corresponding distance as the optimal mapping position for the corresponding single cell; The calculation of the corresponding three-dimensional spatial coordinates based on the Spearman correlation coefficient corresponding to each single-cell sample and the spatial coordinates of the embryo sample includes: selecting the three largest Spearman correlation coefficients among the Spearman correlation coefficients corresponding to each single-cell sample; and, in combination with the three-dimensional spatial coordinates of the embryo sample, applying a spatial smoothing algorithm to calculate the three-dimensional spatial coordinates of each single cell based on the three largest Spearman correlation coefficients corresponding to each single cell and the corresponding three-dimensional spatial coordinates of the embryo sample.

2. The method for constructing a three-dimensional single-cell spatial transcriptome according to claim 1, characterized in that, The method utilizes spatial transcriptomics technology to segment embryos at the target developmental stage, capture location-specific spatial transcriptomics data, and accumulate a target number of embryonic spatial transcriptomics samples, including: Embryos at the target developmental stage were horizontally cut along their proximal and distal axes to obtain m-layer slices arranged in sequence. Based on the developmental state of the germ layers, each layer slice was laser micro-dissected to obtain n embryo samples. Transcriptome sequencing was performed on each embryo sample to obtain spatial transcriptome data for each embryo sample. The process of laser microdissection of each germ layer slice based on its developmental state includes: Based on the segmentation rules of the first layer slice located at the farthest end of the proximal and distal axes, laser microscopy is performed on the ectoderm and endoderm according to the developmental state of the first layer slice, with each segmented area serving as an embryo sample; wherein, the segmentation rules of the first layer slice include: segmenting the anterior and posterior ends of the ectoderm as segmentation areas; and segmenting the entire endoderm as segmentation areas. Based on the segmentation rules of the second-layer slice, laser microscopy is used to cut the ectoderm, mesoderm, and endoderm according to their developmental state, with each cut area representing an embryo sample. The segmentation rules of the second-layer slice include: segmenting the ectoderm into the anterior, posterior, left, and right sides; segmenting the mesoderm into the anterior and posterior sides; and segmenting the endoderm into the anterior and posterior sides. Based on standard segmentation rules, laser microscopy is used to cut the ectoderm, mesoderm, and endoderm according to the developmental state of the layers other than the first and second layers, with each cut region serving as an embryo sample. The standard segmentation rules include: cutting the ectoderm into segments at the anterior, posterior, left anterior, left posterior, right anterior, and right posterior ends; cutting the mesoderm into segments at the anterior and posterior ends; and cutting the endoderm into segments at the anterior and posterior ends.

3. The method for constructing a three-dimensional single-cell spatial transcriptome according to claim 1, characterized in that, The process of obtaining the two-dimensional spatial expression patterns of each gene based on the spatial transcriptome data of each embryo sample includes: For each layer of laser microdissection samples, the embryo samples are arranged from left to right on the plane according to the anterior-posterior axis of the embryo. Adjust the spacing between each embryo sample and assign planar coordinates to each embryo sample; Based on the expression levels of each gene in each sample, a corresponding color is drawn for each embryo sample to obtain the two-dimensional spatial expression pattern of each gene.

4. The method for constructing a three-dimensional spatial transcriptome of a single cell according to claim 1, characterized in that, The aforementioned spatial simulation model simulates the shape of different germ layers in early embryos, and further simulates and obtains embryonic spatial atlases with single-cell precision based on the optimal mapping position of a single cell, including: The germ layer shape of the embryo at the target developmental stage is simulated to obtain a model of the target embryo shape; A circular model was established and the circular space was segmented to simulate embryo samples segmented by spatial transcriptomics technology; Based on the optimal mapping location of a single cell, the single cell is located at the corresponding embryo sample location. According to the cell type, the single cell is drawn with the corresponding color. At the same time, according to the expression level of each gene in the single cell, the corresponding color is drawn for each single cell. Based on gene expression in single cells, the bubble sort algorithm is used to spatially rearrange the single cells in the annular spatial region where each embryo sample is located, thereby obtaining an embryo spatial map with single-cell precision.

5. The method for constructing a three-dimensional single-cell spatial transcriptome according to claim 4, characterized in that, The calculation of Euclidean distances between single cells based on the spatial optimization model to simulate the three-dimensional spatial distribution of single cells includes: Extract single-cell gene expression profiles from each annular spatial region in the embryo spatial atlas at single-cell precision and perform logarithmic transformation; Based on the single-cell gene expression profiles in each annular spatial region after logarithmic transformation, the Euclidean distance between every two cells in each annular spatial region is calculated, and the corresponding Euclidean distance matrix is ​​established. The farthest transcriptome distance between single cells and the maximum distance within the annular space region are extracted within each Euclidean distance matrix. The Euclidean distance matrix is ​​then standardized to obtain a standardized Euclidean distance matrix. Using the functions of each annular spatial region as constraints, the least squares method is applied to simulate the optimal coordinates of each cell based on the corresponding standardized Euclidean distance matrix, so as to obtain the three-dimensional spatial distribution of a single cell.

6. A three-dimensional construction system for a single-cell spatial transcriptome, characterized in that, The system includes: The segmentation module is used to segment embryos at a target developmental stage using spatial transcriptomics technology, capture location-specific spatial transcriptomics data, and accumulate a target number of embryonic spatial transcriptomics samples. A two-dimensional spatial expression module, connected to the segmentation module, is used to obtain the two-dimensional spatial expression pattern of each gene based on the spatial transcriptome data of each embryo sample. A three-dimensional spatial expression module, connected to the segmentation module, is used to obtain the three-dimensional spatial expression pattern of each gene based on the spatial transcriptome data of each embryo sample and the three-dimensional maize map model. The multidimensional single-cell mapping module connects the segmentation module and the three-dimensional spatial expression module. It is used to map each single cell onto the target developmental stage embryo based on single-cell transcriptome data and spatial transcriptome data of each embryo sample, using the multidimensional single-cell mapping model to determine the optimal mapping position of the single cell. The spatial simulation module connects the multidimensional single-cell mapping module and the three-dimensional spatial expression module. It is used to simulate the shape of different germ layers of early embryos based on the spatial simulation model, and further simulate and obtain a single-cell precision embryonic spatial map according to the optimal mapping position of the single cell. The spatial optimization module, connected to the spatial simulation module, is used to calculate the Euclidean distance between single cells based on the spatial optimization model, so as to simulate the three-dimensional spatial distribution of single cells. The process of obtaining the three-dimensional spatial expression pattern of each gene based on the spatial transcriptome data of each embryo sample specifically includes: Each embryo sample is mapped onto a 3D model used to simulate the shape of the embryo, and each embryo sample is assigned 3D spatial coordinates; based on the expression values ​​of each gene in each sample, the corresponding color is drawn for each embryo sample to obtain the 3D spatial expression pattern of each gene. The method of mapping each single cell onto a target developmental stage embryo using a multidimensional single-cell mapping model, based on single-cell transcriptome data and spatial transcriptome data of each embryo sample, to determine the optimal mapping position for a single cell includes: extracting spatially specific marker genes based on single-cell transcriptome data, wherein the marker genes divide the target embryo into multiple spatial domains; calculating the Spearman correlation coefficient between each single cell and the embryo sample; calculating the three-dimensional spatial coordinates of each single cell based on the Spearman correlation coefficient corresponding to each single cell sample and the three-dimensional spatial coordinates of the embryo sample; calculating the distance between the corresponding three-dimensional spatial coordinates and the three-dimensional spatial coordinates of each embryo sample, and taking the three-dimensional spatial coordinates of the embryo sample with the smallest corresponding distance as the optimal mapping position for the corresponding single cell; The calculation of the corresponding three-dimensional spatial coordinates based on the Spearman correlation coefficient corresponding to each single-cell sample and the spatial coordinates of the embryo sample includes: selecting the three largest Spearman correlation coefficients among the Spearman correlation coefficients corresponding to each single-cell sample; and, in combination with the three-dimensional spatial coordinates of the embryo sample, applying a spatial smoothing algorithm to calculate the three-dimensional spatial coordinates of each single cell based on the three largest Spearman correlation coefficients corresponding to each single cell and the corresponding three-dimensional spatial coordinates of the embryo sample.

7. A three-dimensional construction terminal for a single-cell spatial transcriptome, characterized in that, include: One or more memories and one or more processors; The one or more memories are used to store computer programs; The one or more processors are connected to the memory and are used to run the computer program to perform the method as described in any one of claims 1 to 5.

Images