Machine learning based method for correlating gene expression and cellular microphenotypes

Abstract

The application provides a method for correlating gene expression and cell microphenotype based on machine learning, wherein the method comprises: obtaining cell imaging data and spatial transcriptome data, wherein the cell imaging data comprises a nucleus image of a target cell; performing single cell segmentation based on the nucleus image and the spatial transcriptome data by fine-tuning a large model in the field of vision to obtain a cell segmentation image of the target cell; determining a cell mask of the target cell based on the cell segmentation image; quantifying the cell mask to obtain a microphenotype of the target cell; determining gene expression of the target cell based on the cell mask of the target cell and the spatial transcriptome data; and performing correlation analysis on the microphenotype of the target cell and the gene expression of the target cell based on a preset machine learning model to obtain a correlation result of the microphenotype and the gene expression of the target cell.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of cell biology technology, and in particular to a method for correlating gene expression and cell microphenotype based on machine learning. BACKGROUND

[0002] A series of subtle traits exhibited by plants, whether physical, chemical or biological properties, are the result of the expression of specific genes at specific times, collectively known as "microphenotype".

[0003] In the past 20 years, there has been great progress in macro-level phenotypic traits (organs, plants and canopies), but little attention has been paid to micro-level phenotypic traits and their correlation with macro-level phenotypic traits. At the same time, due to the fact that gene expression data is usually very complex and contains a large amount of information, it is difficult to interpret and analyze the data, which requires professional bioinformatics knowledge and skills.

[0004] Therefore, the trait phenotype identification method in the related art has the technical problem of difficult data analysis. SUMMARY

[0005] The present application provides a method for correlating gene expression and cell microphenotype based on machine learning, which solves the problem of difficult data analysis in the prior art trait phenotype identification method and realizes the correlation of single cell microphenotype and gene expression.

[0006] The present application provides a method for correlating gene expression and cell microphenotype based on machine learning, comprising the following steps.

[0007] Obtain cell imaging data and spatial transcriptome data, wherein the cell imaging data includes a nucleus image of a target cell; perform single cell segmentation based on the nucleus image and the spatial transcriptome data by fine-tuning a large model in the field of vision to obtain a cell segmentation image of the target cell; determine a cell mask of the target cell based on the cell segmentation image; quantify the cell mask of the target cell to obtain a microphenotype of the target cell; determine gene expression of the target cell based on the cell mask of the target cell and the spatial transcriptome data; and perform correlation analysis based on the microphenotype of the target cell and the gene expression of the target cell by a pre-set machine learning model to obtain a microphenotype and gene expression correlation result of the target cell.

[0008] According to the present invention, a method for associating gene expression and cellular microphenotype based on machine learning, wherein the step of performing single-cell segmentation based on the cell imaging data and the spatial transcriptome data by fine-tuning a large visual domain model to obtain a cell image of the target cell includes: determining the complete image data of the target cell on the cell imaging data based on the cell nuclear image and the spatial transcriptome data; performing single-cell segmentation based on the complete image data by fine-tuning a large visual domain model to obtain cell boundary localization; and determining the cell segmentation image of the target cell based on the cell boundary localization.

[0009] According to the present invention, a method for associating gene expression and cell microphenotype based on machine learning is provided, wherein the microphenotype of the target cell includes at least one of the following: cell area, cell perimeter, radius of the smallest circumscribed circle of the cell, major axis of the smallest circumscribed ellipse of the cell, minor axis of the smallest circumscribed ellipse of the cell, length of the smallest circumscribed rectangle of the cell, and width of the smallest circumscribed rectangle of the cell.

[0010] According to a method for associating gene expression and cellular microphenotype based on machine learning provided by the present invention, the method for determining the gene expression of the target cell based on the cell mask of the target cell and the spatial transcriptome data includes: determining each cell mask position of the target cell; aligning and integrating each cell mask position with the spatial transcriptome data to obtain transcriptome data corresponding to each cell mask position; using the transcriptome data corresponding to each cell mask position as the gene expression of each cell mask; and determining the gene expression of the target cell based on the gene expression of each cell mask of the target cell.

[0011] According to a machine learning-based method for associating gene expression and cellular microphenotype provided by the present invention, after performing association analysis on the microphenotype and gene expression of the target cell using a preset machine learning model to obtain the association result between the microphenotype and gene expression of the target cell, the method further includes: determining the target microphenotype of the target cell; and based on the association result between the microphenotype and gene expression, determining gene expressions with an association degree greater than a preset association degree threshold as candidate gene expressions, wherein the candidate gene expressions are the target gene expressions of the target microphenotype.

[0012] According to the method for associating gene expression and cell microphenotype based on machine learning provided in the application, after the preset machine learning model is used to perform association analysis on the microphenotype of the target cell and the gene expression of the target cell, and the association result of the microphenotype and the gene expression of the target cell is obtained, the method further comprises: obtaining single-cell gene expression input by a user; and predicting the single-cell gene expression based on the association result of the microphenotype and the gene expression to obtain a predicted microphenotype corresponding to the single-cell gene expression.

[0013] The application further provides a device for associating gene expression and cell microphenotype based on machine learning, comprising the following modules: an acquisition module, configured to acquire cell imaging data and spatial transcriptome data, wherein the cell imaging data comprises a nucleus image of a target cell; a segmentation module, configured to perform single-cell segmentation on the nucleus image and the spatial transcriptome data based on fine-tuning of a large model in the field of vision, to obtain a cell segmentation image of the target cell; a post-segmentation processing module, configured to determine a cell mask of the target cell based on the cell segmentation image; a microphenotype quantification module, configured to quantify the cell mask of the target cell, to obtain a microphenotype of the target cell; a gene expression quantification module, configured to determine gene expression of the target cell based on the cell mask of the target cell and the spatial transcriptome data; and an association module, configured to perform association analysis on the microphenotype of the target cell and the gene expression of the target cell based on a preset machine learning model, to obtain an association result of the microphenotype and the gene expression of the target cell.

[0014] The application further provides an electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the method for associating gene expression and cell microphenotype based on machine learning according to any one of the above-described methods when executing the program.

[0015] The application further provides a non-transitory computer-readable storage medium having a computer program stored thereon, wherein the computer program is executable on a processor to implement the method for associating gene expression and cell microphenotype based on machine learning according to any one of the above-described methods.

[0016] The application further provides a computer program product comprising a computer program, wherein the computer program is executable on a processor to implement the method for associating gene expression and cell microphenotype based on machine learning according to any one of the above-described methods.

[0017] The application provides a method for correlating gene expression and cell microphenotype based on machine learning, which can significantly improve the accuracy and efficiency of cell segmentation by fine-tuning large models in the field of vision, using cell nucleus images and spatial transcriptome data for single cell segmentation; this method combining multiple data sources can capture the fine structure and boundary of cells better than traditional methods, thereby obtaining more accurate cell segmentation images; the cell mask determined based on the cell segmentation image and the quantification of the cell mask can obtain the microphenotype of the target cell; the gene expression information in the spatial transcriptome data can be accurately matched with the specific cell mask, avoiding the problem of mismatch between gene expression data and cell position in traditional methods; by using a preset machine learning model, the microphenotype of the target cell is correlated with the gene expression to obtain the correlation result of the microphenotype of the target cell and the gene expression, which can reveal the potential relationship between the two, helping to understand how gene expression affects the morphology and function of cells; and the technical problem of difficult data analysis in the related art is solved. BRIEF DESCRIPTION OF DRAWINGS

[0018] In order to more clearly illustrate the technical solutions in the application or the prior art, the following will briefly introduce the drawings needed to be used in the embodiments or the prior art description one by one. Obviously, the drawings in the following description are some embodiments of the application, and those skilled in the art can obtain other drawings according to these drawings without creative labor.

[0019] Figure 1 is a structural schematic diagram of the internal environment and the external environment provided by the application.

[0020] Figure 2 is a flowchart of the method for correlating gene expression and cell microphenotype based on machine learning provided by the application.

[0021] Figure 3 is an example diagram of cell imaging data and spatial transcriptome data provided by the application.

[0022] Figure 4 is a structural schematic diagram of cell boundary positioning provided by the application.

[0023] Figure 5 is an integrated schematic diagram of cell segmentation images and spatial transcriptome data provided by the application.

[0024] Figure 6 is a whole flowchart of the method for correlating gene expression and cell microphenotype based on machine learning provided by the application.

[0025] Figure 7is a structural schematic diagram of the device for associating gene expression and cell microphenotype based on machine learning provided by the present application.

[0026] Figure 8 is a physical structure schematic diagram of the electronic device provided by the present application. DETAILED DESCRIPTION

[0027] To make the objectives, technical solutions and advantages of the present application clearer, the technical solutions in the present application will be described clearly and completely below with reference to the drawings in the present application. Obviously, the described embodiments are some of the embodiments of the present application, rather than all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without creative work fall within the scope of protection of the present application.

[0028] In the past 20 years, great progress has been made in macro-level phenotypic traits (organs, plants and canopies), but little attention has been paid to micro-level phenotypic traits and their correlation with macro-level phenotypic traits. Phenotypic identification of these traits at the micro level will help to accurately characterize, accurately identify and systematically understand beneficial complex traits at the macro level.

[0029] REFERENCE Figure 1 , Figure 1 is a structural schematic diagram of the internal environment and the external environment provided by the present application, which includes an internal environment, an external environment, a genome, a transcriptome, a proteome, a metabolome, a microphenotype and a macrophenotype.

[0030] A series of subtle traits exhibited by plants, whether physical, chemical or biological properties, are the result of the expression of genotypes at a specific period, which are collectively referred to as "microphenotypes". Plants capture environmental stimuli at different levels from canopies to organs and exhibit plastic responses. These stimuli from the external environment are converted into stimuli in the internal environment through energy and mass exchange and transport at the cellular and tissue levels. The influence of such internal cues is brought about and amplified through signal transduction at the organelle level. Therefore, microphenotypes play a key role in bridging the gap between genotypes and complex macrophenotypes, and are very important for precision design breeding of crops.

[0031] With the development of molecular biology and cell biology, scientists have begun to focus on microphenotypes at the cellular level, such as cell size, shape, arrangement and tissue structure. These microphenotype characteristics can provide more detailed biological information and help to better understand the genetic characteristics and physiological mechanisms of crops.

[0032] Machine learning algorithms can learn and extract features from large amounts of data, establish a correlation model between gene expression and cell microphenotype. Through these models, the microphenotype performance of crops can be more accurately predicted, providing a scientific basis for breeding decisions.

[0033] Based on the spatial transcriptome data set, high-precision cell segmentation can be achieved through related algorithms, and the cell microphenotype can be quantified. Using the spatial transcriptome data set, the gene expression information at the corresponding location can be obtained, thereby obtaining the gene expression of single cells. Subsequently, through certain machine learning algorithms, the correlation between single cell microphenotype and gene expression is realized.

[0034] Alternatively, the method for correlating gene expression and cell microphenotype based on machine learning according to the embodiments of the present application can be executed by a server, or by a terminal device, or by both the server and the terminal device. Taking the method for correlating gene expression and cell microphenotype based on machine learning in the embodiments of the present application executed by the server as an example.

[0035] Figure 2 is a flowchart of the method for correlating gene expression and cell microphenotype based on machine learning provided by the present application, as shown in Figure 2 , the method comprises the following steps.

[0036] Step 201, obtaining cell imaging data and spatial transcriptome data, wherein the cell imaging data comprises a nucleus image of a target cell.

[0037] In order to observe the nucleus, it is usually necessary to use specific dyes to stain the nucleus to obtain the stained nucleus, and then use a fluorescence microscope or a confocal microscope to image the stained nucleus to obtain cell imaging data including the nucleus image of the target cell.

[0038] Spatial transcriptomics data (ST data) is a type of biological data that combines transcriptomics and spatial information. Transcriptome refers to the expression of all genes in a cell under specific conditions and at a specific time point, i.e. the sum of RNA molecules produced by the process of DNA transcription into RNA.

[0039] Spatial transcriptomics is based on transcriptomics with added spatial information, i.e. retaining the spatial location information of gene expression in tissues. This allows observation of the specific expression location of certain genes within the tissue, thereby understanding the relationship between gene expression and tissue structure.

[0040] Referring to Figure 3 , Figure 3 is an example diagram of cell imaging data and spatial transcriptome data provided by the present application.

[0041] As shown in Figure 3 , Figure 3 The left picture is a spatial transcriptome data visualization. Based on the spatial transcriptome data, the gene information and expression information at each coordinate point are determined. Figure 3 The right picture is a nucleus image.

[0042] With reference to Figure 2 , step 202, fine-tune the visual field large model based on the nucleus image and the spatial transcriptome data to perform single cell segmentation, and obtain a cell segmentation image of the target cell.

[0043] In the embodiments of the present application, based on the spatial transcriptome data and the imaging data, the gene expression information and the image information are comprehensively utilized, and the fine-tuned visual field large model Segment Anything is used to realize high-precision single cell segmentation.

[0044] The spatial transcriptome data of the tissue section is obtained, including the gene expression matrix and the corresponding spatial position information. The spatial transcriptome data is preprocessed, including removing noise, standardizing expression values, etc.

[0045] High-resolution nucleus images are obtained, which include the morphology and position of the nucleus. The nucleus images are preprocessed, such as denoising, enhancing contrast, correcting color, etc., to improve the accuracy of segmentation.

[0046] Align the nucleus image and the spatial transcriptome data so that each spatial transcriptome data point can correspond to the corresponding position in the imaging data.

[0047] A pre-trained visual field large model, such as Segment Anything, is determined, which has strong image segmentation capability. A part of single cells in the imaging data is manually labeled as a training set, and these labels accurately reflect the boundaries of the cells. The labeled training set is used to fine-tune the Segment Anything model, so that it can more accurately segment single cells. In the fine-tuning process, the gene expression information in the spatial transcriptome data is combined as an additional feature input to improve the segmentation performance of the model.

[0048] The preprocessed imaging data is input into the fine-tuned Segment Anything model (i.e., the fine-tuned visual field large model). The Segment Anything model predicts the probability of each pixel belonging to the target cell (i.e., the prediction result) based on the input imaging data and the additional feature input (i.e., the gene expression information in the spatial transcriptome data). The prediction result of the Segment Anything model is post-processed, such as threshold segmentation, morphological operation, etc., to obtain a cell segmentation image of the target cell.

[0049] Through the embodiment of the present application, by fine-tuning the large model Segment Anything in the field of vision and combining spatial transcriptome data and imaging data, high-precision single-cell segmentation can be realized.

[0050] Step 203, determining the cell mask of the target cell based on the cell segmentation image.

[0051] In the embodiment of the present application, the cell segmentation image is converted into a binary image, i.e., a cell mask. For example, a threshold is set, and the pixel values higher (or lower) than the threshold are set to white (or non-zero value), and the remaining pixel values are set to black (or zero value).

[0052] In the binary image, the white area represents the cell area. According to the analysis requirements, specific target cells can be further labeled or identified.

[0053] Step 204, quantifying the cell mask of the target cell to obtain the microscopic phenotype of the target cell.

[0054] In the embodiment of the present application, OpenCV is used to quantify the microscopic phenotype of the segmented cell image.

[0055] Among them, OpenCV (Open Source Computer Vision Library) is an open source computer vision and machine learning software library for processing and analyzing image and video data.

[0056] Step 205, determining the gene expression of the target cell based on the cell mask of the target cell and the spatial transcriptome data.

[0057] In the embodiment of the present application, the position of each cell mask can be obtained through the cell segmentation image, and the gene expression information of each position can be obtained through the spatial transcriptome data. Therefore, by combining spatial transcriptome and cell mask, the gene expression information of each cell can be obtained.

[0058] Through the embodiment of the present application, the gene expression information in the spatial transcriptome data is accurately corresponding to the specific cell mask, avoiding the problem of mismatch between gene expression data and cell position in the traditional method.

[0059] Step 206, performing correlation analysis on the microscopic phenotype of the target cell and the gene expression of the target cell based on the pre-set machine learning model to obtain the correlation result of the microscopic phenotype and the gene expression of the target cell.

[0060] In the embodiment of the present application, a machine learning algorithm such as Random Forest and Support Vector Machine (SVM) is used to realize the correlation analysis of single-cell gene expression-microphenotype, obtain the microphenotype-gene expression correlation result of the target cell, and further realize the mining of key genes determining the microphenotype of the cell and the single-cell microphenotype prediction analysis.

[0061] Through the above steps of the embodiment of the present application, the cell imaging data and the spatial transcriptome data are obtained, wherein the cell imaging data includes the nucleus image of the target cell; the single-cell segmentation is performed based on the nucleus image and the spatial transcriptome data by fine-tuning the visual field large model to obtain the cell segmentation image of the target cell; the quantitative segmentation is performed on the cell segmentation image to obtain the microphenotype of the target cell; the cell mask of the target cell is determined based on the cell segmentation image; the gene expression of the target cell is determined based on the cell mask of the target cell and the spatial transcriptome data; the correlation analysis is performed based on the microphenotype of the target cell and the gene expression of the target cell by the preset machine learning model to obtain the microphenotype-gene expression correlation result of the target cell; it is helpful to understand how the gene expression affects the morphology and function of the cell; and further, the technical problem that the data analysis is difficult in the trait phenotype identification mode in the related art is solved.

[0062] According to the method for correlating gene expression and cell microphenotype based on machine learning provided by the present application, the single-cell segmentation is performed based on the cell imaging data and the spatial transcriptome data by fine-tuning the visual field large model to obtain the cell image of the target cell, which comprises:

[0063] Based on the nucleus image and the spatial transcriptome data, the complete image data of the target cell on the cell imaging data is determined;

[0064] The single-cell segmentation is performed based on the complete image data by fine-tuning the visual field large model to obtain the cell boundary positioning;

[0065] Based on the cell boundary positioning, the cell segmentation image of the target cell is determined.

[0066] In the embodiment of the present application, based on the nucleus image, the position and morphology of the nucleus are recognized and positioned, wherein the nucleus image can be obtained by a fluorescence microscope; the spatial transcriptome data is registered with the nucleus image to ensure the consistency of the two in space; the spatial transcriptome data provides detailed information of gene expression in the cell, including the expression level of each gene in each cell or cell group; by comparing the nucleus position in the nucleus image and the gene expression of the spatial transcriptome data, the position of the target cell in the spatial transcriptome data is determined. According to the position information of the target cell in the nucleus image and the spatial transcriptome data, the complete image containing the target cell is extracted from the cell imaging data.

[0067] The pre-trained deep learning model is fine-tuned using the cell imaging dataset to obtain a fine-tuned visual field large model, which can accurately identify and segment individual cells.

[0068] The fine-tuned visual field large model is used to perform single-cell segmentation on the complete image data to obtain a segmentation result image, a segmentation mask is performed based on the segmentation result image to obtain the boundary positioning of each cell, and a cell segmentation image is generated according to the boundary positioning of each cell, in which each cell is clearly segmented.

[0069] Reference Figure 4 , Figure 4 is a structural diagram of cell boundary positioning provided by the application.

[0070] As Figure 4 indicated, the left image in Figure 4 is the output of the fine-tuned visual field large model (i.e., the output of the model), and here, the output of the fine-tuned visual field large model is visualized (e.g., segmentation mask) to obtain the boundary positioning of each cell (refer to the right image of Figure 4 ).

[0071] Through the embodiments of the application, the complete image data of the target cell on the cell imaging data is determined based on the nucleus image and the spatial transcriptome data, and single-cell segmentation is performed through the fine-tuned visual field large model, and finally the cell boundary positioning and the cell segmentation image are obtained.

[0072] According to the method for associating gene expression and cell microphenotype provided by the application, the microphenotype of the target cell includes at least one of the following: cell area, cell perimeter, cell minimum circumscribed circle radius, cell minimum circumscribed ellipse major axis, cell minimum circumscribed ellipse minor axis, cell minimum circumscribed rectangle length, and cell minimum circumscribed rectangle width.

[0073] In the embodiments of the application, the microphenotype of the segmented cell image is quantitatively segmented by using OpenCV, such as cell area, cell perimeter, cell minimum circumscribed circle radius, cell minimum circumscribed ellipse major axis, cell minimum circumscribed ellipse minor axis, cell minimum circumscribed rectangle length, and cell minimum circumscribed rectangle width.

[0074] The segmented cell image is read by using the cv2.imread() function of OpenCV; the image is converted to a grayscale image by using the cv2.cvtColor() function for subsequent processing; and the grayscale image is converted to a binary image by using the cv2.threshold() or adaptive threshold method (such as cv2.adaptiveThreshold()) to more clearly identify the cell boundary.

[0075] Find contours of cells on the binary image using the cv2.findContours() function; calculate the perimeter of each contour using the cv2.arcLength() function; find the minimum enclosing circle of each contour using the cv2.minEnclosingCircle() function and calculate its radius; fit the minimum enclosing ellipse of each contour using the cv2.fitEllipse() function and get the length of its major and minor axes; find the minimum enclosing rotated rectangle of each contour using the cv2.minAreaRect() function and get its length and width.

[0076] Draw the contours of cells and the minimum enclosing shapes on the original image or a new image using the drawing functions of OpenCV (such as cv2.circle(), cv2.ellipse(), cv2.rectangle(), etc.) to visually display the results.

[0077] By the embodiments of the present application, the segmented cell images can be quantitatively analyzed for microscopic phenotypes using OpenCV.

[0078] According to the method for associating gene expression and cell microscopic phenotype based on machine learning provided by the present application, the gene expression of the target cell is determined based on the cell mask of the target cell and the spatial transcriptome data, and the method comprises the following steps:

[0079] Determine the cell mask position of each cell mask of the target cell;

[0080] Align and integrate each cell mask position with the spatial transcriptome data to obtain the transcriptome data corresponding to each cell mask position;

[0081] Take the transcriptome data corresponding to each cell mask position as the gene expression of each cell mask;

[0082] Determine the gene expression of the target cell based on the gene expression of each cell mask of the target cell.

[0083] In the embodiments of the present application, the cell mask image corresponding to the target cell of the cell segmentation image is determined, wherein the area of the target cell is marked as 1 (white) and the other areas are marked as 0 (black); all pixel positions marked as 1 are determined as cell mask positions by traversing the cell mask image, which represent the boundaries or the inside of the target cell.

[0084] Map each cell mask position to the coordinate system of the spatial transcriptome data and extract the gene expression value of the corresponding position to obtain the transcriptome data corresponding to each cell mask position.

[0085] For example, if the spatial transcriptome data is in pixel units, the cell mask positions are converted to corresponding pixel coordinates. If the spatial transcriptome data is in area units (e.g., small areas in a tissue section), the cell mask is matched with the above-mentioned areas.

[0086] The extracted (spatial) transcriptome data is associated with the cell mask positions, ensuring that each cell mask has a corresponding gene expression dataset.

[0087] Reference Figure 5 , Figure 5 is a schematic diagram of integrating a cell segmentation image with spatial transcriptome data provided by the present application.

[0088] As shown in Figure 5 , the left picture in Figure 5 integrates the cell segmentation image with the spatial transcriptome data to extract transcriptome information at each cell position, for example, the spatial transcriptome information corresponding to the position of the cell segmentation image for cell 1 (reference Figure 5 ) is extracted as the gene expression of the cell.

[0089] Through the embodiments of the present application, the position of each cell mask of the target cell can be determined, and these positions can be aligned and integrated with the spatial transcriptome data, thereby obtaining the transcriptome data corresponding to each cell mask, and finally determining the gene expression of the target cell.

[0090] According to a method for associating gene expression and cell microphenotype based on machine learning provided by the present application, after a preset machine learning model is used to perform association analysis based on the microphenotype of a target cell and the gene expression of the target cell, obtaining a microphenotype and gene expression association result of the target cell, the method further comprises:

[0091] determining a target microphenotype of the target cell;

[0092] based on the microphenotype and gene expression association result, determining a gene expression having an association degree greater than a preset association degree threshold as a candidate gene expression, wherein the candidate gene expression is a target gene expression of the target microphenotype.

[0093] In the embodiments of the present application, the microphenotype refers to the morphological, structural or functional characteristics exhibited by the cell under specific conditions; the gene expression refers to the process of gene transcription into mRNA and further translation into protein, and these proteins are the basis for the morphological, structural or functional characteristics exhibited by the cell.

[0094] The gene expression data of each cell is matched with the corresponding microphenotype data, and the correlation between the gene expression and the microphenotype is analyzed by using a statistical method (such as correlation analysis, regression analysis, a machine learning algorithm, etc.); according to the result of the correlation analysis, the gene expression that is significantly correlated with the microphenotype of the target cell is screened out as the target gene expression, wherein the target gene expression directly participates in the regulation of the microphenotype of the target cell.

[0095] According to the method for correlating gene expression and cell microphenotype based on machine learning provided by the application, after the preset machine learning model is used to perform correlation analysis on the microphenotype of the target cell and the gene expression of the target cell, the microphenotype of the target cell and the gene expression correlation result of the target cell are obtained, the method further comprises the following steps.

[0096] Obtaining single-cell gene expression input by a user;

[0097] Based on the microphenotype and gene expression correlation result, the single-cell gene expression is predicted to obtain a predicted microphenotype corresponding to the single-cell gene expression.

[0098] In the embodiment of the application, the single-cell microphenotype input by the user at least includes one of the morphology, size, structural characteristics and functional state of the cell.

[0099] The known single-cell gene expression data and the corresponding microphenotype data are collected as a training set, and a machine learning algorithm (such as a support vector machine, a random forest, a neural network, etc.) or a deep learning model (such as a convolutional neural network, a recurrent neural network, etc.) is used to perform model training, with the gene expression features as the input and the microphenotype as the output. The trained model is used to predict the single-cell gene expression input by the user to obtain a predicted microphenotype.

[0100] Through the embodiment of the application, by obtaining the single-cell microphenotype input by the user and based on the correlation result of the microphenotype and the gene expression, a predicted microphenotype corresponding to the single-cell gene expression can be obtained, which helps us to better understand the function and characteristics of the cell.

[0101] Reference Figure 6 , Figure 6 is the overall flowchart of the method for correlating gene expression and cell microphenotype based on machine learning provided by the application, and specifically comprises the following steps.

[0102] High-precision single-cell segmentation: based on spatial transcriptome data and imaging data, gene expression information and image information are comprehensively utilized, and high-precision single-cell segmentation is realized by fine-tuning the large model Segment Anything in the field of vision.

[0103] Single-cell microphenotype quantification: OpenCV is used to quantitatively segment the microphenotype of the cell image after segmentation, such as cell area, cell perimeter, cell minimum circumscribed circle radius, cell minimum circumscribed ellipse major axis, cell minimum circumscribed ellipse minor axis, cell minimum circumscribed rectangle length, cell minimum circumscribed rectangle width, and other microphenotypes.

[0104] Single-cell gene expression quantification: the position of each cell mask can be obtained through the segmented cell image, and the gene expression information of each position can be understood through the spatial transcriptome data. Therefore, combined with the spatial transcriptome and the cell mask picture, the gene expression information of each cell can be obtained.

[0105] Here, based on the CellMask position and spatial transcriptome data, single-cell gene expression quantification is determined.

[0106] Gene expression-microphenotype correlation analysis: machine learning algorithms such as random forest and SVM are used to realize the correlation analysis of single-cell gene expression-microphenotype, realize the mining of key genes that determine cell microphenotype, single-cell microphenotype prediction analysis, etc.

[0107] The following describes the gene expression and cell microphenotype correlation device based on machine learning provided by the application, and the gene expression and cell microphenotype correlation device based on machine learning described below can be mutually corresponding and referred to.

[0108] Reference Figure 7 , Figure 7 is a structural schematic diagram of the gene expression and cell microphenotype correlation device based on machine learning provided by the application, which comprises: an acquisition module 701, a segmentation module 702, a post-segmentation processing module 703, a microphenotype quantification module 704, a gene expression quantification module 705, and a correlation module 706.

[0109] The acquisition module 701 is used to acquire cell imaging data and spatial transcriptome data, wherein the cell imaging data comprises a nucleus image of a target cell;

[0110] The segmentation module 702 is used to perform single-cell segmentation based on the nucleus image and the spatial transcriptome data by fine-tuning a large model in the vision field, to obtain a cell segmentation image of the target cell;

[0111] The post-segmentation processing module 703 is used to determine a cell mask of the target cell based on the cell segmentation image;

[0112] The microphenotype quantification module 704 is used to quantitatively analyze the cell mask of the cell segmentation image, to obtain a microphenotype of the target cell;

[0113] The gene expression quantification module 705 is configured to determine gene expression of the target cell based on the cell mask of the target cell and the spatial transcriptome data.

[0114] The association module 706 is configured to perform association analysis on the microphenotype and the gene expression of the target cell based on a preset machine learning model, to obtain an association result of the microphenotype and the gene expression of the target cell.

[0115] Specifically, the above-mentioned device for associating gene expression and cell microphenotype based on machine learning can realize all method steps of the above-mentioned method embodiment for associating gene expression and cell microphenotype based on machine learning, and can achieve the same technical effects. Here, the same parts and beneficial effects in the method embodiment will not be described in detail.

[0116] Figure 8 is a schematic diagram of the physical structure of the electronic device provided by the present application, as shown in Figure 8 The electronic device can include a processor 810, a communications interface 820, a memory 830, and a communications bus 840, wherein the processor 810, the communications interface 820, and the memory 830 can communicate with each other through the communications bus 840. The processor 810 can invoke the logic instructions in the memory 830 to execute the method for associating gene expression and cell microphenotype based on machine learning, which includes: obtaining cell imaging data and spatial transcriptome data, wherein the cell imaging data includes a nucleus image of a target cell; performing single cell segmentation based on the nucleus image and the spatial transcriptome data by fine-tuning a large model in the field of vision to obtain a cell segmentation image of the target cell; performing quantitative segmentation on the cell segmentation image to obtain a microphenotype of the target cell; determining a cell mask of the target cell based on the cell segmentation image; performing quantitative segmentation on the cell mask of the target cell and the cell segmentation image to obtain the microphenotype of the target cell; determining gene expression of the target cell based on the cell mask of the target cell and the spatial transcriptome data; and performing association analysis on the microphenotype and the gene expression of the target cell based on a preset machine learning model, to obtain an association result of the microphenotype and the gene expression of the target cell.

[0117] In addition, the logical instructions in the memory 830 described above can be implemented in the form of a software function unit and sold or used as an independent product, and can be stored in a computer readable storage medium. Based on such understanding, the technical solutions of the present application essentially or the part that contributes to the prior art or part of the technical solutions can be embodied in the form of a software product. The computer software product is stored in a storage medium, and includes several instructions for making a computer device (which can be a personal computer, a server, or a network device, etc.) execute all or part of the steps of the embodiments of the present application. The aforementioned storage medium includes: a U disk, a mobile hard disk, a read-only memory (ROM, Read-Only Memory), a random access memory (RAM, Random Access Memory), a magnetic disk or an optical disk, and various media that can store program codes.

[0118] In another aspect, the present application also provides a computer program product, the computer program product comprising a computer program, the computer program being stored on a non-transitory computer readable storage medium, and the computer program being executable by a processor to cause a computer to perform the method of associating gene expression and cell microphenotype based on machine learning provided by the above-mentioned methods, the method comprising: obtaining cell imaging data and spatial transcriptome data, wherein the cell imaging data comprises a nucleus image of a target cell; performing single cell segmentation based on the nucleus image and the spatial transcriptome data by fine-tuning a large model in the field of vision to obtain a cell segmentation image of the target cell; performing quantitative segmentation on the cell segmentation image to obtain a microphenotype of the target cell; determining a cell mask of the target cell based on the cell segmentation image; performing quantitative segmentation on the cell mask cell segmentation image of the target cell to obtain a microphenotype of the target cell; determining gene expression of the target cell based on the cell mask of the target cell and the spatial transcriptome data; and performing correlation analysis on the microphenotype of the target cell and the gene expression of the target cell based on a preset machine learning model to obtain a correlation result of the microphenotype of the target cell and the gene expression of the target cell.

[0119] In yet another aspect, the present application also provides a non-transitory computer-readable storage medium having stored thereon a computer program, which, when executed by a processor, implements a method for associating gene expression and cell microphenotype based on machine learning, the method comprising: obtaining cell imaging data and spatial transcriptome data, wherein the cell imaging data comprises a nucleus image of a target cell; performing single cell segmentation based on the nucleus image and the spatial transcriptome data by fine-tuning a large model in the field of vision to obtain a cell segmentation image of the target cell; performing quantitative segmentation on the cell segmentation image to obtain a microphenotype of the target cell; determining a cell mask of the target cell based on the cell segmentation image; performing quantitative segmentation on the cell segmentation image of the cell mask of the target cell to obtain a microphenotype of the target cell; determining gene expression of the target cell based on the cell mask of the target cell and the spatial transcriptome data; and performing association analysis on the microphenotype of the target cell and the gene expression of the target cell based on a preset machine learning model to obtain an association result of the microphenotype of the target cell and the gene expression.

[0120] The device embodiments described above are merely illustrative, wherein the units shown as separate components can or can not be physically separated, and the components shown as units can or can not be physical units, i.e., they can be located in one place or distributed on multiple network units. Part or all of the modules can be selected to achieve the purpose of the embodiment according to actual needs. Those skilled in the art can understand and implement without creative labor.

[0121] From the above description of the embodiments, those skilled in the art can clearly understand that the embodiments can be realized by means of software plus necessary general hardware platforms, and of course can also be realized by hardware. Based on such understanding, the above technical solutions, essentially or in terms of the contribution to the prior art, can be embodied in the form of a software product. The computer software product can be stored in a computer readable storage medium, such as a ROM / RAM, a magnetic disk, an optical disk, etc., and includes a plurality of instructions to cause a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the methods described in each embodiment or some parts of the embodiments.

[0122] Finally, it should be noted that: the above embodiments are only used to illustrate the technical solutions of the present application, and not to limit them; although the present application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that: it can still modify the technical solutions recorded in the foregoing embodiments, or make equivalent replacement to some technical features; and these modifications or replacements do not make the corresponding technical solutions deviate from the spirit and scope of the technical solutions of the embodiments of the present application.

Claims

1. A method of correlating gene expression and cellular microphenotypes based on machine learning, characterized by, The method comprises: acquiring cell imaging data and spatial transcriptome data, wherein the cell imaging data comprises a nucleus image of a target cell; performing single-cell segmentation on the nucleus image and the spatial transcriptome data based on a fine-tuned visual domain large model to obtain a cell segmentation image of the target cell; determining a cell mask of the target cell based on the cell segmentation image; quantifying the cell mask of the target cell to obtain a microscopic phenotype of the target cell; determining gene expression of the target cell based on the cell mask of the target cell and the spatial transcriptome data; performing correlation analysis on the microscopic phenotype of the target cell and the gene expression of the target cell based on a preset machine learning model to obtain a correlation result of the microscopic phenotype and the gene expression of the target cell; the microscopic phenotype of the target cell comprises at least one of the following: cell area, cell perimeter, cell minimum circumscribed circle radius, cell minimum circumscribed ellipse major axis, cell minimum circumscribed ellipse minor axis, cell minimum circumscribed rectangle length, and cell minimum circumscribed rectangle width; the determination of the gene expression of the target cell based on the cell mask of the target cell and the spatial transcriptome data comprises: determining each cell mask position of the target cell; aligning and integrating each cell mask position with the spatial transcriptome data to obtain transcriptome data corresponding to each cell mask position; using the transcriptome data corresponding to each cell mask position as the gene expression of each cell mask; determining the gene expression of the target cell based on the gene expression of each cell mask of the target cell.

2. The method of claim 1, wherein the method is based on machine learning. the single-cell segmentation on the nucleus image and the spatial transcriptome data based on the fine-tuned visual domain large model to obtain the cell image of the target cell comprises: determining complete image data of the target cell on the cell imaging data based on the nucleus image and the spatial transcriptome data; performing single-cell segmentation on the complete image data based on the fine-tuned visual domain large model to obtain cell boundary positioning; determining the cell segmentation image of the target cell based on the cell boundary positioning.

3. The method of claim 1, wherein the method is based on machine learning. after the correlation analysis on the microscopic phenotype of the target cell and the gene expression of the target cell based on the preset machine learning model to obtain the correlation result of the microscopic phenotype and the gene expression of the target cell, the method further comprises: determining a target microscopic phenotype of the target cell; determining a candidate gene expression as the target gene expression of the target microscopic phenotype based on the correlation result of the microscopic phenotype and the gene expression, wherein the correlation degree of the candidate gene expression with the target microscopic phenotype is greater than a preset correlation degree threshold.

4. The method of claim 1, wherein the method is based on machine learning. after the correlation analysis on the microscopic phenotype of the target cell and the gene expression of the target cell based on the preset machine learning model to obtain the correlation result of the microscopic phenotype and the gene expression of the target cell, the method further comprises: acquiring single-cell gene expression input by a user; The single-cell gene expression is predicted based on the microphenotype-gene expression association result, to obtain a predicted microphenotype corresponding to the single-cell gene expression.

5. An apparatus for associating gene expression and cellular microphenotypes based on machine learning, comprising: The method comprises the following steps: An acquisition module is configured to acquire cell imaging data and spatial transcriptome data, wherein the cell imaging data comprises a nucleus image of a target cell; A segmentation module is configured to perform single-cell segmentation based on the nucleus image and the spatial transcriptome data by fine-tuning a large model in the field of vision, to obtain a cell segmentation image of the target cell; A post-segmentation processing module is configured to determine a cell mask of the target cell based on the cell segmentation image; A microphenotype quantification module is configured to quantify the cell mask of the target cell, to obtain a microphenotype of the target cell; A gene expression quantification module is configured to determine gene expression of the target cell based on the cell mask of the target cell and the spatial transcriptome data; An association module is configured to perform association analysis on the microphenotype of the target cell and the gene expression of the target cell based on a preset machine learning model, to obtain a microphenotype-gene expression association result of the target cell; The microphenotype of the target cell comprises at least one of the following: cell area, cell perimeter, cell minimum circumscribed circle radius, cell minimum circumscribed ellipse major axis, cell minimum circumscribed ellipse minor axis, cell minimum circumscribed rectangle length, and cell minimum circumscribed rectangle width. The determination of the gene expression of the target cell based on the cell mask of the target cell and the spatial transcriptome data comprises the following steps: Determining each cell mask position of the target cell; Aligning and integrating each cell mask position with the spatial transcriptome data, to obtain transcriptome data corresponding to each cell mask position; Taking the transcriptome data corresponding to each cell mask position as gene expression of each cell mask; Determining the gene expression of the target cell based on the gene expression of each cell mask of the target cell.

6. An electronic device comprising a memory, a processor, and a computer program stored on the memory and executable on the processor, characterized in that, The processor executes the computer program to implement the method for associating gene expression and cell microphenotype based on machine learning according to any one of claims 1 to 4.

7. A non-transitory computer-readable storage medium having stored thereon a computer program, characterized in that, The computer program is executed by the processor to implement the method for associating gene expression and cell microphenotype based on machine learning according to any one of claims 1 to 4.

8. A computer program product comprising a computer program, characterized in that, The computer program is executed by the processor to implement the method for associating gene expression and cell microphenotype based on machine learning according to any one of claims 1 to 4.

Images