Construction method and system of rice root single cell type annotation model
By integrating rice root single-cell sequencing data through double normalization and gradient boosting decision tree, a high-precision rice root single-cell type annotation model was constructed, which solved the problem of inaccurate rice root cell type identification in existing technologies and improved the automation and accuracy of research.
Patent Information
- Application Number
- CN202510790809.7
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-06-13
- Publication Date
- 2025-09-26
AI Technical Summary
Existing technologies lack cell type prediction tools that integrate multiple datasets, are highly automated, and are specific to plant roots, resulting in inaccurate identification of rice root cell types and poor research reproducibility.
A double normalization strategy and gradient boosting decision tree were used to integrate rice root single-cell sequencing data through a machine learning algorithm. The model was trained to learn different types of rice root single cells and construct a high-precision rice root single-cell type annotation model.
It has achieved efficient and accurate identification of single-cell types of rice roots, improved the degree of research automation and the ability to migrate datasets across datasets, and enhanced the efficiency and accuracy of annotation of single-cell types of rice roots.
Smart Images

Figure CN120708710A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the interdisciplinary technical field of bioinformatics and plant genomics, and in particular to a method and system for constructing a rice root single cell type annotation model. Background Art
[0002] With the rapid development of single-cell sequencing technology, it has shown great potential in analyzing the heterogeneity of complex plant tissue cells. As an important food crop, the identification of root cell types in rice is crucial for studying nutrient absorption and stress resistance mechanisms. However, existing plant single-cell analysis methods have significant limitations: traditional annotation methods rely on manual screening of marker genes and subjective clustering, resulting in inconsistent results across data sets; mainstream annotation tools (such as Seurat, scPred, etc.) are mainly developed for animal cells and are not adaptable enough to plant cell wall interference and a higher proportion of environmental response genes; in addition, rice root cells have highly dynamic differentiation characteristics, and existing models have difficulty in accurately identifying transitional cell types. Although the recently published rice root single-cell atlas provides basic data for research, the lack of a cell type prediction tool that integrates multiple data sets, is highly automated, and is dedicated to plant roots has seriously restricted the reproducibility and large-scale application of related research.
[0003] Therefore, the present invention proposes a method for constructing a single cell type annotation model for rice roots. Summary of the Invention
[0004] The present invention provides a method and system for constructing a single-cell type annotation model for rice roots, which are used to address the current lack of a cell type prediction tool that integrates multiple data sets, has a high degree of automation and is dedicated to plant roots.
[0005] The present invention provides a method for constructing a rice root single cell type annotation model, comprising:
[0006] Obtain rice root single-cell sequencing data from different data sources, wherein the rice root single-cell sequencing data is annotated with corresponding rice root single-cell type labels;
[0007] A double normalization strategy was used to integrate rice root single-cell sequencing data from different data sources to obtain double-normalized rice root single-cell sequencing data.
[0008] Based on the double-normalized rice root single-cell sequencing data, the gradient boosting decision tree was used to train the model to learn the rice root single-cell sequencing data corresponding to different types of rice root single cells, and obtain the rice root single-cell type annotation model.
[0009] According to a method for constructing a rice root single-cell type annotation model provided by the present invention, the method adopts a dual normalization strategy to integrate rice root single-cell sequencing data from different data sources to obtain dual-normalized rice root single-cell sequencing data, including:
[0010] Using a machine learning algorithm for data integration, we performed deep batch correction on rice root single-cell sequencing data from different data sources to eliminate technical variation interference and obtain batch-effect-free rice root single-cell sequencing data.
[0011] The single-cell sequencing data analysis toolkit was used to perform plant-specific preprocessing on the rice root single-cell sequencing data after eliminating batch effects, and the double-normalized rice root single-cell sequencing data was obtained.
[0012] According to the present invention, a method for constructing a rice root single-cell type annotation model is provided, wherein the method uses a single-cell sequencing data analysis toolkit to perform plant-specific preprocessing on rice root single-cell sequencing data after eliminating batch effects to obtain double-normalized rice root single-cell sequencing data, including:
[0013] Based on the rice root single-cell sequencing data after eliminating batch effects, the total expression level of each single cell was normalized to 1×10 4 (1e4) to eliminate the differences in sequencing depth between single cells and obtain normalized rice root single-cell sequencing data;
[0014] The normalized rice root single-cell sequencing data was log-transformed to make the expression distribution closer to normal, thus obtaining log-transformed rice root single-cell sequencing data;
[0015] Using logarithmically transformed rice root single-cell sequencing data, we screened genes with significant variation (highly variable genes) based on the dispersion-mean distribution of gene expression to reduce noise in subsequent analyses.
[0016] The expression levels of genes with significant variation were characteristically scaled (mean = 0, variance = 1) to ensure comparability between different genes, and double-normalized rice root single-cell sequencing data were obtained.
[0017] According to a method for constructing a rice root single-cell type annotation model provided by the present invention, the method adopts a dual normalization strategy to integrate rice root single-cell sequencing data from different data sources to obtain dual-normalized rice root single-cell sequencing data, and further includes:
[0018] Random downsampling is performed on clusters with more than 5,000 cells, while synthetic minority class oversampling technology is used for rare cell types (such as quiescent central cells) to improve the class imbalance problem.
[0019] According to a method for constructing a rice root single cell type annotation model provided by the present invention, a gradient boosting decision tree adopts a CatBoost enhanced architecture (CatBoost is a gradient boosting framework for symmetric decision trees).
[0020] The method uses a gradient boosting decision tree to train a model based on the double-normalized rice root single-cell sequencing data to learn the rice root single-cell sequencing data corresponding to different types of rice root single cells, thereby obtaining a rice root single-cell type annotation model, including:
[0021] Based on the double-normalized rice root single-cell sequencing data, target encoding is performed on the rice root single-cell sequencing data corresponding to different types of rice root single cells, and feature combinations are generated;
[0022] Based on the feature combination, a symmetric tree is constructed in each round to fit the negative gradient (residual) of the current model and update the gradient using an ordered boosting strategy;
[0023] The prediction results of all trees were weighted and combined as the model output to obtain the rice root single cell type annotation model.
[0024] According to the present invention, a method for constructing a rice root single-cell type annotation model is provided. The method comprises: using double-normalized rice root single-cell sequencing data, utilizing a gradient boosting decision tree, training a model to learn rice root single-cell sequencing data corresponding to different types of rice root cells, and obtaining a rice root single-cell type annotation model. The method comprises:
[0025] The double-normalized rice root single-cell sequencing data were divided into training and test sets at a ratio of 8:2;
[0026] When training the rice root single cell type annotation model, the model hyperparameters were optimized through ten-fold cross-validation, and the area under the characteristic curve (AUROC), confusion matrix weighted F1 score, and cross-dataset migration accuracy were used to verify the model performance.
[0027] The present invention also provides a method for annotating single cell types of rice roots, comprising:
[0028] Obtaining rice root single-cell sequencing data of a rice root single cell to be tested, wherein the cell type of the rice root single cell to be tested is unknown;
[0029] Obtaining, based on the rice root single-cell sequencing data of the rice root single cell to be tested, a probability that the rice root single cell to be tested belongs to a certain rice root single-cell type using a rice root single-cell type annotation model obtained by any of the above methods for constructing a rice root single-cell type annotation model;
[0030] According to the probability that the rice root single cell to be tested belongs to a certain rice root single cell type, the rice root single cell sequencing data of the rice root single cell to be tested is annotated with the rice root single cell type.
[0031] The present invention also provides a rice root single cell type annotation system, comprising:
[0032] A data acquisition module is used to: acquire rice root single-cell sequencing data of a rice root single cell to be tested, wherein the cell type of the rice root single cell to be tested is unknown;
[0033] a prediction module, configured to: obtain, based on the rice root single cell sequencing data of the rice root single cell to be tested and the rice root single cell type annotation model obtained by any of the above-described methods for constructing a rice root single cell type annotation model, a probability that the rice root single cell to be tested belongs to a certain rice root single cell type;
[0034] The annotation module is used to annotate the rice root single cell type of the rice root single cell sequencing data of the rice root single cell to be tested according to the probability that the rice root single cell to be tested belongs to a certain rice root single cell type.
[0035] The present invention also provides an electronic device comprising a processor and a memory storing a computer program, wherein when the processor executes the computer program, the method for constructing a rice root single cell type annotation model and / or the method for annotating a rice root single cell type described above is implemented.
[0036] The present invention also provides a non-transitory computer-readable storage medium having a computer program stored thereon. When the computer program is executed by a processor, it implements any of the above-mentioned methods for constructing a rice root single cell type annotation model and / or a rice root single cell type annotation method.
[0037] The present invention also provides a computer program product, which includes a computer program. The computer program can be stored on a non-transitory computer-readable storage medium. When the computer program is executed by a processor, the computer can execute any of the above-mentioned methods for constructing a rice root single cell type annotation model and / or the rice root single cell type annotation method.
[0038] The present invention provides a method for constructing a rice root single-cell type annotation model and a system thereof. The method adopts a dual normalization strategy to integrate rice root single-cell sequencing data from different data sources, thereby eliminating batch effects in multi-source data integration, optimizing plant-specific high-noise features, and accurately dividing the dynamic continuous lineage of rice root single-cell types. The method also utilizes a gradient boosting decision tree to train a model to learn rice root single-cell sequencing data corresponding to different types of rice root cells through an ordered boosting strategy and a classification feature processing mechanism. This effectively overcomes the class imbalance and high-dimensional sparsity problems existing in plant single-cell sequencing data, thereby obtaining a high-precision rice root single-cell type annotation model. The model can accurately identify the probability that a rice root cell to be tested belongs to a certain rice root single-cell type based on the rice root single-cell sequencing data of the rice root single cell to be tested and perform rice root single-cell type annotation, thereby greatly improving the efficiency, accuracy, and automation of rice root single-cell type annotation and promoting plant cell research. BRIEF DESCRIPTION OF THE DRAWINGS
[0039] In order to more clearly illustrate the technical solutions in the present invention or the prior art, a brief introduction will be given below to the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.
[0040] Figure 1 This is one of the flow charts for the construction and application of a rice root single cell type annotation model provided by the present invention.
[0041] Figure 2 This is the second flow chart of the construction and application of a rice root single cell type annotation model provided by the present invention.
[0042] Figure 3 Schematic diagram of using AUROC to measure the performance of the rice root single-cell type annotation model. The area under the curve shows the relationship between cells that were incorrectly assigned to clusters and cells that were correctly assigned by the rice root single-cell type annotation model.
[0043] Figure 4 Automatic annotation of custom-generated scRNA-seq clusters in rice root tips for UMAP visualization.
[0044] Figure 5 Validation of novel vascular cell type marker genes annotated by the rice root single-cell type annotation model is shown for UMAP visualization (left) and in situ hybridization (right). Scale bar, 50 μm.
[0045] Figure 6 This is a schematic diagram of the structure of a rice root single cell type annotation system provided by the present invention.
[0046] Figure 7 This is a schematic structural diagram of the electronic device provided by the present invention. DETAILED DESCRIPTION
[0047] In order to make the purpose, technical solutions and advantages of the present invention clearer, the technical solutions of the present invention will be clearly and completely described below in conjunction with the drawings in the present invention. Obviously, the embodiments described are part of the embodiments of the present invention, not all of the embodiments, and they should not be understood as limitations on the present invention. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present invention. In the description of the present invention, it should be understood that the terms used are only for descriptive purposes and cannot be understood as indicating or implying relative importance.
[0048] Figure 1 and Figure 2 A schematic diagram illustrating the construction and application of a rice root single-cell type annotation model provided by the present invention. The methods for constructing a rice root single-cell type annotation model and the rice root single-cell type annotation method provided by the present invention can be executed by any applicable terminal-side or network-side device, such as a rice root single-cell type annotation device.
[0049] See also Figure 1 and Figure 2 The present invention provides a method for constructing a rice root single cell type annotation model, which may include:
[0050] S110 . Obtain rice root single-cell sequencing data from different data sources, wherein the rice root single-cell sequencing data is annotated with corresponding rice root single-cell type labels.
[0051] In one embodiment, single-cell sequencing data of rice roots can be obtained from different experimental platforms (10x Genomics, Smart-seq2, etc.). The single-cell types of rice roots can include cell groups such as epidermis, exodermis, sclerenchyma, cortex, endodermis, pericycle, meristem, and vascular tissue.
[0052] S120. A double normalization strategy is used to integrate rice root single-cell sequencing data from different data sources to obtain double-normalized rice root single-cell sequencing data.
[0053] In one embodiment, S120 may include:
[0054] S1201. Using a machine learning algorithm for data integration, perform deep batch correction on rice root single-cell sequencing data from different data sources to eliminate technical variation interference, thereby obtaining rice root single-cell sequencing data after eliminating batch effects.
[0055] S1202. Using a single-cell sequencing data analysis toolkit, perform plant-specific preprocessing on the rice root single-cell sequencing data after eliminating batch effects to obtain double-normalized rice root single-cell sequencing data.
[0056] S1203. Random downsampling is performed on clusters with more than 5000 cells, while synthetic minority class oversampling technology is used for rare cell types (such as quiescent central cells) to improve the class imbalance problem.
[0057] In one embodiment, S1202 may include:
[0058] Based on the rice root single-cell sequencing data after eliminating batch effects, the total expression level of each single cell was normalized to 1×10 4 (1e4) to eliminate the differences in sequencing depth between single cells and obtain normalized rice root single-cell sequencing data;
[0059] The normalized rice root single-cell sequencing data was log-transformed to make the expression distribution closer to normal, thus obtaining log-transformed rice root single-cell sequencing data;
[0060] Using log-transformed rice root single-cell sequencing data, we screened genes with significant variation (hypervariable genes, specifically those with high expression in some cells but low expression in others) based on the dispersion-mean distribution of gene expression to reduce noise in subsequent analyses.
[0061] The expression levels of genes with significant variation were characteristically scaled (mean = 0, variance = 1) to ensure comparability between different genes, and double-normalized rice root single-cell sequencing data were obtained.
[0062] S130. Based on the double-normalized rice root single-cell sequencing data, a gradient boosting decision tree is used to train a model to learn the rice root single-cell sequencing data corresponding to different types of rice root single cells, and obtain a rice root single-cell type annotation model.
[0063] In one embodiment, the gradient boosting decision tree adopts the CatBoost enhanced architecture (CatBoost is a gradient boosting framework for symmetric decision trees), and S130 may include:
[0064] Based on the double-normalized rice root single-cell sequencing data, target encoding is performed on the rice root single-cell sequencing data corresponding to different types of rice root single cells, and feature combinations are generated;
[0065] Based on the feature combination, a symmetric tree is constructed in each round to fit the negative gradient (residual) of the current model and update the gradient using an ordered boosting strategy;
[0066] The prediction results of all trees were weighted and combined as the model output to obtain the rice root single cell type annotation model.
[0067] In one embodiment, S130 may further include:
[0068] The double-normalized rice root single-cell sequencing data were divided into training and test sets at a ratio of 8:2;
[0069] When training the rice root single cell type annotation model, the model hyperparameters were optimized through ten-fold cross-validation, and the area under the characteristic curve (AUROC), confusion matrix weighted F1 score, and cross-dataset migration accuracy were used to verify the model performance.
[0070] After obtaining the rice root single cell type annotation model, it can be applied to rice root single cell type annotation to form a rice root single cell type annotation method, including:
[0071] Obtaining rice root single-cell sequencing data of a rice root single cell to be tested, wherein the cell type of the rice root single cell to be tested is unknown;
[0072] Based on the rice root single-cell sequencing data of the rice root single cell to be tested, the probability that the rice root single cell to be tested belongs to a certain rice root single-cell type is obtained using the rice root single-cell type annotation model;
[0073] According to the probability that the rice root single cell to be tested belongs to a certain rice root single cell type, the rice root single cell sequencing data of the rice root single cell to be tested is annotated with the rice root single cell type.
[0074] In one embodiment, the rice root single cell type annotation model can also be used to mine new cell group-specific marker genes through feature importance analysis, specifically the feature importance of the screening model, which corresponds to the marker genes of each cell type.
[0075] This embodiment uses 10×G enomics The platform's rice single-cell sequencing data was used to verify the construction method and application of a rice root single-cell type annotation model provided by the present invention. The rice root single-cell type annotation model achieved an AUROC value of over 0.96 in cross-dataset testing, an improvement of approximately 15% over existing tools, a confusion matrix weighted F1 score of 0.94±0.03, a cross-dataset migration accuracy of 94.1%, and supported the continuous inclusion of new datasets for incremental training.
[0076] The present invention provides a method for constructing a rice root single-cell type annotation model and a system thereof. The method adopts a dual normalization strategy to integrate rice root single-cell sequencing data from different data sources, thereby eliminating batch effects in multi-source data integration, optimizing plant-specific high-noise features, and accurately dividing the dynamic continuous lineage of rice root single-cell types. The method also utilizes a gradient boosting decision tree to train a model to learn rice root single-cell sequencing data corresponding to different types of rice root cells through an ordered boosting strategy and a classification feature processing mechanism. This effectively overcomes the class imbalance and high-dimensional sparsity problems existing in plant single-cell sequencing data, thereby obtaining a high-precision rice root single-cell type annotation model. The model can accurately identify the probability that a rice root cell to be tested belongs to a certain rice root single-cell type based on the rice root single-cell sequencing data of the rice root single cell to be tested and perform rice root single-cell type annotation, thereby greatly improving the efficiency, accuracy, and automation of rice root single-cell type annotation and promoting plant cell research.
[0077] The rice root single cell type annotation system provided by the present invention is described below. The rice root single cell type annotation system described below and the rice root single cell type annotation method described above can be referenced to each other.
[0078] See also Figure 6 The present invention provides a rice root single cell type annotation system, which may include:
[0079] A data acquisition module is used to: acquire rice root single-cell sequencing data of a rice root single cell to be tested, wherein the cell type of the rice root single cell to be tested is unknown;
[0080] a prediction module, configured to: obtain, based on the rice root single cell sequencing data of the rice root single cell to be tested and the rice root single cell type annotation model obtained by any of the above-described methods for constructing a rice root single cell type annotation model, a probability that the rice root single cell to be tested belongs to a certain rice root single cell type;
[0081] The annotation module is used to annotate the rice root single cell type of the rice root single cell sequencing data of the rice root single cell to be tested according to the probability that the rice root single cell to be tested belongs to a certain rice root single cell type.
[0082] Figure 7 An example of a physical structure diagram of an electronic device is shown below. Figure 7As shown, the electronic device may include: a processor 810, a communications interface 820, a memory 830, and a communication bus 840, wherein the processor 810, the communications interface 820, and the memory 830 communicate with each other via the communication bus 840. The processor 810 may call the logic instructions in the memory 830 to perform the following steps:
[0083] Obtaining rice root single-cell sequencing data of a rice root single cell to be tested, wherein the cell type of the rice root single cell to be tested is unknown;
[0084] Obtaining, based on the rice root single-cell sequencing data of the rice root single cell to be tested, a probability that the rice root single cell to be tested belongs to a certain rice root single-cell type using a rice root single-cell type annotation model obtained by any of the above methods for constructing a rice root single-cell type annotation model;
[0085] According to the probability that the rice root single cell to be tested belongs to a certain rice root single cell type, the rice root single cell sequencing data of the rice root single cell to be tested is annotated with the rice root single cell type.
[0086] In addition, the logic instructions in the above-mentioned memory 830 can be implemented in the form of a software functional unit and can be stored in a computer-readable storage medium when sold or used as an independent product. Based on this understanding, the technical solution of the present invention, or the part that contributes to the prior art, or the part of the technical solution, 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 enabling a computer device (which can be a personal computer, server, or network device, etc.) to perform all or part of the steps of the method described in each embodiment of the present invention. The aforementioned storage medium includes: various media that can store program codes, such as a USB flash drive, a mobile hard disk, a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk.
[0087] In another aspect, the present invention further provides a computer program product, comprising a computer program, which may be stored on a non-transitory computer-readable storage medium, and wherein when the computer program is executed by a processor, the computer is capable of performing the following steps:
[0088] Obtaining rice root single-cell sequencing data of a rice root single cell to be tested, wherein the cell type of the rice root single cell to be tested is unknown;
[0089] Obtaining, based on the rice root single-cell sequencing data of the rice root single cell to be tested, a probability that the rice root single cell to be tested belongs to a certain rice root single-cell type using a rice root single-cell type annotation model obtained by any of the above methods for constructing a rice root single-cell type annotation model;
[0090] According to the probability that the rice root single cell to be tested belongs to a certain rice root single cell type, the rice root single cell sequencing data of the rice root single cell to be tested is annotated with the rice root single cell type.
[0091] In another aspect, the present invention further provides a non-transitory computer-readable storage medium having a computer program stored thereon, which is configured to execute the following steps when executed by a processor:
[0092] Obtaining rice root single-cell sequencing data of a rice root single cell to be tested, wherein the cell type of the rice root single cell to be tested is unknown;
[0093] Obtaining, based on the rice root single-cell sequencing data of the rice root single cell to be tested, a probability that the rice root single cell to be tested belongs to a certain rice root single-cell type using a rice root single-cell type annotation model obtained by any of the above methods for constructing a rice root single-cell type annotation model;
[0094] According to the probability that the rice root single cell to be tested belongs to a certain rice root single cell type, the rice root single cell sequencing data of the rice root single cell to be tested is annotated with the rice root single cell type.
[0095] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate, and the components shown as units may or may not be physical units, i.e., they may be located in one location or distributed across multiple network units. Some or all of the modules may be selected based on actual needs to achieve the objectives of the present embodiment. Persons of ordinary skill in the art will be able to understand and implement the present invention without inventive effort.
[0096] Through the description of the above embodiments, those skilled in the art can clearly understand that each embodiment can be implemented by means of software plus a necessary general hardware platform, or of course, by hardware. Based on this understanding, the essence of the above technical solution or the part that contributes to the existing technology 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 ROM / RAM, a magnetic disk, an optical disk, etc., and includes a number of instructions for enabling a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the methods described in each embodiment or certain parts of the embodiments.
[0097] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the various embodiments of the present invention.
Claims
1. A method for constructing a single cell type annotation model for rice roots, characterized in that: include: Obtain rice root single-cell sequencing data from different data sources, wherein the rice root single-cell sequencing data is annotated with corresponding rice root single-cell type labels; A double normalization strategy was used to integrate rice root single-cell sequencing data from different data sources to obtain double-normalized rice root single-cell sequencing data. Based on the double-normalized rice root single-cell sequencing data, the gradient boosting decision tree was used to train the model to learn the rice root single-cell sequencing data corresponding to different types of rice root single cells, and obtain the rice root single-cell type annotation model.
2. The method for constructing a rice root single cell type annotation model according to claim 1, characterized in that: The double normalization strategy is used to integrate rice root single-cell sequencing data from different data sources to obtain double-normalized rice root single-cell sequencing data, including: Using a machine learning algorithm for data integration, we performed deep batch correction on rice root single-cell sequencing data from different data sources to eliminate technical variation interference and obtain batch-effect-free rice root single-cell sequencing data. The single-cell sequencing data analysis toolkit was used to perform plant-specific preprocessing on the rice root single-cell sequencing data after eliminating batch effects, and the double-normalized rice root single-cell sequencing data was obtained.
3. The method for constructing a rice root single cell type annotation model according to claim 2, characterized in that: The single-cell sequencing data analysis toolkit is used to perform plant-specific preprocessing on the rice root single-cell sequencing data after eliminating batch effects to obtain double-normalized rice root single-cell sequencing data, including: Based on the rice root single-cell sequencing data after eliminating batch effects, the total expression level of each single cell was normalized to 1×10 4 , obtain normalized rice root single-cell sequencing data; Performing logarithmic transformation on the normalized rice root single-cell sequencing data to obtain logarithmically transformed rice root single-cell sequencing data; Based on the logarithm-transformed rice root single-cell sequencing data, genes with significant variation were screened based on the dispersion-mean distribution of gene expression; The expression levels of genes with significant variation were feature scaled to obtain double-normalized rice root single-cell sequencing data.
4. The method for constructing a rice root single cell type annotation model according to claim 2, wherein: The double normalization strategy is used to integrate rice root single-cell sequencing data from different data sources to obtain double-normalized rice root single-cell sequencing data, including: Clusters with more than 5000 cells were randomly downsampled, while rare cell types were treated with synthetic minority class oversampling.
5. The method for constructing a rice root single cell type annotation model according to any one of claims 1 to 4, characterized in that: The gradient boosting decision tree uses the CatBoost enhancement architecture. The method uses a gradient boosting decision tree to train a model based on the double-normalized rice root single-cell sequencing data to learn the rice root single-cell sequencing data corresponding to different types of rice root single cells, thereby obtaining a rice root single-cell type annotation model, including: Based on the double-normalized rice root single-cell sequencing data, target encoding is performed on the rice root single-cell sequencing data corresponding to different types of rice root single cells, and feature combinations are generated; Based on the feature combination, a symmetric tree is constructed in each round, the negative gradient of the current model is fitted, and the gradient is updated using an ordered boosting strategy; The prediction results of all trees were weighted and combined as the model output to obtain the rice root single cell type annotation model.
6. The method for constructing a rice root single cell type annotation model according to claim 5, characterized in that: The method uses a gradient boosting decision tree to train a model based on the double-normalized rice root single-cell sequencing data to learn the rice root single-cell sequencing data corresponding to different types of rice root single cells, thereby obtaining a rice root single-cell type annotation model, including: The double-normalized rice root single-cell sequencing data were divided into training and test sets at a ratio of 8:2; When training the rice root single cell type annotation model, the model hyperparameters were optimized through ten-fold cross-validation, and the area under the characteristic curve, confusion matrix weighted F1 score, and cross-dataset migration accuracy were used to verify the model performance.
7. A method for annotating single cell types in rice roots, characterized in that: Obtaining rice root single-cell sequencing data of a rice root single cell to be tested, wherein the cell type of the rice root single cell to be tested is unknown; Obtaining, based on rice root single-cell sequencing data of the rice root single cell to be tested, a probability that the rice root single cell to be tested belongs to a certain rice root single-cell type using a rice root single-cell type annotation model obtained by the method for constructing a rice root single-cell type annotation model according to any one of claims 1 to 6; According to the probability that the rice root single cell to be tested belongs to a certain rice root single cell type, the rice root single cell sequencing data of the rice root single cell to be tested is annotated with the rice root single cell type.
8. A rice root single cell type annotation system, characterized in that: include: A data acquisition module is used to: acquire rice root single-cell sequencing data of a rice root single cell to be tested, wherein the cell type of the rice root single cell to be tested is unknown; a prediction module for: obtaining, based on rice root single-cell sequencing data of a rice root single-cell to be tested and a rice root single-cell type annotation model obtained by the method for constructing a rice root single-cell type annotation model according to any one of claims 1 to 6, a probability that the rice root single-cell to be tested belongs to a certain rice root single-cell type; The annotation module is used to annotate the rice root single cell type of the rice root single cell sequencing data of the rice root single cell to be tested according to the probability that the rice root single cell to be tested belongs to a certain rice root single cell type.
9. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein: When the processor executes the program, the method for constructing a rice root single cell type annotation model according to any one of claims 1 to 6 and / or the method for annotating rice root single cell types according to claim 7 is implemented.
10. A non-transitory computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the method for constructing a rice root single cell type annotation model according to any one of claims 1 to 6 and / or the method for annotating rice root single cell types according to claim 7 is implemented.
