Crop growth state monitoring method, system, equipment and medium
By constructing a crop growth status monitoring model based on high variable genes, the problem of difficult to identify factors affecting crop abnormal phenotypes in the prior art is solved, and precise monitoring of crop growth status and improvement of agricultural management are achieved.
Patent Information
- Application Number
- CN202411416579.X
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-10-11
- Publication Date
- 2025-02-07
- Estimated Expiration
- Not applicable · inactive patent
AI Technical Summary
The prior art is difficult to accurately identify the influencing factors that lead to abnormal crop phenotypes, and the lack of explanatory and versatility, limiting the application value of precise crop management.
By obtaining the crop gene expression data set, an independent data subset is formed according to the sample attribute classification, random sampling is used to form a training set, the error discovery rate FDR value is calculated, and a high variable gene set is filtered to obtain, and a machine learning model is constructed for crop growth status monitoring.
Accurate monitoring of crop growth status is achieved, providing biological explanatory and rapid screening of highly variable genes, and improving the accuracy and versatility of agricultural management.
Smart Images
Figure CN119400248A_ABST
Abstract
Description
Technical Field
[0001] The present invention belongs to the field of bioinformatics, and in particular relates to a crop growth status monitoring method, system, equipment and medium. Background Art
[0002] In recent years, with global climate change, extreme weather has occurred frequently, and food security issues have become increasingly prominent. In order to further improve the efficiency of crop growth status monitoring, machine learning and artificial intelligence technologies have been introduced into crop production safety detection. In agricultural practice, a certain abnormal phenotype of crops may come from a variety of situations. For example, yellowing of plant leaves may be caused by insufficient light, water and fertilizer problems, abnormal ambient temperature, and pests and diseases.
[0003] Existing technologies mainly use field plant phenotypic data to model and predict plant growth status. However, this technology can only be used to distinguish whether the plant growth status is normal. If it is not combined with other environmental factors, crop physiological index measurement and other methods, it is difficult to identify the influencing factors that cause its abnormal phenotypic traits. Therefore, in agricultural practice, the application value of this method for precise crop management is very limited. Some studies screened some differentially expressed genes from a few transcriptome sequencing samples and used them to predict crop phenotypes. However, this method is often sensitive to the selected samples, making the constructed model lack generalization ability. With the development of high-throughput sequencing technology and the reduction of sequencing costs, a more advanced method is to perform feature screening on large-scale gene expression profiles under different conditions of a certain crop. However, this method is very sensitive to the selection of feature screening thresholds and requires a large amount of statistical analysis of gene expression data. It is not easy to implement for people who lack statistical knowledge. At the same time, the selected genes are only statistically meaningful and lack interpretability in biological terms. There are also some phenotype-based machine learning models that can distinguish whether the crop phenotype is abnormal or not, but cannot further explain the molecular mechanism of abnormal phenotypes, so they cannot provide accurate agricultural decision-making and guidance. In addition, although some studies have attempted to evaluate growth status through differentially expressed genes obtained from individual studies, this method is less efficient and lacks universality for different crops, considering the complex changes in gene expression of crops in different environments and the impact of crop genetic diversity on gene expression. Summary of the invention
[0004] In order to solve the problem of applying gene expression data in crop phenotype and growth status monitoring, the present invention provides a crop growth status monitoring method and system.
[0005] In order to achieve the above object, the present invention provides the following technical solutions:
[0006] A method for monitoring crop growth status comprises the following steps:
[0007] Acquire a crop gene expression data set; classify the crop gene expression data set according to sample attributes to form different independent data subsets;
[0008] Randomly sampling each of the independent data subsets in proportion to form a training set, calculating the false discovery rate FDR value of the data in the training set, filtering the false discovery rate FDR value to obtain a highly variable gene set;
[0009] Using the sample data in the training set as the response variable and the highly variable gene set as the predictor variable, training the monitoring machine learning model to obtain a crop growth status monitoring model;
[0010] The crop data to be detected is input into the crop growth status monitoring model to predict the crop phenotype, and the growth status of the crop is obtained according to the crop phenotype.
[0011] Preferably, the monitoring machine learning model includes a support vector machine model, an extreme gradient boosting XGboost model, a deep neural network model, a random forest model and a naive Bayes model.
[0012] Preferably, the false discovery rate FDR value is filtered to obtain a highly variable gene set, specifically by fitting the false discovery rate FDR value of each gene with the average expression of the gene, and finally selecting genes that significantly deviate from the fitting line as the highly variable gene set.
[0013] Preferably, it also includes evaluating the accuracy of the monitoring machine learning model by drawing a receiver operating characteristic curve (ROC), that is, drawing a false positive and false negative correlation curve, and then calculating the AUC value to evaluate the accuracy of the model.
[0014] Preferably, the gene expression data set is obtained by collecting RNA-Seq data from NCBI and PlantExp public databases, and converting the RNA-Seq data into TMP gene expression levels.
[0015] Preferably, the crop gene expression dataset is classified according to sample attributes, and specific classification criteria include tissue type, different varieties or strains, developmental stage and environmental stress.
[0016] The present invention provides a crop growth status monitoring system, which specifically comprises:
[0017] The data module is used to obtain a crop gene expression data set; classify the crop gene expression data set according to sample attributes to form different independent data subsets.
[0018] The training module is used to randomly sample each of the independent data subsets in proportion to form a training set, calculate the false discovery rate FDR value of the data in the training set, and filter the false discovery rate FDR value to obtain a highly variable gene set.
[0019] The machine learning module is used to use the sample data in the training set as the response variable and the highly variable gene set as the predictor variable to train the monitoring machine learning model to obtain a crop growth status monitoring model.
[0020] The monitoring module is used to input the crop data to be detected into the crop growth status monitoring model to predict the crop phenotype and obtain the crop growth status according to the crop phenotype.
[0021] The present invention also provides a computer device, comprising a memory, a processor and a computer program stored in the memory, wherein the processor executes the computer program to implement the steps described in the crop growth status monitoring method.
[0022] The present invention also provides a computer-readable storage medium, on which a computer program is stored. When the computer program is loaded by a processor, the steps described in the crop growth status monitoring method can be executed.
[0023] The crop growth status monitoring method provided by the present invention has the following beneficial effects:
[0024] The present invention proposes to obtain crop gene expression data sets from a large-scale gene expression database, and classify them according to sample types to form different independent data subsets, considering the impact of gene data on plant growth status. Each independent data subset is randomly sampled in proportion to form a training set, the gene expression variation of the training set is calculated, and the data is filtered to obtain a highly variable gene set. The obtained gene data is not only highly interpretable in a biological sense, but also can quickly select a meaningful characteristic gene set. A crop growth status monitoring model is constructed using highly variable genes as predictive variables, and the growth status of crops is monitored based on gene expression data. While intelligently detecting crop growth, agricultural resources are saved, and agricultural management and application value can be guided. BRIEF DESCRIPTION OF THE DRAWINGS
[0025] In order to more clearly illustrate the embodiment of the present invention and its design scheme, the following briefly introduces the drawings required for this embodiment. The drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.
[0026] Figure 1Schematic diagram of the modeling and evaluation process of corn genes in Example 1 of the present invention. Figure 1 A is the original maize gene expression matrix obtained. Figure 1 B is the classification of the data set, Figure 1 C is a schematic diagram of the modeling process using a specific data set.
[0027] Figure 2 The prediction and evaluation results of three machine learning models based on three data sets. Figure 2 A. Prediction accuracy of five machine learning models for corn tissue types, Figure 2 B Confusion matrix of the prediction results of the maize tissue type identification model, Figure 2 The prediction accuracy of five machine learning models for corn development period, Figure 2 The D confusion matrix shows the prediction results of the corn development period prediction model. Figure 2 The prediction accuracy of five machine learning models for the stress type of corn growth environment, Figure 2 The F confusion matrix shows the prediction results of the stress type of the maize growth environment.
[0028] Figure 3 The radar chart shows the prediction accuracy of the five machine learning models in different rice tissue types, growth and development stages, different strain types, and stress response types.
[0029] Figure 4 The present invention is a flow chart of a crop growth status monitoring method. DETAILED DESCRIPTION
[0030] In order to enable those skilled in the art to better understand the technical solution of the present invention and implement it, the present invention is described in detail below in conjunction with the accompanying drawings and specific embodiments. The following embodiments are only used to more clearly illustrate the technical solution of the present invention, and cannot be used to limit the scope of protection of the present invention.
[0031] Example 1
[0032] The present invention provides a crop growth status monitoring method. In this embodiment, corn is taken as an example. The growth status of corn is predicted by establishing large-scale gene expression data of corn for machine learning modeling. Figure 1 As shown, it mainly includes the following steps:
[0033] S1. Large-scale transcriptome data collection of maize: 21,612 maize RNA-Seq sample information and the corresponding expression value (TPM) of each gene of maize were collected from the NCBI and PlantEXP public databases to form a gene-sample expression matrix, such as Figure 1 As shown in A.
[0034] S2. Classify the gene expression profiles according to sample attributes to form different independent data subsets, including classification according to tissue type (root, stem, leaf, flower, fruit, seed, endosperm, etc.), classification according to developmental period (V1, V2, V3, V4, etc.), and classification according to environmental stress type (low temperature stress, drought stress, high temperature stress, and normal conditions, etc.) to form data subsets I-III respectively, such as Figure 1 As shown in B.
[0035] S3. Randomly sample the above different data subsets according to the ratio of 7:3 to form training sets and test sets respectively. Then, use the Brennecke_getVariableGenes function in the R package M3Drop to calculate the FDR value of the above training sets, and filter the highly variable gene set through the FDR value to form a highly variable gene expression matrix, such as Figure 1 As shown in C.
[0036] S4. Based on the training set and test set obtained from each of the above data subsets, five machine learning models (support vector machine, xgboost, deep neural network, random forest and naive Bayes, etc.) were built with the sample phenotype in the training set as the response variable or prediction target, and the expression of the highly variable gene set as the prediction variable, and then the attributes of the samples were predicted and the model was evaluated in the test set. The support vector machine model uses the svm function in the R language extension library; the XGboost model is built using the xgboost function in the R language extension library; the deep neural network model is built using the Keras and tensorflow extension packages in the R language extension library; the random forest model is built using the randomRorest function in the R language extension library; and the naive Bayes model is built using the naiveBayes function in the R language extension library. After the model is built, the test set is predicted, and then the confusion matrix is calculated using the confus ionMatrix function in the R package caret; the AUC value is calculated using the R package pROC to evaluate the accuracy of the model.
[0037] S5. By modeling the data subset I, a classifier for predicting corn tissue type can be obtained; by modeling the data set II, a classifier for identifying corn development period can be obtained; by modeling the data set III, a classifier for identifying corn stress type can be obtained. The model prediction and evaluation results are shown in Figure 2. Figure 2 As shown in Figure 1, the prediction and evaluation results of three machine learning models based on three data sets are obtained. According to the corresponding model established for each data subset, the prediction of crop phenotype is output, and the growth status of the crop is obtained based on the crop phenotype.
[0038] Example 2
[0039] The growth status of rice is predicted by machine learning modeling based on large-scale rice gene expression data, including the following steps:
[0040] S1. Large-scale rice transcriptome data collection: 9965 rice RNA-Seq sample information and the corresponding expression value (TPM) of each gene in corn were collected from the NCBI and PlantExp public databases to form a gene-sample expression matrix.
[0041] S2. Classify the gene expression profiles according to sample attributes to form different independent data subsets, including classification by tissue type (roots, stems, leaves, flowers, fruits, seeds, endosperm, etc.), classification by developmental period (V1, V2, V3, V4, etc.), classification by environmental stress type (low temperature stress, drought stress, high temperature stress, and normal conditions, etc.), and classification by variety and strain to form data subsets I-IV.
[0042] S3. Randomly sample the above different data subsets to form training sets and test sets, and then perform feature screening on the above training sets. The screening method is based on the degree of gene expression variation.
[0043] S4. Based on the training and test sets obtained from each of the above data subsets, five machine learning models (support vector machine, xgboost, deep neural network, random forest, and naive Bayes, etc.) were used to model the sample phenotype in the training set as the response variable or prediction target, and the expression level of the highly variable gene set as the predictor variable. Then, the attributes of the samples in the test set were predicted and the model was evaluated.
[0044] S5. By modeling the data subset I, a classifier for predicting rice tissue types can be obtained; by modeling the data set II, a classifier for identifying rice development periods can be obtained; by modeling the data set III, a classifier for identifying rice environmental stress types can be obtained; by modeling the data set IV, a prediction model for information on different rice varieties can be obtained. Except for individual models, the overall prediction accuracy and AUC value are as high as 95%. Figure 3 The radar chart shows the prediction accuracy of the five machine learning models in different rice tissue types, growth and development stages, different strain types, and stress response types.
[0045] The present invention also includes a crop phenotype and growth status monitoring system, comprising:
[0046] The data module is used to obtain the crop gene expression data set; the crop gene expression data set is classified according to sample attributes to form different independent data subsets.
[0047] The training module is used to randomly sample each independent data subset in proportion to form a training set, calculate the false discovery rate FDR value of the data in the training set, filter the false discovery rate FDR value, and obtain a highly variable gene set.
[0048] The machine learning module is used to train the monitoring machine learning model by taking the sample data in the training set as the response variable and the highly variable gene set as the predictor variable to obtain the crop growth status monitoring model.
[0049] The monitoring module is used to input the crop data to be detected into the crop growth status monitoring model to predict the crop phenotype and obtain the crop growth status according to the crop phenotype.
[0050] Each module in the above crop growth status monitoring system can be implemented in whole or in part by software, hardware, or a combination thereof. Each module can be embedded in or independent of a processor in a computer device in the form of hardware, or can be stored in a memory in a computer device in the form of software, so that the processor can call and execute operations corresponding to each module.
[0051] The present invention also provides a computer device, including a memory, a processor and a computer program stored in the memory, wherein the processor executes the computer program to implement the steps in an embodiment of a crop growth status monitoring method. The specific implementation method can be found in the method embodiment, which will not be described in detail here.
[0052] Furthermore, the present invention also provides a non-temporary computer-readable storage medium containing instructions, and a computer program is stored on the storage medium. For example, a memory containing instructions, the above instructions can be executed by a processor of a computer device to complete the above method. For example, a non-temporary computer-readable storage medium can be a ROM, a random access memory (RAM), a CD-ROM, a magnetic tape, a floppy disk, and an optical data storage device. When the computer program is executed by the processor, the steps in an embodiment of a crop growth status monitoring method can be implemented. The specific implementation method can be found in the method embodiment, which will not be repeated here.
[0053] It will be appreciated by those skilled in the art that embodiments of the present invention may provide methods, systems or computer program products. Therefore, the present invention may take the form of a complete hardware embodiment, a complete software embodiment or an embodiment combining software and hardware. Moreover, the present invention may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.
[0054] The present invention is described with reference to flowcharts and / or block diagrams of methods, apparatus (systems) and computer program products according to embodiments of the present invention. It should be understood that each process and / or block in the flowchart and / or block diagram, as well as a combination of processes and / or blocks in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.
[0055] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.
[0056] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.
[0057] It should be pointed out that the specific implementation methods described above can enable those skilled in the art to understand the invention more comprehensively, but do not limit the invention in any way. Therefore, although the invention has been described in detail in this specification and embodiments, those skilled in the art should understand that the invention can still be modified or replaced by equivalents; and all technical solutions and improvements that do not deviate from the spirit and scope of the invention are included in the protection scope of the patent for the invention. Any figure mark in the claims should not be regarded as limiting the claims involved. Any simple change or equivalent replacement of the technical solution that can be obviously obtained by any technician familiar with the field within the technical scope disclosed in the present invention belongs to the protection scope of the present invention.
Claims
1. A method for monitoring crop growth status, characterized in that: The following steps are involved: Acquire a crop gene expression data set; classify the crop gene expression data set according to sample attributes to form different independent data subsets; Randomly sampling each of the independent data subsets in proportion to form a training set, calculating the false discovery rate FDR value of the data in the training set, filtering the false discovery rate FDR value to obtain a highly variable gene set; Using the sample data in the training set as the response variable and the highly variable gene set as the predictor variable, training the monitoring machine learning model to obtain a crop growth status monitoring model; The crop data to be detected is input into the crop growth status monitoring model to predict the crop phenotype, and the growth status of the crop is obtained according to the crop phenotype.
2. A crop growth status monitoring method according to claim 1, characterized in that: The monitoring machine learning models include support vector machine models, extreme gradient boosting XGboost models, deep neural network models, random forest models and naive Bayes models.
3. A crop growth status monitoring method according to claim 1, characterized in that: The false discovery rate FDR value is filtered to obtain a highly variable gene set, specifically by fitting the false discovery rate FDR value of each gene with the average gene expression, and finally selecting genes that significantly deviate from the fitting line as the highly variable gene set.
4. A crop growth status monitoring method according to claim 2, characterized in that: It also includes evaluating the accuracy of the monitoring machine learning model by drawing a receiver operating characteristic curve (ROC), that is, drawing a false positive and false negative correlation curve, and then calculating the AUC value to evaluate the accuracy of the model.
5. A crop growth status monitoring method according to claim 1, characterized in that: The gene expression data set is obtained by collecting RNA-Seq data from NCBI and PlantExp public databases, and converting the RNA-Seq data into TMP gene expression levels.
6. A crop growth status monitoring method according to claim 1, characterized in that: The crop gene expression dataset is classified according to sample attributes, and specific classification criteria include tissue type, different varieties or strains, developmental period and environmental stress.
7. A crop growth status monitoring system, characterized in that: include: Data module, used to obtain crop gene expression datasets; Classifying the crop gene expression dataset according to sample attributes to form different independent data subsets; A training module, used for randomly sampling each of the independent data subsets in proportion to form a training set, calculating the false discovery rate FDR value of the data in the training set, filtering the false discovery rate FDR value, and obtaining a highly variable gene set; A machine learning module, used to train a monitoring machine learning model using the sample data in the training set as a response variable and the highly variable gene set as a predictor variable to obtain a crop growth status monitoring model; The monitoring module is used to input the crop data to be detected into the crop growth status monitoring model to predict the crop phenotype and obtain the crop growth status according to the crop phenotype.
8. A computer device comprising a memory, a processor and a computer program stored in the memory, characterized in that: The processor executes the computer program to implement the steps of the method according to any one of claims 1 to 6.
9. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is loaded into a processor, it can execute the steps of the method according to any one of claims 1 to 6.