Intelligent breeding system for hybrid rice

Abstract

The invention relates to the crossing field of agricultural biotechnology and information technology, and discloses a smart breeding system for hybrid rice, comprising: a smart breeding data acquisition and processing system for collecting and processing genotype data, environmental factors and seed phenotype data of rice parents and offspring; the intelligent breeding model parallel optimization system is used for optimizing the parallel calculation process, and the invention further discloses a hybrid rice breeding method of the intelligent breeding system, and the method comprises the following steps: collecting rice genotype, environmental factor and seed phenotype data; processing and analyzing data by using the intelligent breeding model, and training a prediction model; predicting hybrid rice characters through a prediction system, and generating a report; and formulating a breeding scheme according to the report and performing field verification. By integrating an advanced information technology, comprehensive intelligentization and high efficiency of hybrid rice breeding are realized, the breeding period is shortened, the cost is reduced, and the system performance is continuously improved through a closed-loop iterative optimization mechanism.

Description

Technical Field

[0001] The present invention relates to the intersection of agricultural biotechnology and information technology, and specifically to an intelligent breeding system for hybrid rice. Background Art

[0002] In the development of modern agricultural science, the innovation of crop breeding technology is one of the key factors in promoting the improvement of agricultural productivity. With the rapid development of cutting-edge technologies such as genomics, bioinformatics, big data analysis and artificial intelligence, crop breeding is gradually shifting from traditional phenotypic selection to precision breeding based on molecular genetic information. As one of the most important food crops in the world, the innovation of rice breeding technology is of great significance to increasing food production and ensuring global food security.

[0003] Traditional rice breeding methods mainly rely on field trials and phenotypic selection, which is not only long-term and inefficient, but also difficult to accurately capture complex genetic variations and traits controlled by multiple genes. In addition, due to the uncertainty of genetic recombination of hybrid offspring, traditional breeding faces many challenges when utilizing hybrid vigor, such as the difficulty in accurately predicting the trait performance of hybrid offspring. Therefore, in view of the above shortcomings, the present invention proposes an intelligent breeding system for hybrid rice. Summary of the invention

[0004] In view of the shortcomings of the prior art, the present invention provides an intelligent breeding system for hybrid rice, which solves the problems of low efficiency, difficulty in accurately capturing complex genetic variations and inaccurate prediction of hybrid offspring traits in the traditional hybrid rice breeding process.

[0005] To achieve the above objectives, the present invention is implemented through the following technical solutions: a smart breeding system for hybrid rice, comprising:

[0006] Intelligent breeding data collection and processing system, which is used to collect and process genotype data, environmental factors and seed phenotype data of rice parents and offspring;

[0007] Intelligent breeding model parallel optimization system, which is used to optimize the parallel computing process and improve the training efficiency and stability of the intelligent breeding model;

[0008] The rice intelligent breeding prediction system is used to build and evaluate intelligent breeding prediction models and improve the accuracy of phenotypic prediction;

[0009] Intelligent breeding supercomputing cloud-native open innovation platform, which is used to integrate and manage the computing resources, algorithms and data required for breeding.

[0010] Preferably, the intelligent breeding data collection and processing system includes:

[0011] A genotype data collection unit, which is used to collect genotype data of rice parents and progeny based on the indica rice core site liquid phase 10K breeding chip and deep sequencing technology;

[0012] A phenotypic data collection unit, which is used to collect rice phenotypic data through field intelligent environment collection equipment and seed testing equipment;

[0013] The data processing unit is used to clean, organize and integrate the collected data.

[0014] Preferably, the intelligent breeding model parallel optimization system comprises:

[0015] Parallel computing environment unit, which is used to build a parallel computing environment on the domestically produced heterogeneous multi-core supercomputer system and configure the necessary software and hardware resources;

[0016] Data segmentation and storage unit, which is used to segment large data sets into small blocks and assign them to different computing nodes for parallel processing;

[0017] The model training unit is used to implement a parallel gradient algorithm for transfer learning based on incremental iterations, improve model training efficiency, and monitor data consistency and model convergence.

[0018] Preferably, the rice intelligent breeding prediction system comprises:

[0019] A prediction model building unit, which is used to build an intelligent breeding prediction model based on the attention mechanism;

[0020] Data processing unit, which is used to clean, complete and feature-screen the original data to improve data quality;

[0021] A model training and evaluation unit, which is used to use the processed data to train the model and conduct a comprehensive evaluation of the model;

[0022] The prediction result output unit is used to display the prediction results through the user interaction interface and generate a prediction report.

[0023] Preferably, the intelligent breeding supercomputing cloud native open innovation platform includes:

[0024] Platform architecture unit, which is used to design platform architecture using microservice architecture and container technology to achieve decoupling and independent deployment of services;

[0025] Resource management and scheduling unit, which is used to implement model and data adaptation strategies, communication optimization strategies and serverless framework;

[0026] Security control and permission management unit, which is used to control access to and protect platform resources;

[0027] The user interface is used to display prediction results, receive user input, and provide operation guidance.

[0028] The method for hybrid rice breeding using the smart breeding system comprises the following steps:

[0029] S1. Collect genotype data, environmental factors and seed phenotype data of rice parents and offspring using the intelligent breeding data collection and processing system;

[0030] S2. Use the intelligent breeding model parallel optimization system to process and analyze the collected data and train the intelligent breeding prediction model;

[0031] S3. Predict the characteristics of hybrid rice through the rice intelligent breeding prediction system and generate a prediction report

[0032] S4. Formulate a hybrid rice breeding plan based on the forecast report, and implement and verify the optimal plan in the field. The S1 step specifically includes the following steps:

[0033] S1.1. Genotype data collection:

[0034] Genotype data were collected using the indica rice core site liquid phase 10K breeding chip;

[0035] Obtain SNP information of rice parents and offspring through whole genome resequencing technology;

[0036] S1.2. Phenotypic data collection:

[0037] Use intelligent field environment collection equipment to collect data on environmental factors during rice growth, including temperature, humidity, and light intensity;

[0038] Use seed testing equipment, such as automatic seed testing machines and image recognition systems, to measure and record rice phenotypic characteristics, such as plant height, number of panicles, growth period, 1,000-grain weight, number of tillers, and seed setting rate;

[0039] S1.3 Data processing:

[0040] Perform quality control on the collected genotypic and phenotypic data and remove low-quality, missing or abnormal data points;

[0041] Integrate environmental parameters and phenotypic data to ensure data accuracy and consistency;

[0042] Data storage and management form a unified database to provide basic data for subsequent analysis.

[0043] Preferably, the step S2 specifically includes the following steps:

[0044] S2.1, Parallel computing environment:

[0045] Build a parallel computing environment on a domestically produced heterogeneous multi-core supercomputer system and configure the necessary software and hardware resources;

[0046] S2.2, Data segmentation and storage:

[0047] Divide large-scale data sets into small pieces and assign them to different computing nodes for parallel processing;

[0048] Adopt efficient data storage and access strategies to ensure fast data transmission and access during parallel computing;

[0049] S2.3, Model training:

[0050] Implement a parallel gradient algorithm for transfer learning based on incremental iteration to accelerate the model training process;

[0051] Monitor data consistency and model convergence during training to ensure the accuracy and stability of training results;

[0052] S2.4, Optimization:

[0053] Optimize inter-node communication strategies and data storage methods to reduce communication delays and memory consumption.

[0054] Preferably, the S3 step specifically includes the following contents:

[0055] S3.1. Prediction model construction:

[0056] Build an intelligent breeding prediction model based on the attention mechanism to model the mutual influence between genes;

[0057] S3.2 Data processing:

[0058] Clean, complete and feature-screen the original data to improve data quality;

[0059] S3.3, Model training and evaluation:

[0060] Use the processed data to train the prediction model and adjust the model parameters to optimize the prediction performance;

[0061] Develop a full-process model evaluation method for machine learning, including cross-validation and performance evaluation indicators, to comprehensively evaluate the predictive model;

[0062] S3.4. Prediction result output

[0063] The prediction results are displayed through the user interface, including the predicted values ​​and confidence intervals of hybrid rice traits;

[0064] Generate prediction reports to facilitate breeders to review and analyze prediction results.

[0065] Preferably, the step S4 specifically includes the following steps:

[0066] S4.1. Develop breeding plan:

[0067] According to the forecast report and combined with actual breeding needs, formulate a detailed hybrid rice breeding plan;

[0068] S4.2. Field implementation:

[0069] Implement hybrid rice breeding programs in the field and monitor various breeding operations;

[0070] S4.3, Verification and Adjustment:

[0071] Monitor and record breeding programs implemented in the field, and dynamically adjust and optimize breeding programs based on test results;

[0072] Collect data generated during the implementation process to provide basic data for subsequent analysis and optimization.

[0073] The present invention provides an intelligent breeding system for hybrid rice, which has the following beneficial effects:

[0074] 1. The present invention realizes comprehensive intelligence and efficiency of hybrid rice breeding by integrating advanced information technology. The intelligent breeding data acquisition and processing system can efficiently and accurately collect and process large-scale genotype, phenotype and environmental data, providing comprehensive data support for breeding decisions.

[0075] 2. The parallel optimization method of the intelligent breeding model of the present invention significantly reduces communication delay and memory consumption by optimizing communication strategies and data processing methods, accelerates the model training process, and thus shortens the breeding cycle.

[0076] 3. The rice intelligent breeding prediction system of the present invention uses the attention mechanism to model the mutual influence between genes, and conducts a comprehensive evaluation through a machine learning full-process model evaluation method to ensure the accuracy and reliability of phenotypic prediction.

[0077] 4. The present invention reduces the experimental and labor costs in traditional breeding methods through an intelligent and efficient breeding process. At the same time, the intelligent breeding supercomputing cloud-native open innovation platform adopts microservice architecture and container technology to achieve on-demand allocation, elastic expansion and efficient utilization of resources, thereby improving the development efficiency and quality of intelligent breeding software.

[0078] 5. The present invention forms a closed-loop iterative optimization mechanism of "data-feature-model-evaluation", which improves the accuracy of prediction and the overall performance of the system by continuously optimizing the model and data processing. BRIEF DESCRIPTION OF THE DRAWINGS

[0079] Figure 1 It is a framework diagram of the system of the present invention;

[0080] Figure 2 It is a process framework diagram of the system of the present invention;

[0081] Figure 3 It is a framework diagram of the intelligent breeding data collection and processing system of the present invention;

[0082] Figure 4 It is a framework diagram of the parallel optimization system of the intelligent breeding model of the present invention;

[0083] Figure 5 It is a framework diagram of the rice intelligent breeding prediction system of the present invention;

[0084] Figure 6 This is a framework diagram of the intelligent breeding supercomputing cloud native open innovation platform of the present invention;

[0085] Figure 7 Flow chart of the method of the present invention. DETAILED DESCRIPTION

[0086] The technical solution of the present invention will be described clearly and completely below in conjunction with the accompanying drawings of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0087] Example:

[0088] Please see attached Figure 1 -Attached Figure 2 The embodiment of the present invention provides a smart breeding system for hybrid rice, comprising:

[0089] Intelligent breeding data collection and processing system, which is used to collect and process genotype data, environmental factors and seed phenotype data of rice parents and offspring;

[0090] Intelligent breeding model parallel optimization system, which is used to optimize the parallel computing process and improve the training efficiency and stability of the intelligent breeding model;

[0091] The rice intelligent breeding prediction system is used to build and evaluate intelligent breeding prediction models and improve the accuracy of phenotypic prediction;

[0092] Intelligent breeding supercomputing cloud-native open innovation platform, which is used to integrate and manage the computing resources, algorithms and data required for breeding.

[0093] Specifically, the intelligent breeding data collection and processing system can efficiently and accurately collect and process large-scale genotype, phenotype and environmental data, providing comprehensive data support for breeding decisions.

[0094] Effectively lay a solid foundation for subsequent intelligent analysis and ensure the accuracy and consistency of data.

[0095] It solves the problem that traditional breeding methods rely on field trials and phenotypic selection, have long cycles, low efficiency, and difficulty in accurately capturing complex genetic variations.

[0096] Specific solutions:

[0097] Genotype data collection: The genotype data of rice parents and offspring were collected using the indica rice core site liquid phase 10K breeding chip and deep sequencing technology.

[0098] Phenotypic data collection: Collect environmental factors and phenotypic data of rice through intelligent field environment collection equipment and seed testing equipment.

[0099] Data processing: quality control, cleaning, organization and integration of collected data to form a unified database.

[0100] The intelligent breeding model parallel optimization system effectively reduces communication delays and memory consumption, and significantly improves the training efficiency and stability of the intelligent breeding model.

[0101] Ensure data consistency and model convergence by accelerating the model training process.

[0102] It solves the problem that traditional methods of large-scale data processing and model training are inefficient and cannot meet the needs of efficient breeding.

[0103] Specific solutions:

[0104] Parallel computing environment: Build a parallel computing environment on a domestically produced independent heterogeneous multi-core supercomputer system and configure the necessary software and hardware resources.

[0105] Data segmentation and storage: large-scale data sets are divided into small pieces and assigned to different computing nodes for parallel processing.

[0106] Model training: Implement a parallel gradient algorithm for transfer learning based on incremental iterations, and monitor data consistency and model convergence in real time during the training process.

[0107] The rice intelligent breeding prediction system effectively improves the accuracy of phenotypic prediction and ensures the reliability of prediction results. It builds and evaluates intelligent breeding prediction models to provide a scientific basis for breeding decisions.

[0108] It solves the problem that traditional methods are difficult to accurately predict the trait performance of hybrid offspring.

[0109] Specific solutions:

[0110] Prediction model construction: Build an intelligent breeding prediction model based on the attention mechanism to model the interaction between genes.

[0111] Data processing: Clean, complete and feature-screen the raw data to improve data quality.

[0112] Model training and evaluation: Use the processed data to train the model and develop a full-process model evaluation method for machine learning to comprehensively evaluate the predictive model.

[0113] Forecast result output: The forecast results are displayed through the user interaction interface and a forecast report is generated.

[0114] The intelligent breeding supercomputing cloud-native open innovation platform realizes on-demand allocation, elastic expansion and efficient utilization of resources, and improves the development efficiency and quality of intelligent breeding software.

[0115] Integrate and manage the computing resources, algorithms and data required for breeding to support diverse breeding application scenarios.

[0116] It solves the problem that traditional platforms have low resource management efficiency and are difficult to adapt to the needs of intelligent breeding.

[0117] Specific solutions:

[0118] Platform architecture: The platform architecture is designed using microservice architecture and container technology to achieve decoupling and independent deployment of services.

[0119] Resource management and scheduling: Implement model and data adaptation strategies, communication optimization strategies, and serverless frameworks to ensure efficient use of resources.

[0120] Security control and permission management: Perform access control and security protection on platform resources to ensure the security of platform operation.

[0121] Please see attached Figure 3 , the intelligent breeding data collection and processing system includes:

[0122] A genotype data collection unit, which is used to collect genotype data of rice parents and progeny based on the indica rice core site liquid phase 10K breeding chip and deep sequencing technology;

[0123] A phenotypic data collection unit, which is used to collect rice phenotypic data through field intelligent environment collection equipment and seed testing equipment;

[0124] The data processing unit is used to clean, organize and integrate the collected data.

[0125] Specifically, the genotype data collection unit can efficiently and accurately obtain the genotype data of rice parents and offspring through the indica rice core site liquid phase 10K breeding chip and deep sequencing technology.

[0126] Ensure the comprehensiveness and accuracy of genotype data and provide a reliable data basis for subsequent analysis and breeding decisions.

[0127] It solves the problem that traditional breeding methods are inefficient in collecting genotype data and have difficulty in capturing genetic information comprehensively and accurately.

[0128] Specific solutions:

[0129] The key loci in the rice genome were determined using a high-coverage breeding array.

[0130] A large amount of genetic variation information such as SNPs (single nucleotide polymorphisms) is obtained through deep sequencing technology.

[0131] The phenotypic data collection unit automatically collects rice phenotypic data through intelligent equipment, improving the efficiency and accuracy of data collection.

[0132] Provide high-quality phenotypic data and provide a reliable basis for breeding analysis.

[0133] It solves the problems of cumbersome data collection process, high labor cost and low accuracy in traditional field trials.

[0134] Specific solutions:

[0135] Use intelligent field environment collection equipment (such as weather stations and soil moisture sensors) to monitor environmental parameters in real time.

[0136] Use seed testing equipment (such as automatic seed testing machines, image recognition systems) to measure and record rice phenotypic characteristics.

[0137] The data processing unit ensures the accuracy and consistency of data through data cleaning, organization and integration, providing a high-quality data foundation for intelligent analysis.

[0138] Improve the quality and availability of data and reduce the impact of data noise and outliers.

[0139] It solves the problem that traditional methods have low data quality, a large amount of noise and missing values, which affect the accuracy of analysis results.

[0140] Specific solutions:

[0141] The collected genotypic and phenotypic data were quality controlled to remove low-quality, missing or abnormal data points.

[0142] Clean and organize data to ensure data integrity and consistency.

[0143] Integrate environmental parameters and phenotypic data to form a unified database.

[0144] Please see attached Figure 4 , the intelligent breeding model parallel optimization system includes:

[0145] Parallel computing environment unit, which is used to build a parallel computing environment on the domestically produced heterogeneous multi-core supercomputer system and configure the necessary software and hardware resources;

[0146] Data segmentation and storage unit, which is used to segment large data sets into small blocks and assign them to different computing nodes for parallel processing;

[0147] The model training unit is used to implement a parallel gradient algorithm for transfer learning based on incremental iterations, improve model training efficiency, and monitor data consistency and model convergence.

[0148] Specifically, the parallel computing environment unit improves computing efficiency and system stability by building a parallel computing environment on a domestically produced independent heterogeneous multi-core supercomputing system.

[0149] Supports large-scale data processing and complex model training, and improves computing performance.

[0150] It solves the problem that traditional single-node computing power is limited and difficult to handle large-scale breeding data and complex models.

[0151] Specific solutions:

[0152] Configure necessary software and hardware resources to build a high-performance parallel computing environment.

[0153] Design and optimize inter-node communication strategies to reduce communication delays and conflicts.

[0154] The data segmentation and storage unit divides large-scale data sets into small blocks and assigns them to different computing nodes for parallel processing to improve data processing efficiency.

[0155] Accelerate the data processing process and ensure the efficiency of data transmission and storage.

[0156] It solves the problems of low efficiency in large-scale data processing and slow data transmission and storage speeds.

[0157] Specific solutions:

[0158] Divide large-scale data sets into small pieces and assign them to different computing nodes for parallel processing.

[0159] Adopt efficient data storage and access strategies to ensure fast transmission and access of data during parallel computing.

[0160] The model training unit accelerates the model training process and improves the efficiency and stability of model training through a parallel gradient algorithm based on incremental iteration and transfer learning.

[0161] Improve the training speed and accuracy of the model and ensure the convergence and consistency of the model.

[0162] It solves the problems of long training time, low efficiency and difficulty in quickly responding to breeding needs of traditional models.

[0163] Specific solutions:

[0164] Implement a parallel gradient algorithm for transfer learning based on incremental iteration to accelerate the model training process.

[0165] Monitor the convergence of the model and data consistency in real time during the training process to ensure the accuracy and stability of the training results.

[0166] In summary, the specific implementation method of the intelligent breeding model parallel optimization method adopted by the intelligent breeding model parallel optimization system is as follows:

[0167] 1. Parallel computing environment construction:

[0168] Build a parallel computing environment on the domestically-produced independent heterogeneous multi-core supercomputing system and configure the necessary software and hardware resources.

[0169] Design and optimize inter-node communication strategies to reduce communication delays and conflicts.

[0170] 2. Data segmentation and storage:

[0171] Divide large-scale data sets into multiple small blocks and assign them to different computing nodes for parallel processing. Adopt efficient data storage and access strategies to ensure fast data transmission and access during parallel computing.

[0172] 3. Model training:

[0173] Implement a parallel gradient algorithm for transfer learning based on incremental iterations, and use parallel computing resources to accelerate the model training process.

[0174] During the training process, the convergence of the model and data consistency are monitored in real time to ensure the accuracy and stability of the training results.

[0175] Please see attached Figure 5 , the rice intelligent breeding prediction system includes:

[0176] A prediction model building unit, which is used to build an intelligent breeding prediction model based on the attention mechanism;

[0177] Data processing unit, which is used to clean, complete and feature-screen the original data to improve data quality;

[0178] A model training and evaluation unit, which is used to use the processed data to train the model and conduct a comprehensive evaluation of the model;

[0179] The prediction result output unit is used to display the prediction results through the user interaction interface and generate a prediction report.

[0180] Specifically, the prediction model construction unit can more accurately model the interactions between genes and improve the accuracy of the prediction model through an intelligent breeding prediction model based on the attention mechanism.

[0181] Ensure that the complex relationships between genes can be accurately captured when making hybrid rice breeding predictions.

[0182] It solves the problem that traditional breeding methods have difficulty in accurately modeling the complex interactions between genes, resulting in inaccurate prediction results.

[0183] Specific solutions:

[0184] An intelligent breeding prediction model is constructed using the attention mechanism to capture and model the interactions between genes and improve the prediction accuracy of the model.

[0185] The data processing unit improves data quality and ensures the accuracy and completeness of the input data for the prediction model.

[0186] Improve data quality and consistency through deep cleaning, completion, and feature screening.

[0187] It solves the problem of low data quality, missing data and noise in traditional methods, which affects the accuracy of the prediction model.

[0188] Specific solutions:

[0189] Perform deep cleaning on raw data to remove noise and outliers.

[0190] Complete missing data to ensure data integrity.

[0191] Perform feature screening, extract key features, and improve data effectiveness.

[0192] The model training and evaluation unit ensures the accuracy and stability of the prediction model by comprehensively evaluating the model performance.

[0193] Optimize model parameters, improve prediction performance, and provide reliable breeding decision support.

[0194] It solves the problem of incomplete evaluation of traditional breeding models, which leads to unstable model performance.

[0195] Specific solutions:

[0196] Use the processed high-quality data for model training and continuously adjust model parameters to optimize performance.

[0197] Develop a full-process model evaluation method for machine learning, including cross-validation and performance evaluation indicators, to comprehensively evaluate the model and ensure its stability and accuracy.

[0198] The prediction result output unit displays the prediction results through a user interactive interface, making it easy for breeders to intuitively understand the prediction situation and generate detailed prediction reports.

[0199] Provides clear prediction result display and report generation function to support breeding decision-making process.

[0200] Existing technical problems solved: In traditional methods, the prediction results are not intuitive and lack detailed report support, which affects breeding decisions.

[0201] Specific solutions:

[0202] The prediction results, including trait prediction values ​​and confidence intervals, are displayed through a user interface.

[0203] It provides a prediction report generation function to facilitate breeders to review and analyze prediction results and support subsequent breeding decisions.

[0204] In summary, the specific implementation of the rice intelligent breeding prediction system is as follows:

[0205] 1. Prediction model construction:

[0206] Construct an intelligent breeding prediction model based on the attention mechanism and use the attention mechanism to model the interaction between genes.

[0207] The data processing of the seed industry vertical big model is introduced to conduct in-depth cleaning, completion and feature screening of the original data to improve data quality.

[0208] 2. Model training and evaluation:

[0209] Use the processed data to train the prediction model and continuously adjust the model parameters to optimize the prediction performance.

[0210] Develop a full-process model evaluation method for machine learning, including cross-validation, performance evaluation indicators, etc., to objectively and comprehensively evaluate the predictive model.

[0211] 3. Prediction result output:

[0212] The prediction results, including the predicted values ​​and confidence intervals of hybrid rice traits, are displayed through a user interactive interface.

[0213] Provides prediction report generation function to facilitate breeders to review and analyze prediction results.

[0214] Please see attached Figure 6 The intelligent breeding supercomputing cloud native open innovation platform includes:

[0215] Platform architecture unit, which is used to design platform architecture using microservice architecture and container technology to achieve decoupling and independent deployment of services;

[0216] Resource management and scheduling unit, which is used to implement model and data adaptation strategies, communication optimization strategies and serverless framework;

[0217] Security control and permission management unit, which is used to control access to and protect platform resources;

[0218] The user interface is used to display prediction results, receive user input, and provide operation guidance.

[0219] Specifically, the platform architecture unit achieves decoupling and independent deployment of services by adopting microservice architecture and container technology, thereby improving the flexibility and maintainability of the system.

[0220] Ensure that different service modules are independent of each other to improve the stability and scalability of the system.

[0221] It solves the problem that the traditional system has a single architecture and is difficult to adapt to complex breeding application scenarios and rapidly changing needs.

[0222] Specific solutions:

[0223] The platform is designed with a microservice architecture to ensure that each service module can be independently developed, deployed and maintained.

[0224] Use container technology to improve service deployment efficiency and resource utilization.

[0225] The resource management and scheduling unit ensures efficient utilization and on-demand allocation of resources by implementing model and data adaptation strategies, communication optimization strategies, and serverless frameworks.

[0226] Optimize resource management and improve the system's computing efficiency and response speed.

[0227] It solves the problem that traditional resource management methods are inefficient, have low resource utilization, and are unable to meet large-scale data processing needs.

[0228] Specific solutions:

[0229] Design and implement model and data adaptation strategies to ensure seamless connection and efficient use of different resources and services.

[0230] Optimize communication strategies to reduce network latency and bandwidth usage.

[0231] A serverless framework is adopted to achieve on-demand allocation, elastic expansion and automatic recycling of resources.

[0232] The security control and permission management unit ensures the security of data and systems by performing access control and security protection on platform resources.

[0233] Ensure the secure operation of the platform and prevent unauthorized access and data leakage.

[0234] It solves the problem that the security control measures of traditional platforms are insufficient and prone to data leakage and system security issues.

[0235] Specific solutions:

[0236] Deploy security control modules to control access to platform resources.

[0237] Implement permission management functions to ensure permission allocation and access control for different user roles and prevent unauthorized operations.

[0238] The user interface displays prediction results, receives user input and provides operation guidance to improve user experience and system usability.

[0239] It is convenient for breeders to view prediction results, input breeding parameters and obtain operational guidance to support breeding decisions.

[0240] It solves the problem that the traditional system's user interface design is unfriendly, lacks interactivity and guidance, and affects user experience and usage efficiency.

[0241] Specific solutions:

[0242] Design an intuitive and easy-to-use user interface to display prediction results and related information.

[0243] Provide user input function to support breeders to input breeding parameters and adjust breeding plans.

[0244] Provide operation guidance to help users understand system functions and operation steps and improve usage efficiency.

[0245] In summary, the specific implementation methods of the intelligent breeding supercomputing cloud native open innovation platform are as follows:

[0246] 1. Platform architecture design:

[0247] The platform architecture is designed using microservice architecture and container technology to achieve decoupling and independent deployment of services.

[0248] Integrate various resources such as computing power, algorithms, data, and applications to build an open and shared innovation environment.

[0249] 2. Resource management and scheduling:

[0250] Design model and data adaptation strategies to ensure seamless connection and efficient use of different resources and services.

[0251] Implement communication optimization strategies to reduce network latency and bandwidth usage.

[0252] A serverless framework is adopted to achieve on-demand allocation, elastic expansion and automatic recycling of resources.

[0253] 3. Security control and authority management:

[0254] Deploy security control modules to perform access control and security protection on platform resources.

[0255] Implement permission management functions to ensure permission allocation and access control for different user roles.

[0256] Please see attached Figure 7 , a method for hybrid rice breeding using a smart breeding system, comprising the following steps:

[0257] S1. Collect genotype data, environmental factors and seed phenotype data of rice parents and offspring using the intelligent breeding data collection and processing system;

[0258] S2. Use the intelligent breeding model parallel optimization system to process and analyze the collected data and train the intelligent breeding prediction model;

[0259] S3. Predict the characteristics of hybrid rice through the rice intelligent breeding prediction system and generate a prediction report

[0260] S4. Based on the forecast report, formulate a hybrid rice breeding plan and implement and verify it in the field.

[0261] The S1 step specifically includes the following steps:

[0262] S1.1. Genotype data collection:

[0263] Genotype data were collected using the indica rice core site liquid phase 10K breeding chip;

[0264] Obtain SNP information of rice parents and offspring through whole genome resequencing technology;

[0265] S1.2. Phenotypic data collection:

[0266] Use intelligent field environment collection equipment to collect data on environmental factors during rice growth, including temperature, humidity, and light intensity;

[0267] Use seed testing equipment, such as automatic seed testing machines and image recognition systems, to measure and record rice phenotypic characteristics, such as plant height, number of panicles, growth period, 1,000-grain weight, number of tillers, and seed setting rate;

[0268] S1.3 Data processing:

[0269] Perform quality control on the collected genotypic and phenotypic data and remove low-quality, missing or abnormal data points;

[0270] Integrate environmental parameters and phenotypic data to ensure data accuracy and consistency;

[0271] Data storage and management form a unified database to provide basic data for subsequent analysis.

[0272] The S2 step specifically includes the following steps:

[0273] S2.1, Parallel computing environment:

[0274] Build a parallel computing environment on a domestically produced heterogeneous multi-core supercomputer system and configure the necessary software and hardware resources;

[0275] S2.2, Data segmentation and storage:

[0276] Divide large-scale data sets into small pieces and assign them to different computing nodes for parallel processing;

[0277] Adopt efficient data storage and access strategies to ensure fast data transmission and access during parallel computing;

[0278] S2.3, Model training:

[0279] Implement a parallel gradient algorithm for transfer learning based on incremental iteration to accelerate the model training process;

[0280] Monitor data consistency and model convergence during training to ensure the accuracy and stability of training results;

[0281] S2.4, Optimization:

[0282] Optimize inter-node communication strategies and data storage methods to reduce communication delays and memory consumption.

[0283] The S3 step specifically includes the following:

[0284] S3.1、Prediction model construction:

[0285] Build an intelligent breeding prediction model based on the attention mechanism to model the mutual influence between genes;

[0286] S3.2 Data processing:

[0287] Clean, complete and feature-screen the original data to improve data quality;

[0288] S3.3, Model training and evaluation:

[0289] Use the processed data to train the prediction model and adjust the model parameters to optimize the prediction performance;

[0290] Develop a full-process model evaluation method for machine learning, including cross-validation and performance evaluation indicators, to comprehensively evaluate the predictive model;

[0291] S3.4. Prediction result output

[0292] The prediction results are displayed through the user interface, including the predicted values ​​and confidence intervals of hybrid rice traits;

[0293] Generate prediction reports to facilitate breeders to review and analyze prediction results.

[0294] The S4 step specifically includes the following steps:

[0295] S4.1. Develop breeding plan:

[0296] According to the forecast report and combined with actual breeding needs, formulate a detailed hybrid rice breeding plan;

[0297] S4.2. Field implementation:

[0298] Implement hybrid rice breeding programs in the field and monitor various breeding operations;

[0299] S4.3, Verification and Adjustment:

[0300] Monitor and record breeding programs implemented in the field, and dynamically adjust and optimize breeding programs based on test results;

[0301] Collect data generated during the implementation process to provide basic data for subsequent analysis and optimization.

[0302] Although embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions and variations may be made to the embodiments without departing from the principles and spirit of the present invention, and that the scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. An intelligent breeding system for hybrid rice, characterized in that: include: Intelligent breeding data collection and processing system, which is used to collect and process genotype data, environmental factors and seed phenotype data of rice parents and offspring; Intelligent breeding model parallel optimization system, which is used to optimize the parallel computing process and improve the training efficiency and stability of the intelligent breeding model; The rice intelligent breeding prediction system is used to build and evaluate intelligent breeding prediction models and improve the accuracy of phenotypic prediction; Intelligent breeding supercomputing cloud-native open innovation platform, which is used to integrate and manage the computing resources, algorithms and data required for breeding.

2. The intelligent breeding system for hybrid rice according to claim 1, characterized in that: The intelligent breeding data collection and processing system comprises: A genotype data collection unit, which is used to collect genotype data of rice parents and progeny based on the indica rice core site liquid phase 10K breeding chip and deep sequencing technology; A phenotypic data collection unit, which is used to collect rice phenotypic data through field intelligent environment collection equipment and seed testing equipment; The data processing unit is used to clean, organize and integrate the collected data.

3. The intelligent breeding system for hybrid rice according to claim 1, characterized in that: The intelligent breeding model parallel optimization system comprises: Parallel computing environment unit, which is used to build a parallel computing environment on the domestically produced heterogeneous multi-core supercomputer system and configure the necessary software and hardware resources; Data segmentation and storage unit, which is used to segment large data sets into small blocks and assign them to different computing nodes for parallel processing; The model training unit is used to implement a parallel gradient algorithm for transfer learning based on incremental iterations, improve model training efficiency, and monitor data consistency and model convergence.

4. The intelligent breeding system for hybrid rice according to claim 1, characterized in that: The rice intelligent breeding prediction system comprises: A prediction model building unit, which is used to build an intelligent breeding prediction model based on the attention mechanism; Data processing unit, which is used to clean, complete and feature-screen the original data to improve data quality; A model training and evaluation unit, which is used to use the processed data to train the model and conduct a comprehensive evaluation of the model; The prediction result output unit is used to display the prediction results through the user interaction interface and generate a prediction report.

5. The intelligent breeding system for hybrid rice according to claim 1, characterized in that: The intelligent breeding supercomputing cloud native open innovation platform includes: Platform architecture unit, which is used to design platform architecture using microservice architecture and container technology to achieve decoupling and independent deployment of services; Resource management and scheduling unit, which is used to implement model and data adaptation strategies, communication optimization strategies and serverless framework; Security control and permission management unit, which is used to control access to and protect platform resources; The user interface is used to display prediction results, receive user input, and provide operation guidance.

6. A method for hybrid rice breeding using a smart breeding system, characterized in that: A smart breeding system for hybrid rice according to any one of claims 1 to 5, comprising the following steps: S1. Collect genotype data, environmental factors and seed phenotype data of rice parents and offspring using the intelligent breeding data collection and processing system; S2. Use the intelligent breeding model parallel optimization system to process and analyze the collected data and train the intelligent breeding prediction model; S3. Predict the traits of hybrid rice through the rice intelligent breeding prediction system and generate a prediction report; S4. Based on the forecast report, formulate a hybrid rice breeding plan and implement and verify it in the field.

7. The method for hybrid rice breeding using the smart breeding system according to claim 6, characterized in that: The S1 step specifically includes the following steps: S1.

1. Genotype data collection: Genotype data were collected using the indica rice core site liquid phase 10K breeding chip; Obtain SNP information of rice parents and offspring through whole genome resequencing technology; S1.

2. Phenotypic data collection: Use intelligent field environment collection equipment to collect data on environmental factors during rice growth, including temperature, humidity, and light intensity; Use seed testing equipment, such as automatic seed testing machines and image recognition systems, to measure and record rice phenotypic characteristics, such as plant height, number of panicles, growth period, 1,000-grain weight, number of tillers, and seed setting rate; S1.3 Data processing: Perform quality control on the collected genotypic and phenotypic data and remove low-quality, missing or abnormal data points; Integrate environmental parameters and phenotypic data to ensure data accuracy and consistency; Data is stored and managed to form a unified database, providing basic data for subsequent analysis.

8. The method for hybrid rice breeding using the smart breeding system according to claim 6, characterized in that: The S2 step specifically includes the following steps: S2.1, Parallel computing environment: Build a parallel computing environment on a domestically produced heterogeneous multi-core supercomputer system and configure the necessary software and hardware resources; S2.2, Data segmentation and storage: Divide large-scale data sets into small pieces and assign them to different computing nodes for parallel processing; Adopt efficient data storage and access strategies to ensure fast data transmission and access during parallel computing; S2.3, Model training: Implement a parallel gradient algorithm for transfer learning based on incremental iteration to accelerate the model training process; Monitor data consistency and model convergence during training to ensure the accuracy and stability of training results; S2.4, Optimization: Optimize inter-node communication strategies and data storage methods to reduce communication delays and memory consumption.

9. The method for hybrid rice breeding using the smart breeding system according to claim 6, characterized in that: The S3 step specifically includes the following contents: S3.1、Prediction model construction: Build an intelligent breeding prediction model based on the attention mechanism to model the mutual influence between genes; S3.2 Data processing: Clean, complete and feature-screen the original data to improve data quality; S3.3, Model training and evaluation: Use the processed data to train the prediction model and adjust the model parameters to optimize the prediction performance; Develop a full-process model evaluation method for machine learning, including cross-validation and performance evaluation indicators, to comprehensively evaluate the predictive model; S3.

4. Prediction result output The prediction results are displayed through the user interface, including the predicted values ​​and confidence intervals of hybrid rice traits; Generate prediction reports to facilitate breeders to review and analyze prediction results.

10. The method for hybrid rice breeding using the smart breeding system according to claim 6, characterized in that: The S4 step specifically includes the following steps: S4.

1. Develop breeding plan: According to the forecast report and combined with actual breeding needs, formulate a detailed hybrid rice breeding plan; S4.

2. Field implementation: Implement hybrid rice breeding programs in the field and monitor various breeding operations; S4.3, Verification and Adjustment: Monitor and record breeding programs implemented in the field, and dynamically adjust and optimize breeding programs based on test results; Collect data generated during the implementation process to provide basic data for subsequent analysis and optimization.

Images