A decision support system and method for whole genome selection-assisted rice breeding

Abstract

The present invention relates to the field of bioinformatics, and in particular to a decision support system and method for whole genome selection-assisted breeding of rice. The content includes: collecting genome data, environmental data and time series data of rice samples, constructing a multidimensional feature association matrix, deducing the change of genome data over time, and updating the genome data, predicting the breeding target traits of rice based on the updated genome data; optimizing the weights of breeding target traits and calculating the comprehensive evaluation score; decomposing the comprehensive evaluation score, performing dynamic constraint optimization based on the quantitative indicators of the breeding target obtained by decomposition, and generating the optimal breeding decision plan. It solves the problem that the traditional method relies on manual experiments and experience, has a long breeding cycle and high cost, and is difficult to achieve fast and efficient breeding goals; it is difficult to achieve an effective balance between multiple breeding goals, resulting in insufficient breeding results and greatly limiting the improvement of breeding efficiency.

Description

Technical Field

[0001] The present invention relates to the field of bioinformatics, and in particular to a decision support system and method for whole genome selection-assisted rice breeding. Background Art

[0002] Traditional rice breeding methods mainly rely on artificial selection and hybridization experiments. Although they have achieved certain success in the past, with the increase in demand and the complexity of environmental changes, this traditional method has been unable to meet the high efficiency, high yield and stress resistance requirements of modern agriculture. In recent years, with the rapid development of genomics technology, especially the widespread application of high-throughput sequencing technology, scientists can use genomic data to understand the genetic information of plants, thereby providing a more accurate and efficient means for the breeding process.

[0003] Whole genome selection technology has become a breeding method that has attracted much attention. It predicts the performance of target traits of plants by analyzing genomic data, thereby accelerating the selection of excellent varieties. However, selection based on genomic data alone is not enough to fully predict the performance of plants under different environmental conditions. The growth of rice is affected by a variety of external factors, such as environmental conditions such as climate, soil and water, as well as meteorological changes at different times during the growth cycle. The interactive effects of environmental factors and gene expression play a vital role in the final performance of rice traits such as yield, disease resistance and drought tolerance. Therefore, in the process of rice breeding, the three key dimensions of genes, environment and time must be combined to accurately predict the performance of crops under real planting conditions and make optimization decisions.

[0004] Traditional rice breeding methods have the following technical problems: they rely on manual experiments and experience and lack intelligent data analysis and decision-making support tools, which not only leads to long breeding cycles and high costs, but also large uncertainty and blindness in the test results, making it difficult to achieve fast and efficient breeding goals; it is difficult to achieve an effective balance between multiple breeding goals (such as disease resistance, yield, drought resistance, etc.), and it is easy for one breeding goal to be overly dominant, while ignoring the importance of other breeding goals, resulting in incomplete breeding results and weak comprehensive traits of the varieties, which greatly limits the improvement of breeding efficiency. Summary of the invention

[0005] The present invention provides a decision support system and method for whole genome selection-assisted rice breeding, so as to solve the problems that the traditional rice breeding method relies on manual experiments and experience, lacks intelligent data analysis and decision support tools, not only leads to a long breeding cycle and high cost, but also has great uncertainty and blindness in the test results, making it difficult to achieve fast and efficient breeding goals; it is difficult to achieve an effective balance between multiple breeding goals (such as disease resistance, yield, drought resistance, etc.), and it is easy for a certain breeding goal to be overly dominant, while ignoring the importance of other breeding goals, resulting in incomplete breeding results, weak comprehensive traits of the variety, and greatly limiting the improvement of breeding efficiency.

[0006] The present invention provides a decision support system and method for whole genome selection-assisted rice breeding, which specifically includes the following technical solutions:

[0007] A decision support method for whole genome selection-assisted rice breeding, comprising the following steps:

[0008] S1. Collect genomic data, environmental data and time series data of rice samples, and construct a multidimensional feature association matrix; based on the multidimensional feature association matrix, deduce the change of genomic data over time, and update the genomic data; based on the updated genomic data, predict the breeding target traits of rice and obtain the predicted values ​​of the breeding target traits;

[0009] S2. Optimize the weights of breeding target traits based on the predicted values ​​of breeding target traits to obtain optimized weights of breeding target traits; calculate the comprehensive evaluation score based on the optimized weights of breeding target traits; decompose the comprehensive evaluation score to obtain quantitative indicators of breeding targets; perform dynamic constraint optimization based on the quantitative indicators of breeding targets to generate the optimal breeding decision plan.

[0010] Preferably, the S1 specifically includes:

[0011] The genomic data, environmental data and time series data were integrated, and the genomic data of each rice sample was matched with the corresponding environmental data and time series data to construct a multidimensional feature association matrix.

[0012] Preferably, the S1 specifically includes:

[0013] Based on the multidimensional feature association matrix, the dynamic equation is used to simulate and deduce the changes of genome data over time to obtain the gene change rate. The formula is as follows:

[0014]

[0015] in, It is genome data in time The rate of change on represents the rate of gene change; Indicates at time No. genome data; is the adjustment coefficient; are the elements of the multidimensional feature association matrix; and are the dimensions of environmental data and time series data, respectively; Indicates at time No. Environmental data; Indicates the rate of change over time; Represents time series data; Indicates a point in time.

[0016] Preferably, the S1 specifically includes:

[0017] Based on the gene change rate, the genome data is updated; based on the updated genome data, a nonlinear gene-trait mapping model is introduced, and the environmental data and time series data are combined to obtain the predicted values ​​of breeding target traits.

[0018] Preferably, the S2 specifically includes:

[0019] Based on the predicted values ​​of breeding target traits, a progressive weight transfer mechanism is introduced to dynamically optimize the trait weights of different breeding targets to obtain the optimized breeding target trait weights.

[0020] Preferably, the S2 specifically includes:

[0021] Based on the optimized breeding target trait weights and breeding target trait prediction values, the breeding targets are comprehensively evaluated to obtain a comprehensive evaluation score; the comprehensive evaluation formula is:

[0022]

[0023] in, It's in time The comprehensive evaluation score of all breeding objectives and traits; is the total number of breeding objectives; is the total number of traits; Indicates at time At that time, Breeding goal The weight of personality traits; It's in time Previous Predicted value of breeding target traits; Indicates at time At that time, Breeding goal The weight of personality traits; and All are ordinal numbers of breeding objectives.

[0024] Preferably, the S2 specifically includes:

[0025] The comprehensive evaluation score is decomposed, breeding indicators are generated for different breeding goals, and the quantitative indicators of the breeding goals are obtained based on the breeding target trait weights and breeding target trait prediction values. The specific decomposition formula is as follows:

[0026]

[0027] in, Indicates at time At that time, Quantitative indicators of breeding goals.

[0028] Preferably, the S2 specifically includes:

[0029] Based on the quantitative indicators of breeding objectives, dynamic constraint optimization is performed to obtain the optimal target performance.

[0030] Preferably, the S2 specifically includes:

[0031] Based on the optimal target performance, combination optimization is performed to generate the optimal breeding decision plan.

[0032] A decision support system for whole genome selection-assisted rice breeding, comprising the following parts:

[0033] Data acquisition module, data preprocessing and integration module, trait prediction module, comprehensive evaluation module, dynamic constraint optimization module, and combination optimization module;

[0034] The data collection module collects the genome data, environmental data, and time series data of rice samples at different growth cycles; and passes the genome data, environmental data, and time series data to the data preprocessing and integration module;

[0035] The data preprocessing and integration module integrates genomic data, environmental data and time series data to generate a multidimensional feature association matrix; the multidimensional feature association matrix is ​​passed to the trait prediction module;

[0036] The trait prediction module, based on the multidimensional feature association matrix, uses dynamic equations to deduce the changes in genome data over time and updates the genome data; based on the updated genome data, a nonlinear gene-trait mapping model is used to predict the target traits of rice and obtain the predicted values ​​of the breeding target traits; the predicted values ​​of the breeding target traits are passed to the comprehensive evaluation module;

[0037] The comprehensive evaluation module dynamically optimizes the trait weights of different breeding objectives through a progressive weight transfer mechanism to obtain the optimized breeding objective trait weights; based on the optimized breeding objective trait weights and the breeding objective trait prediction values, the breeding objectives are comprehensively evaluated to obtain a comprehensive evaluation score; the comprehensive evaluation score is passed to the dynamic constraint optimization module;

[0038] The dynamic constraint optimization module decomposes the comprehensive evaluation score to obtain the quantitative indicators of the breeding objectives; performs dynamic constraint optimization based on the quantitative indicators of the breeding objectives to obtain the optimal target performance; and passes the optimal target performance to the combination optimization module;

[0039] The combination optimization module optimizes the combination of genomic data, environmental data and time series data based on the optimal target performance to generate the optimal breeding decision plan.

[0040] The beneficial effects of the technical solution of the present invention are:

[0041] 1. By constructing a multidimensional feature association matrix and combining genomic data, environmental data and time series data, the dynamic changes of genes under different environmental and temporal conditions can be accurately simulated; by adopting a nonlinear gene-trait mapping model and combining the dynamic changes of environmental data and time series data, the decision support system can accurately predict rice traits (such as yield, disease resistance, drought resistance, etc.), improve breeding efficiency, avoid the blindness and high-cost experiments in traditional methods, and greatly improve the success rate of breeding work;

[0042] 2. By introducing a progressive weight transfer mechanism, the trait weights under different breeding goals are dynamically optimized to achieve balanced optimization among multiple goals. The decision support system dynamically transfers the trait weights between breeding goals through a progressive adjustment mechanism, so that breeders can achieve a balance among multiple breeding goals such as yield, disease resistance, and stress tolerance, avoiding the situation where one goal is overly dominant or other breeding goals are neglected. The balanced optimization of breeding goals can ensure the comprehensive excellence of rice varieties, thereby coping with complex farmland environments and climate changes.

[0043] 3. By combining the actual breeding environment and breeders' needs, dynamic constraint optimization is performed on the resources, costs, land area and other constraints that may exist in the breeding process; during the optimization process, the decision support system can obtain the optimal breeding combination plan based on different resource allocation situations and the priority weights set by the breeder; the dynamic constraint optimization function makes the decision support system highly adaptable in practical applications, and can generate the optimal breeding decision plan based on the resource limitations and environmental conditions of breeding. BRIEF DESCRIPTION OF THE DRAWINGS

[0044] Figure 1This is a structural diagram of a decision support system for whole genome selection-assisted breeding of rice according to the present invention;

[0045] Figure 2 This is a flow chart of a decision support method for whole genome selection-assisted rice breeding according to the present invention. DETAILED DESCRIPTION

[0046] In order to further explain the technical means and effects adopted by the present invention to achieve the predetermined invention purpose, the technical scheme in the embodiment of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiment of the present invention. Obviously, the described embodiment is only a part of the embodiment of the present invention, not all 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.

[0047] Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention belongs.

[0048] The specific scheme of a decision support system and method for whole genome selection-assisted breeding of rice provided by the present invention is described in detail below with reference to the accompanying drawings.

[0049] See attached Figure 1 , which shows a structure diagram of a decision support system for whole genome selection-assisted breeding of rice provided by an embodiment of the present invention, the system includes the following parts:

[0050] Data acquisition module, data preprocessing and integration module, trait prediction module, comprehensive evaluation module, dynamic constraint optimization module, and combination optimization module;

[0051] The data collection module collects the genome data, environmental data, and time series data of rice samples at different growth cycles; and passes the genome data, environmental data, and time series data to the data preprocessing and integration module;

[0052] The data preprocessing and integration module integrates genomic data, environmental data and time series data to generate a multidimensional feature association matrix; the multidimensional feature association matrix is ​​passed to the trait prediction module;

[0053] The trait prediction module, based on the multidimensional feature association matrix, uses dynamic equations to deduce the changes in genome data over time and updates the genome data; based on the updated genome data, a nonlinear gene-trait mapping model is used to predict the target traits of rice and obtain the predicted values ​​of the breeding target traits; the predicted values ​​of the breeding target traits are passed to the comprehensive evaluation module;

[0054] The comprehensive evaluation module dynamically optimizes the trait weights of different breeding objectives through a progressive weight transfer mechanism to obtain the optimized breeding objective trait weights to ensure the balanced optimization of multiple objectives; it comprehensively evaluates the breeding objectives by combining the optimized breeding objective trait weights and the predicted breeding objective trait values ​​to obtain a comprehensive evaluation score; and it passes the comprehensive evaluation score to the dynamic constraint optimization module;

[0055] The dynamic constraint optimization module decomposes the comprehensive evaluation score, generates breeding indicators for different breeding objectives, and obtains quantitative indicators of the breeding objectives, which are presented in a quantitative form; combines the actual breeding environment and breeder needs, performs dynamic constraint optimization based on the quantitative indicators of the breeding objectives, and obtains the optimal target performance; and passes the optimal target performance to the combination optimization module;

[0056] The combination optimization module further optimizes the combination of genomic data, environmental data and time series data to generate the optimal breeding decision plan.

[0057] See attached Figure 2 , which shows a flow chart of a decision support method for whole genome selection-assisted breeding of rice provided by an embodiment of the present invention, the method comprising the following steps:

[0058] S1. Collect genomic data, environmental data and time series data of rice samples, and construct a multidimensional feature association matrix; based on the multidimensional feature association matrix, deduce the change of genomic data over time, and update the genomic data; based on the updated genomic data, predict the breeding target traits of rice and obtain the predicted values ​​of the breeding target traits;

[0059] The rice samples are sequenced using high-throughput sequencing technology using common sequencing platforms such as Illumina or PacBio, and a genome data set containing thousands of gene loci is obtained from the rice samples. , genomic data includes single nucleotide polymorphism (SNP), insertion and deletion variation (InDel) and other gene marker information, among which Indicates individual gene markers.

[0060] Environmental factors, such as temperature, humidity, light intensity, soil nutrients, and precipitation, play an important role in rice breeding. Therefore, it is necessary to accurately obtain environmental data related to breeding conditions, such as temperature, humidity, light intensity, soil nutrients, precipitation data, etc. The above environmental data are provided by a variety of sensors or agricultural meteorological stations, which can be real-time collection or based on historical environmental records. Each environmental factor will be combined with genomic data to analyze its impact on rice traits. The environmental data is recorded as ,in Represents different environmental factors.

[0061] The growth process of rice is a dynamic time series process, and different growth stages have different responses to the environment and gene expression. Therefore, it is necessary to collect time series data to represent the gene expression and environmental response of rice at different growth stages. The time series data is marked by the agricultural growth cycle and experimental time point, recorded as ,in Representative time points (such as seedling stage, tillering stage, heading stage, maturity stage, etc.).

[0062] The acquired genomic data, environmental data and time series data are integrated, and the genomic data of each rice sample is matched with the corresponding environmental data and time series data to construct a multidimensional feature association matrix. The purpose of constructing a multidimensional feature association matrix is ​​to transform the complex associations of genomic data, environmental data and time series data into relationships that can reflect real biological mechanisms through mathematical deduction. The construction formula of the multidimensional feature association matrix is ​​as follows:

[0063]

[0064] in, is an element of the multidimensional feature association matrix, representing genomic data , Environmental data and time series data The strength of the correlation between the three; standardize the genomic data, environmental data and time series data of each rice sample, and use the interaction effect of genes, environment and time to reflect the real correlation; is the number of rice samples; , They are the value of genomic data, environmental data, and time series data in rice samples; , , are the standard deviations of genomic data, environmental data, and time series data, respectively; , , are the means of genomic data, environmental data, and time series data, respectively.

[0065] Simulating genomic data through dynamic equations based on multidimensional feature association matrices The changes over time use the coupling relationship between genes, environment and time to predict the performance of genes at different time points and obtain the gene change rate. The core formula is as follows:

[0066]

[0067] in, Indicates genome data in time The rate of change in genes, i.e. the rate of change in genes; Indicates at time No. genome data; is the adjustment coefficient, which indicates the sensitivity of the gene change; and are the dimensions of environmental data and time series data respectively. No. Environmental data and time rate of change The dynamic changes in gene expression will be simulated based on real-time climate, soil conditions, and time series data input during the growth cycle.

[0068] Genomic data can affect the traits of rice (such as disease resistance, yield, stress resistance, etc.). Therefore, in the breeding process, the breeding goal is to achieve certain specific traits that the rice is expected to have during the breeding process, such as increasing yield, enhancing disease resistance, improving drought resistance, etc.; the gene change rate represents the dynamic changes of genomic data under different times and environmental conditions, and is the dynamic response of genes to environmental (such as temperature, humidity, etc.) and time (growth cycle) factors.

[0069] The expression of genome data at each time point is gradually updated according to the gene change rate. The specific update formula is:

[0070]

[0071] in, It's in time The updated genome data will Recorded as ; is the update time step.

[0072] In order to derive the trait values ​​of breeding targets based on the updated genomic data, a nonlinear gene-trait mapping model was introduced to convert the updated genomic data into It is associated with the predicted value of the breeding target traits, combined with environmental data and time series data, to jointly determine the breeding target trait performance and predict the breeding target traits of rice. The formula is as follows:

[0073]

[0074] in, It's in time Previous Predicted value of breeding target traits; Represents the hyperbolic tangent function, ensuring that the output value is between [-1,1] to enhance the ability to handle extreme data; It is the parameter to be estimated in the nonlinear gene-trait mapping model, which is used to represent the contribution of genomic data, environmental data and time series data to the breeding target trait; is the total number of genomic data; Indicates at time No. time series data.

[0075] S2. Optimize the weights of breeding target traits based on the predicted values ​​of breeding target traits to obtain optimized weights of breeding target traits; calculate the comprehensive evaluation score based on the optimized weights of breeding target traits; decompose the comprehensive evaluation score to obtain quantitative indicators of breeding targets; perform dynamic constraint optimization based on the quantitative indicators of breeding targets to generate the optimal breeding decision plan.

[0076] According to the different predicted values ​​of breeding target traits, a progressive weight transfer mechanism is introduced to dynamically optimize the weight of each breeding target trait so that the relative importance of different breeding target traits can be reflected in the comprehensive evaluation; in the progressive weight transfer mechanism, the concepts of "hierarchical progression" and "weight transfer" are introduced. After assigning initial weights to each breeding target trait, the weights of each breeding target trait are progressively adjusted. The weights will gradually transfer among different breeding target traits, thereby achieving balanced optimization of multiple objectives.

[0077] Specifically, in the initial stage, combined with the breeding experience of field experts, the initial weights of various traits under different breeding goals are manually set, and then through a progressive process, the weights of breeding target traits are optimized with time as the level, and the weights of different breeding target traits are transferred in each level. The formula is as follows:

[0078]

[0079] in, Indicates at time At that time, Breeding goal The weight of the individual trait after optimization, that is, the weight of the optimized breeding target trait; Indicates at time At that time, Breeding goal The amount of change in the weight of the personality trait; Indicates at time At that time, Breeding goal The weight of personality traits; is the total number of breeding objectives; and All are ordinal numbers of breeding objectives; Indicates at time At that time, Breeding goal The weight of individual traits. The weight of the current breeding target is affected by the weight ratio of other breeding targets, and finally the trait weights are dynamically transferred between different breeding targets. Through multi-level progressive adjustment, that is, the next level of the current level gradually adjusts the trait weights of each breeding target according to the weight transfer and the breeding target trait weights of the current level, so that the weight of each breeding target trait changes continuously with the progressive optimization, forming a multi-dimensional optimization effect.

[0080] The specific calculation formula for the weight change is:

[0081]

[0082] in, is the adjustment coefficient, which controls the rate at which the weight changes; It's in time At that time, The greatest weight of an individual trait among all breeding objectives. The better the trait, the faster its weight increases.

[0083] The optimization of progressive levels is not limited to the weight optimization within one breeding target, but also includes global optimization across multiple breeding targets. In the progressive level optimization, each level calculates the comprehensive evaluation score of the current level based on the predicted value of the breeding target trait and the weight of the breeding target trait at the current level. The comprehensive evaluation formula is:

[0084]

[0085] in, It's in time The comprehensive evaluation score of all breeding objectives and traits; is the total number of traits. By taking the weighted sum of the predicted values ​​of each breeding target trait and combining it with the logarithmic function of the weight transfer ratio, it is ensured that each breeding target trait is reasonably balanced in the comprehensive evaluation to avoid excessive dominance of a certain breeding target trait.

[0086] The obtained comprehensive evaluation score is decomposed to generate breeding indicators for different breeding goals. The breeding indicators will be presented in a quantitative form to provide specific breeding target values ​​for subsequent decision support. The specific decomposition formula is as follows:

[0087]

[0088] in, Indicates at time At that time, The quantitative indicator of a breeding target is calculated based on the breeding target trait weight and the predicted value of the breeding target trait.

[0089] After generating quantitative indicators of breeding objectives, dynamic constraint optimization is performed based on the quantitative indicators of breeding objectives in combination with the actual breeding environment and breeders' needs to ensure that the optimal breeding decision plan is generated under the trade-offs and constraints between multiple breeding objectives.

[0090] The specific formula for dynamic constraint optimization is:

[0091]

[0092]

[0093] in, It's in time The priority weight of each breeding goal set by the breeder indicates the emphasis of the breeder on certain breeding goals; the constraint condition It is the constraint condition of each breeding goal in the breeding process, such as water resources, cost, land area and other constraints. is the actual target value, It is the threshold value corresponding to each target value, ensuring that the decision plan can be implemented in the actual environment. Under the premise of meeting resource and constraint conditions, maximize the comprehensive value of different breeding goals and output the optimal target performance .

[0094] The optimal objective performance generated in dynamic constrained optimization Based on this, we conduct combination optimization and further consider the complementary relationship between different genes and environments, so as to generate a combination breeding program suitable for the current climate and soil conditions. The core formula of combination optimization is:

[0095]

[0096] in, It is the optimal breeding decision plan; is the coefficient of gene adaptability to the environment. The optimal breeding decision plan includes the following:

[0097] Prioritization of breeding objectives: Assign priorities to different breeding objectives, such as focusing on yield or disease resistance, based on breeder needs and environmental conditions.

[0098] Optimization suggestions for each trait: Based on the calculated predicted values ​​of breeding target traits and weight adjustments, optimization suggestions for each trait (such as height, leaf area, number of ears, etc.) are provided.

[0099] In summary, a decision support system and method for whole genome selection-assisted breeding of rice have been completed.

[0100] The order of the embodiments of the invention is for description only and does not represent the advantages and disadvantages of the embodiments. The processes depicted in the drawings do not necessarily require the specific order or continuous order shown to achieve the desired results. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0101] The various embodiments in this specification are described in a progressive manner, and the same or similar parts between the various embodiments can be referenced to each other, and each embodiment focuses on the differences from other embodiments.

[0102] The above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit the same. Although the present invention has been described in detail with reference to the above embodiments, those skilled in the art should understand that the technical solutions described in the above embodiments may still be modified, or some of the technical features may be replaced by equivalents. Such modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present invention, and should be included in the protection scope of the present invention.

Claims

1. A decision support method for whole genome selection-assisted rice breeding, characterized in that: The following steps are involved: S1. Collect the genome data, environmental data and time series data of rice samples, integrate the genome data, environmental data and time series data, match the genome data of each rice sample with the corresponding environmental data and time series data, and construct a multidimensional feature association matrix. The specific formula is: ； in, is an element of the multidimensional feature association matrix, representing genomic data , Environmental data and time series data the strength of the inter-correlations among the three; is the number of rice samples; , , They are the value of genomic data, environmental data, and time series data in rice samples; , , are the standard deviations of genomic data, environmental data, and time series data, respectively; , , are the means of genomic data, environmental data, and time series data, respectively; Based on the multidimensional feature association matrix, the changes of genome data over time are simulated and derived through dynamic equations to obtain the gene change rate and update the genome data; based on the updated genome data, the breeding target traits of rice are predicted to obtain the predicted values ​​of the breeding target traits; the specific calculation formula for the gene change rate is: ； in, It is Genome data in time The rate of change on represents the rate of gene change; Indicates at time No. genome data; is the adjustment coefficient; and are the dimensions of environmental data and time series data, respectively; Indicates at time No. Environmental data; Indicates the rate of change over time; Represents time series data; Indicates a point in time; S2. Optimize the weights of breeding target traits based on the predicted values ​​of breeding target traits, introduce a progressive weight transfer mechanism, dynamically optimize the trait weights of different breeding targets, and obtain the optimized breeding target trait weights; The progressive weight transfer mechanism first sets the initial weights of each trait under different breeding objectives, then optimizes the weights of breeding objective traits by time through a progressive process, and transfers the weights of different breeding objective traits at each level. The specific formula is as follows: ； in, It's in time At that time, Breeding goal The weight of the individual trait after optimization indicates the weight of the optimized breeding target trait; Indicates at time At that time, Breeding goal The amount of change in the weight of the personality trait; Indicates at time At that time, Breeding goal The weight of personality traits; is the total number of breeding objectives; and All are ordinal numbers of breeding objectives; Indicates at time At that time, Breeding goal The weight of personality traits; The specific calculation formula for the weight change is: ； in, is the adjustment coefficient; It's in time At that time, Individual traits have the greatest weight among all breeding objectives; It's in time Previous Prediction value of breeding target traits; Based on the optimized breeding target trait weights and breeding target trait prediction values, the breeding targets are comprehensively evaluated and the comprehensive evaluation scores are calculated. The comprehensive evaluation formula is: ； in, It's in time The comprehensive evaluation score of all breeding objectives and traits; is the total number of traits; The comprehensive evaluation score is decomposed to obtain the quantitative indicators of breeding objectives; dynamic constraint optimization is performed based on the quantitative indicators of breeding objectives to generate the optimal breeding decision plan.

2. A decision support method for whole genome selection-assisted rice breeding according to claim 1, characterized in that: The S1 specifically includes: Based on the updated genomic data, a nonlinear gene-trait mapping model was introduced, and the environmental data and time series data were combined to obtain the predicted values ​​of breeding target traits.

3. A decision support method for whole genome selection-assisted rice breeding according to claim 1, characterized in that: The S2 specifically includes: The comprehensive evaluation score is decomposed, breeding indicators are generated for different breeding goals, and the quantitative indicators of the breeding goals are obtained based on the breeding target trait weights and the predicted values ​​of the breeding target traits. The specific decomposition formula is as follows: ； in, Indicates at time At that time, A quantitative indicator of a breeding goal.

4. A decision support method for whole genome selection-assisted rice breeding according to claim 3, characterized in that: The S2 specifically includes: Based on the quantitative indicators of breeding objectives, dynamic constraint optimization is performed to obtain the optimal target performance.

5. A decision support method for whole genome selection-assisted rice breeding according to claim 4, characterized in that: The S2 specifically includes: Based on the optimal target performance, combination optimization is performed to generate the optimal breeding decision plan.

6. A decision support system for whole genome selection-assisted rice breeding, applied to a decision support method for whole genome selection-assisted rice breeding as claimed in claim 1, characterized in that: Includes the following parts: Data acquisition module, data preprocessing and integration module, trait prediction module, comprehensive evaluation module, dynamic constraint optimization module, and combination optimization module; The data collection module collects the genome data, environmental data, and time series data of rice samples at different growth cycles; and passes the genome data, environmental data, and time series data to the data preprocessing and integration module; The data preprocessing and integration module integrates genomic data, environmental data, and time series data to generate a multidimensional feature association matrix; Pass the multidimensional feature association matrix to the trait prediction module; The trait prediction module, based on the multidimensional feature association matrix, uses dynamic equations to deduce the changes in genome data over time and updates the genome data; based on the updated genome data, a nonlinear gene-trait mapping model is used to predict the target traits of rice and obtain the predicted values ​​of the breeding target traits; the predicted values ​​of the breeding target traits are passed to the comprehensive evaluation module; The comprehensive evaluation module dynamically optimizes the trait weights of different breeding objectives through a progressive weight transfer mechanism to obtain the optimized breeding objective trait weights; Based on the optimized breeding target trait weights and breeding target trait prediction values, the breeding targets are comprehensively evaluated to obtain a comprehensive evaluation score; Pass the comprehensive evaluation score to the dynamic constraint optimization module; The dynamic constraint optimization module decomposes the comprehensive evaluation score to obtain the quantitative indicators of the breeding objectives; based on the quantitative indicators of the breeding objectives, dynamic constraint optimization is performed to obtain the optimal target performance; Pass the optimal target performance to the combinatorial optimization module; The combination optimization module optimizes the combination of genomic data, environmental data and time series data based on the optimal target performance to generate the optimal breeding decision plan.

Images