A method suitable for cross-population joint genetic assessment

Abstract

The present application relates to the field of animal genomic breeding technology, and particularly relates to a method suitable for cross-population joint genetic evaluation. The method obtains porcine genotype data and phenotype data from multiple source populations, infers population structure by using an unsupervised genomic stratification method, and divides the individuals to be evaluated into multiple genetically homogeneous subpopulations; based on the inferred population structure, a double random genetic effect linear mixed model is constructed, wherein the first random genetic effect is used to represent the genomic effect shared among different subpopulations, and the second random genetic effect is used to represent the subpopulation-specific deviation effect of each subpopulation relative to the shared genomic effect; the shared genomic effect and the subpopulation-specific deviation effect are predicted respectively, and the sum of the two is determined as the genomic estimated breeding value of the individual. The present application can utilize the genetic information of multiple populations without pre-specifying population labels, and improve the accuracy of genetic evaluation of important economic traits of pigs.

Description

Technical Field

[0001] This application relates to the field of animal genome breeding technology, specifically to a method suitable for cross-population joint genetic evaluation. Background Technology

[0002] Genomic genetic assessment is an important technique in animal breeding. It uses genotypic and phenotypic data to predict an individual's genomic potential, estimating its breeding value for screening candidate breeding stock and for genetic improvement of traits such as growth, reproduction, and feed efficiency. The accuracy of genetic assessment depends primarily on the size of the reference population and the genetic relationship between the reference population and the individual being assessed.

[0003] In pig breeding practice, directly merging data from multiple source populations is the simplest way to expand the size of the reference population. However, this method assumes that gene effects are homogeneous and uniformly distributed, ignoring the heterogeneity of data between populations, which contradicts the true genetic results of the population and may reduce the accuracy of trait prediction. Using multi-trait models to fit unbalanced data is one of the effective methods for joint genetic assessment, but multi-trait models are complex, and when there are many subpopulations, the number of estimated parameters is large, the variance components are difficult to converge, and the assessment results are unstable. In addition, artificially defined population division criteria cannot accurately reflect differences in genetic background. Individuals with different genetic structures may exist within the same source population, and individuals with similar genetic backgrounds may exist between different source populations.

[0004] Existing methods struggle to adaptively identify genetically homogeneous subpopulations based on genomic data and simultaneously utilize shared genetic information among different genetically homogeneous subpopulations and specific genetic information within each subpopulation within the same genetic assessment model. Therefore, it is necessary to provide a joint genetic assessment method that can adapt to complex population structures and effectively utilize cross-population genetic information. Summary of the Invention

[0005] To address the aforementioned technical problems, this application provides a method suitable for cross-population joint genetic assessment. This method infers genetically homogeneous subpopulations based on genome-wide markers and constructs two complementary genomic kinship matrices based on the subpopulation division results. The two genomic kinship matrices are then incorporated into a linear mixture model of two random genetic effects to describe the shared genomic effects between different subpopulations and the subpopulation-specific deviation effects of each subpopulation relative to the shared effects. Finally, the sum of the two predicted values ​​is determined as the individual's estimated genomic breeding value.

[0006] To achieve the above-mentioned objectives, the present invention provides the following technical solution: In a first aspect, the present invention provides a method suitable for joint genetic assessment across populations, comprising: Obtain genotypic and phenotypic data of the pigs to be evaluated from multiple source groups; Unsupervised genome stratification was performed based on the genotype data to determine the genetically homogeneous subgroup to which each pig to be evaluated belonged; Based on the classification results of the genetically homogeneous subgroups and the genotype data, a first genomic kinship matrix is ​​constructed to describe the shared genomic kinship between different genetically homogeneous subgroups and within each genetically homogeneous subgroup, and a second genomic kinship matrix is ​​constructed to describe the specific genomic kinship of each genetically homogeneous subgroup. A linear mixture model of two random genetic effects is constructed based on the first genomic kinship matrix and the second genomic kinship matrix. The first random genetic effect is used to characterize the genomic effect shared among different genetically homogeneous subgroups, and the second random genetic effect is used to characterize the subgroup-specific deviation effect of each genetically homogeneous subgroup relative to the shared genomic effect. The phenotypic data were used to solve the linear mixed model of the two random genetic effects to obtain the predicted values ​​of the shared genome effect and the predicted values ​​of the subgroup-specific deviation effect for each pig to be evaluated. The sum of the predicted value of the shared genome effect and the predicted value of the subgroup-specific deviation effect for the same pig to be evaluated is determined as the estimated genomic breeding value of that pig.

[0007] In some specific embodiments of the present invention, the unsupervised genome stratification based on the genotype data includes: Principal component analysis was performed on the genotype data to obtain the principal component characteristics of the pigs to be evaluated; The principal component features are nonlinearly reduced in dimensionality using a uniform manifold approximation and projection method to obtain low-dimensional genome features. The low-dimensional genomic features were clustered using a density-based hierarchical clustering method to obtain multiple genetically homogeneous subgroups.

[0008] In some specific embodiments of the present invention, the first step obtained from unsupervised genome stratification is... The genotype matrix of each genetically homogeneous subgroup is constructed according to the following formula:

[0009] in, It is the first Pigs in a genetically homogeneous subgroup At the site Allele dosage on It is the first loci in genetically homogeneous subpopulations The frequency of encoded alleles.

[0010] In some specific embodiments of the present invention, the first genomic kinship matrix is ​​constructed according to the following steps: The centralized genotype matrix is ​​processed according to the following formula. Standardize:

[0011] in, It is the first Sample size of each genetically homogeneous subpopulation Represents the trace of a matrix. Indicates matrix transpose; Standardized genotype matrix corresponding to each genetically homogeneous subpopulation By vertically piecing together data according to the individual dimension, a combined standardized genotype matrix is ​​obtained:

[0012] The first genome phylogenetic matrix is ​​constructed according to the following formula:

[0013] Among them, the first The blocks are:

[0014] in, It is the first The and the first Sample size of each genetically homogeneous subpopulation , The first genomic kinship matrix represents the total number of the genetically homogeneous subpopulations. It includes genomic kinship between individuals within the same genetically homogeneous subgroup and between individuals from different genetically homogeneous subgroups, and is used to construct the variance-covariance structure of the shared genomic effect.

[0015] In some specific embodiments of the present invention, the second genomic kinship matrix is ​​constructed according to the following steps:

[0016] Among them, the Each diagonal piece is calculated using the following formula:

[0017] The second genomic kinship matrix Genomic kinship among pigs to be evaluated within the same genetically homogeneous subgroup is preserved, while subgroup-specific genomic kinship among individuals from different genetically homogeneous subgroups is set to zero, in order to construct the variance-covariance structure of the subgroup-specific deviation effect.

[0018] In some specific embodiments of the present invention, the linear mixture model of double random genetic effects is expressed as:

[0019] in, This represents a phenotypic record vector from all genetically homogeneous subpopulations. Represents the fixed effects vector. Represents the fixed effects design matrix. This indicates the shared genome effect. This indicates the subgroup-specific deviation effect. These represent the association matrices that link phenotypic records to shared genome effects and subgroup-specific deviation effects, respectively. This represents a random residual vector.

[0020] In some specific embodiments of the present invention, the shared genome effect, the subgroup-specific deviation effect, and the random residual vector respectively obey:

[0021] in, Indicates the variance of shared genome effects. This indicates the variance of the subgroup-specific deviation effect. Represents the residual variance. The unit matrix is ​​represented; the shared genome effect and the subgroup-specific deviation effect are independent of each other, and their covariance is zero.

[0022] In some specific embodiments of the present invention, the first is determined according to the following formula. Genomic estimated breeding value of the pigs to be evaluated:

[0023] in, Indicates the first Predicted values ​​of shared genome effects for 10 pigs to be evaluated. Indicates the first Predicted values ​​of subgroup-specific deviation effects for each pig to be evaluated.

[0024] In some specific embodiments of the present invention, the plurality of source groups are pig herds distinguished according to at least one of breed, strain, breeding farm, breeding enterprise or geographical origin; The phenotypic data include at least one of growth traits, reproductive traits, feed efficiency traits, carcass traits, or meat quality traits.

[0025] In some specific embodiments of the present invention, solving the linear mixture model of two random genetic effects using the phenotypic data includes: The variance components of the linear mixture model of the two random genetic effects were estimated using the average information-restricted maximum likelihood algorithm. The model is considered converged when the absolute value of the difference between the log-likelihood values ​​of two consecutive iterations is less than the preset convergence threshold. The predicted values ​​of the shared genome effect and the predicted values ​​of the subgroup-specific deviation effect are obtained by using the converged variance components.

[0026] In some specific embodiments of the present invention, the step of performing principal component analysis on the genotype data to obtain the principal component characteristics of the pigs to be evaluated includes: A genome kinship matrix was constructed using the genotype data and its features were decomposed. The top 10 principal components were selected as the principal component features. The nonlinear dimensionality reduction of the principal component features using the uniform manifold approximation and projection method includes: using the number of nearest neighbor individuals and the minimum distance as dimensionality reduction parameters to map the principal component features to a two-dimensional low-dimensional genome feature space. The method of using density-based hierarchical clustering to cluster the low-dimensional genomic features includes: constructing a minimum spanning tree based on the distance between individuals, determining stable clusters based on the clustering stability in the hierarchical clustering tree, treating each stable cluster as a genetically homogeneous subgroup, and obtaining the total number of subgroups K and the subgroup affiliation of each pig to be evaluated.

[0027] In some specific embodiments of the present invention, validation of the method applicable to cross-population joint genetic assessment is also included: Methods suitable for joint genetic assessment across populations include: Obtain genotypic and phenotypic data of the pigs to be evaluated from multiple source groups; Unsupervised genome stratification was performed based on the genotype data to determine the genetically homogeneous subgroup to which each pig to be evaluated belonged; Based on the classification results of the genetically homogeneous subgroups and the genotype data, a first genomic kinship matrix is ​​constructed to describe the shared genomic kinship between different genetically homogeneous subgroups and within each genetically homogeneous subgroup, and a second genomic kinship matrix is ​​constructed to describe the specific genomic kinship of each genetically homogeneous subgroup. A linear mixture model of two random genetic effects is constructed based on the first genomic kinship matrix and the second genomic kinship matrix. The first random genetic effect is used to characterize the genomic effect shared among different genetically homogeneous subgroups, and the second random genetic effect is used to characterize the subgroup-specific deviation effect of each genetically homogeneous subgroup relative to the shared genomic effect. The phenotypic data were used to solve the linear mixed model of the two random genetic effects to obtain the predicted values ​​of the shared genome effect and the predicted values ​​of the subgroup-specific deviation effect for each pig to be evaluated. The sum of the predicted value of the shared genome effect and the predicted value of the subgroup-specific deviation effect for the same pig to be evaluated is determined as the estimated genomic breeding value of the pig to be evaluated. The single-population GBLUP method constructed according to the preset source population was used as the control model, and the above-mentioned method applicable to cross-population joint genetic assessment was used as the model to be validated. Repeated Monte Carlo cross-validation was used. 80% of the individuals in each initial source population were randomly selected as the training set, and the remaining 20% ​​of the individuals were selected as the test set. The phenotypic records of the test individuals were masked during model fitting. The two models used the same data partitioning and were independently repeated multiple times. The predictive ability was evaluated by the Pearson correlation coefficient between the estimated breeding values ​​and corrected phenotypic values ​​of the test set genome.

[0028] In some specific embodiments of the present invention, after determining the sum of the predicted value of the shared genome effect and the predicted value of the subgroup-specific deviation effect for the same pig to be evaluated as the estimated genomic breeding value of the pig to be evaluated, the method further includes: Output individual ID, inferred genetically homogeneous subpopulation, predicted value of shared genome effect, predicted value of subpopulation-specific deviation effect, and estimated breeding value of genome; Candidate breeding pigs are ranked and screened based on the estimated breeding values ​​from the genome.

[0029] Secondly, the present invention also provides a system based on the method, suitable for cross-population joint genetic assessment, comprising: The data acquisition module is used to acquire genotypic and phenotypic data of pigs to be evaluated from multiple source groups; An unsupervised stratification module is used to perform unsupervised genome stratification based on the genotype data to determine the genetically homogeneous subgroup to which each pig to be evaluated belongs; The kinship matrix construction module is used to construct, based on the division results of the genetically homogeneous subgroups and the genotype data, a first genomic kinship matrix describing the shared genomic kinship between different genetically homogeneous subgroups and within each genetically homogeneous subgroup, and a second genomic kinship matrix describing the specific genomic kinship of each genetically homogeneous subgroup; The model building module is used to construct a linear mixture model of double random genetic effects based on the first genomic kinship matrix and the second genomic kinship matrix, wherein the first random genetic effect is used to characterize the genomic effect shared among different genetically homogeneous subgroups, and the second random genetic effect is used to characterize the subgroup-specific deviation effect of each genetically homogeneous subgroup relative to the shared genomic effect. The model solving module is used to solve the linear mixed model of double random genetic effects using the phenotypic data to obtain the predicted values ​​of the shared genome effect and the predicted values ​​of the subgroup-specific deviation effect for each pig to be evaluated. The breeding value determination module is used to determine the sum of the predicted value of the shared genome effect and the predicted value of the subgroup-specific deviation effect for the same pig to be evaluated as the estimated genomic breeding value of the pig to be evaluated.

[0030] In some specific embodiments of the present invention, the unsupervised hierarchical module includes: Principal component analysis unit, used to perform principal component analysis on the genotype data to obtain the principal component characteristics of the pigs to be evaluated; The nonlinear dimensionality reduction unit is used to perform nonlinear dimensionality reduction on the principal component features using a uniform manifold approximation and projection method to obtain low-dimensional genome features. Clustering units are used to cluster the low-dimensional genomic features using a density-based hierarchical clustering method to obtain multiple genetically homogeneous subgroups.

[0031] In some specific embodiments of the present invention, the kinship matrix construction module includes: The centralized unit is used to construct a centralized genotype matrix for the k-th genetically homogeneous subpopulation obtained from unsupervised genome stratification, according to the following formula:

[0032] in, It is the first Pigs in a genetically homogeneous subgroup At the site Allele dosage on It is the first loci in genetically homogeneous subpopulations The frequency of encoding alleles; Standardized units are used to normalize the centralized genotype matrix according to the following formula. Standardize:

[0033] in, It is the first Sample size of each genetically homogeneous subpopulation Represents the trace of a matrix. Indicates matrix transpose; The splicing unit is used to assemble the standardized genotype matrix corresponding to each genetically homogeneous subpopulation. By vertically piecing together data according to the individual dimension, a combined standardized genotype matrix is ​​obtained:

[0034] The first matrix construction unit is used to construct the first genomic kinship matrix according to the following formula:

[0035] Among them, the first The blocks are:

[0036] in, It is the first The and the first Sample size of each genetically homogeneous subpopulation , The first genomic kinship matrix represents the total number of the genetically homogeneous subpopulations. It includes genomic kinship between individuals within the same genetically homogeneous subgroup and between individuals from different genetically homogeneous subgroups, and is used to construct the variance-covariance structure of the shared genomic effect; The second matrix construction unit is used to construct the second genome kinship matrix according to the following formula:

[0037] Among them, the Each diagonal piece is calculated using the following formula:

[0038] The second genomic kinship matrix Genomic kinship among pigs to be evaluated within the same genetically homogeneous subgroup is preserved, while subgroup-specific genomic kinship among individuals from different genetically homogeneous subgroups is set to zero, in order to construct the variance-covariance structure of the subgroup-specific deviation effect.

[0039] In some specific embodiments of the present invention, the linear mixture model of double random genetic effects constructed by the model building module is represented as follows:

[0040] in, This represents a phenotypic record vector from all genetically homogeneous subpopulations. Represents the fixed effects vector. Represents the fixed effects design matrix. This indicates the shared genome effect. This indicates the subgroup-specific deviation effect. These represent the association matrices that link phenotypic records to shared genome effects and subgroup-specific deviation effects, respectively. Represents a random residual vector; The shared genome effect, subgroup-specific deviation effect, and random residual vector respectively follow the following:

[0041] in, Indicates the variance of shared genome effects. This indicates the variance of the subgroup-specific deviation effect. Represents the residual variance. The unit matrix is ​​represented; the shared genome effect and the subgroup-specific deviation effect are independent of each other, and their covariance is zero.

[0042] In some specific embodiments of the present invention, the breeding value determination module determines the estimated genomic breeding value of the i-th pig to be evaluated according to the following formula:

[0043] in, Indicates the first Predicted values ​​of shared genome effects for 10 pigs to be evaluated. Indicates the first Predicted values ​​of subgroup-specific deviation effects for each pig to be evaluated.

[0044] Thirdly, the present invention also provides an apparatus suitable for joint genetic assessment across populations, comprising: Memory, used to store computer programs; A processor for implementing the steps of the method as described in the first aspect when executing the computer program.

[0045] Fourthly, the present invention also provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps of the method described in the first aspect.

[0046] Fifthly, the present invention also provides a computer program product, including a computer program that, when executed by a processor, implements the steps of the method described in the first aspect.

[0047] This invention discloses a method suitable for cross-population joint genetic assessment. The method acquires genotypic and phenotypic data of pigs from multiple source populations, infers population structure using an unsupervised genomic stratification method, and divides the individuals to be assessed into multiple genetically homogeneous subpopulations. Based on the inferred population structure, a linear mixture model with two random genetic effects is constructed, wherein a first random genetic effect characterizes the shared genomic effect among different subpopulations, and a second random genetic effect characterizes the subpopulation-specific deviation effect of each subpopulation relative to the shared genomic effect. The shared genomic effect and the subpopulation-specific deviation effect are predicted separately, and their sum is determined as the individual's estimated genomic breeding value. This invention can improve the accuracy of genetic assessment of important economic traits in pigs by utilizing genetic information from multiple populations without pre-specifying population tags. Attached Figure Description

[0048] The above and other objects, features, and advantages of this application will become more apparent from the more detailed description of the embodiments of this application in conjunction with the accompanying drawings. The drawings are provided to further illustrate the embodiments of this application and form part of the specification. They are used together with the embodiments of this application to explain this application and do not constitute a limitation thereof. In the drawings, the same reference numerals generally represent the same components or steps.

[0049] Figure 1 This is a flowchart illustrating a method for cross-population joint genetic evaluation according to an embodiment of this application; Figure 2 The diagram shows a comparison of the predictive capabilities of the method of this application embodiment and the single-population GBLUP method in correcting three traits: backfat thickness at 100 kg, feed conversion ratio, and total number of piglets. Model A represents the single-population GBLUP method constructed according to preset source populations, and Model B represents the method of this application embodiment. Detailed Implementation

[0050] This invention discloses a method suitable for cross-population joint genetic assessment. Those skilled in the art can refer to the content of this document and appropriately modify the process parameters to achieve the desired result. It should be particularly noted that all similar substitutions and modifications are obvious to those skilled in the art and are considered to be included in this invention. The method and application of this invention have been described through preferred embodiments. Those skilled in the art can obviously modify or appropriately change and combine the methods and applications described herein without departing from the content, spirit, and scope of this invention to implement and apply the technology of this invention.

[0051] In the description of this application, it should be understood that the terms "first," "second," etc., are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of technical features indicated. Therefore, a feature defined as "first" or "second" may explicitly or implicitly include one or more of that feature. In the description of this application, "multiple" means two or more, unless otherwise explicitly specified.

[0052] In this application, unless otherwise expressly specified and limited, the terms "installation," "connection," "linking," and "fixing," etc., should be interpreted broadly. For example, they can refer to a fixed connection, a detachable connection, or an integral part; they can refer to a mechanical connection, an electrical connection, or a connection that allows communication between components; they can refer to a direct connection or an indirect connection through an intermediate medium; they can refer to the internal communication between two components or the interaction between two components. Those skilled in the art can understand the specific meaning of the above terms in this application according to the specific circumstances.

[0053] In this application, unless otherwise expressly specified and limited, "above" or "below" the second feature can mean that the first feature is in direct contact with the second feature, or that the first feature is in indirect contact with the second feature through an intermediate medium. Furthermore, "above," "on top of," and "over" the second feature can mean that the first feature is directly above or diagonally above the second feature, or simply that the first feature is at a higher horizontal level than the second feature. "Below," "below," and "under" the second feature can mean that the first feature is directly below or diagonally below the second feature, or simply that the first feature is at a lower horizontal level than the second feature.

[0054] In the description of this specification, the references to terms such as "one embodiment," "some embodiments," "example," "specific example," or "some examples," etc., refer to specific features, structures, materials, or characteristics described in connection with that embodiment or example, which are included in at least one embodiment or example of this application. In this specification, the illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples. Moreover, without contradiction, those skilled in the art can combine and integrate the different embodiments or examples described in this specification, as well as the features of different embodiments or examples.

[0055] Although embodiments of this application have been shown and described above, it is understood that the above embodiments are exemplary and should not be construed as limiting this application. Those skilled in the art can make changes, modifications, substitutions and variations to the above embodiments within the scope of this application.

[0056] "Pigs to be evaluated": refers to individual pigs for which genomic estimation of breeding value prediction is required, including but not limited to breeding boars, breeding sows, replacement boars, replacement gilts, commercial pigs, piglets, fattening pigs, local breed pigs, developed breed pigs, introduced breed pigs, hybrid pigs, and purebred pigs. Pigs to be evaluated can come from one or more source populations, and their genotypic and phenotypic data can be used for joint genetic evaluation.

[0057] "Source population": refers to a pig herd distinguished by at least one of the following: breed, strain, breeding farm, breeding company, geographical origin, country or region, climate zone, feeding method, nutritional level, and health status. The source population can be an artificially defined population label or a genetically homogeneous subgroup inferred from genomic data. In this application, the source population and the genetically homogeneous subgroup can be the same concept or different concepts; when the source population and the genetically homogeneous subgroup are inconsistent, this application redetermines the genetically homogeneous subgroup through unsupervised genomic stratification.

[0058] A "genetically homogeneous subpopulation" refers to a group of pigs with a relatively consistent genetic background, inferred from genomic data. Individuals within the same genetically homogeneous subpopulation share similar allele frequencies, linkage disequilibrium patterns, selection signals, or pedigree structures; different genetically homogeneous subpopulations differ in these aspects. The number of genetically homogeneous subpopulations is denoted by K, where K is an integer greater than or equal to 2.

[0059] "Genographic data" refers to data describing the allele composition at genetic marker loci in the pigs to be evaluated, including but not limited to SNP markers, insertion / deletion markers, copy number variation markers, and structural variation markers. Genotype data can be encoded as 0, 1, and 2, representing homozygous, heterozygous, and homozygous alternative alleles, respectively. Genotype data can be obtained from SNP microarrays, resequencing, low-depth sequencing, high-throughput sequencing, capture sequencing, methylation sequencing, transcriptome sequencing, etc.

[0060] "Phenographic data" refers to data describing the observed economic traits of pigs to be evaluated, including but not limited to growth traits, reproductive traits, feed efficiency traits, carcass traits, and meat quality traits. Phenotypic data can be original observations, phenotypic values ​​corrected for known environmental effects, or phenotypic values ​​processed by standardization, mean removal, transformation, etc.

[0061] Unsupervised genomic stratification refers to the process of inferring population structure and classifying pigs under evaluation into multiple genetically homogeneous subpopulations based solely on genotype data, without relying on pre-specified population labels. Unsupervised genomic stratification may include steps such as principal component analysis, nonlinear dimensionality reduction, and cluster analysis, or may employ other methods capable of inferring population structure based on genotype data.

[0062] "Centered genotype matrix": This refers to the matrix obtained by centering the genotype matrix according to allele frequencies. Centering is used to eliminate the influence of population mean allele frequency differences on kinship estimation.

[0063] "Standardized genotype matrix": refers to the matrix obtained by standardizing the centralized genotype matrix according to the matrix trace and sample size. Standardization makes the average diagonal elements of the genomic kinship matrix within each subgroup equal to 1, thereby placing the genomic kinship of different genetically homogeneous subgroups on a comparable scale.

[0064] "Genomic kinship matrix": refers to a matrix describing the genomic-level kinship between individuals. This application includes two complementary genomic kinship matrices: the first genomic kinship matrix is ​​used to describe the shared genomic kinship between different genetically homogeneous subgroups and within each genetically homogeneous subgroup; the second genomic kinship matrix is ​​used to describe the specific genomic kinship between each genetically homogeneous subgroup.

[0065] "Shared genome effect": refers to the shared genome effect between different genetically homogeneous subpopulations, denoted as... The shared genome effect is used to characterize the shared genetic components among different subpopulations, and its variance-covariance structure is constructed from the first genome kinship matrix.

[0066] "Subgroup-specific deviation effect": refers to the subgroup-specific deviation effect of each genetically homogeneous subgroup relative to the shared genome effect. The subgroup-specific deviation effect is used to characterize the specific genetic components within each subgroup, and its variance-covariance structure is constructed by a second genome kinship matrix.

[0067] "Double random genetic effects linear mixture model": This refers to a linear mixture model that simultaneously incorporates two random genetic effects: shared genome effects and subgroup-specific deviation effects. This model can describe the shared genetic components among different genetically homogeneous subgroups and the specific genetic differences within each genetically homogeneous subgroup within the same genetic assessment model.

[0068] "Genome Estimated Breeding Value": refers to the estimated breeding value at the individual genome level, denoted as GEBV. In this application, the sum of the predicted value of the shared genome effect and the predicted value of the subgroup-specific deviation effect for the same pig to be evaluated is determined as the genome estimated breeding value for that pig.

[0069] "Fixed effects" refer to effects estimated as fixed parameters in a mixture model, including but not limited to field effects, year effects, seasonal effects, monthly effects, sex effects, parity effects, breed effects, strain effects, breeding firm effects, geographic origin effects, birth weight effects, weaning weight effects, age effects, litter effects, maternal effects, paternal effects, birth type effects, lactation effects, ambient temperature effects, ambient humidity effects, feed type effects, stocking density effects, health status effects, and immune status effects.

[0070] "Random effects" refer to effects estimated as random variables in a mixture model, including shared genome effects, subgroup-specific deviation effects, and random residual vectors. Random effects follow a specific distribution, and their variance-covariance structure is constructed from the corresponding genomic kinship matrix.

[0071] "Prediction accuracy" refers to the correlation between the estimated breeding values ​​and the actual breeding values ​​or corrected phenotypic values, and can be evaluated using indicators such as Pearson correlation coefficient, Spearman correlation coefficient, mean squared prediction error, and prediction bias. In this application, prediction accuracy is used to evaluate the effectiveness of cross-population joint genetic assessment.

[0072] Cross-validation refers to the process of dividing data into training and test sets, fitting a model on the training set, and evaluating its predictive ability on the test set. Cross-validation can include K-fold cross-validation, leave-one-out cross-validation, repeated Monte Carlo cross-validation, bootstrap cross-validation, etc. In this application, repeated Monte Carlo cross-validation is used to evaluate the predictive performance of the method.

[0073] "Single-population GBLUP method": refers to the optimal linear unbiased prediction method for single-population genomes constructed according to preset source populations. In this application, the single-population GBLUP method can be used as a control model to evaluate the improvement in prediction accuracy of the method in this application.

[0074] "Multi-trait model": refers to a genetic evaluation model that analyzes multiple traits simultaneously, describing the genetic correlation between traits through the covariance matrix between traits.

[0075] "Generalized linear mixed model": refers to a mixed model used when the response variable does not follow a normal distribution, including logistic regression model, Poisson regression model, survival analysis model, etc. The generalized linear mixed model can extend the double random genetic effects linear mixed model of this application to handle non-normal phenotypes such as binary traits, count traits, and survival traits.

[0076] "Threshold model": refers to a model that compares a latent continuous variable with a threshold to obtain a classification result, used to handle categorical traits. The threshold model can extend the linear mixture model of two random genetic effects in this application to handle categorical traits.

[0077] "Spatial model": refers to a model that incorporates spatial correlation structures into a mixture model to handle phenotypic data with spatial correlation. The spatial model can extend the linear mixture model of double random genetic effects in this application to handle phenotypic data with spatial correlation.

[0078] "Time series model": refers to a model that incorporates time-related structures into a mixture model for processing longitudinal phenotypic data. The time series model can extend the linear mixture model with two random genetic effects proposed in this application for processing longitudinal phenotypic data.

[0079] "Multi-omics model" refers to a model that integrates multi-omics data such as genomics, transcriptomics, metabolomics, and proteomics. The multi-omics model can extend the linear mixture model of double random genetic effects proposed in this application to integrate multi-omics data.

[0080] "Machine learning methods" refers to methods that use random forests, gradient boosting trees, support vector regression, neural networks, deep learning models, etc., to predict random genetic effects. Machine learning methods can be used as an alternative to solving the model in this application.

[0081] "Distributed computing method": refers to a method that divides large-scale data into blocks for processing and obtains a global solution through weighted combination. The distributed computing method can be used to accelerate the solution process of the model in this application.

[0082] "Computer-readable storage medium" refers to a medium capable of storing computer programs, including but not limited to hard disks, floppy disks, optical disks, digital versatile optical disks, flash memory, solid-state drives, read-only memory, and random access memory. A computer-readable storage medium can carry a computer program that implements the method of this application.

[0083] "Computer program product": refers to a product that includes a computer program and can be provided to users through methods such as network download, CD distribution, pre-installation, etc. The computer program product can run on a local computing device or a cloud computing platform, and implements the steps of the method described in this application when executed.

[0084] "Processor": refers to hardware capable of executing computer programs, including but not limited to central processing units, microprocessors, digital signal processors, application-specific integrated circuits, field-programmable gate arrays, or other programmable logic devices. A processor can be used to perform the steps of the methods described in this application.

[0085] "Memory" refers to hardware capable of storing computer programs, including but not limited to read-only memory, random access memory, flash memory, hard disk, optical disk, or other computer-readable storage media. Memory can be used to store computer programs that implement the methods of this application.

[0086] The technical solution adopted in this application includes: According to one aspect of this application, a method suitable for joint genetic assessment across populations is provided, comprising: Obtain genotypic and phenotypic data of the pigs to be evaluated from multiple source groups; Unsupervised genome stratification was performed based on the genotype data to determine the genetically homogeneous subgroup to which each pig to be evaluated belonged; Based on the classification results of the genetically homogeneous subgroups and the genotype data, a first genomic kinship matrix is ​​constructed to describe the shared genomic kinship between different genetically homogeneous subgroups and within each genetically homogeneous subgroup, and a second genomic kinship matrix is ​​constructed to describe the specific genomic kinship of each genetically homogeneous subgroup. A linear mixture model of two random genetic effects is constructed based on the first genomic kinship matrix and the second genomic kinship matrix. The first random genetic effect is used to characterize the genomic effect shared among different genetically homogeneous subgroups, and the second random genetic effect is used to characterize the subgroup-specific deviation effect of each genetically homogeneous subgroup relative to the shared genomic effect. The phenotypic data were used to solve the linear mixed model of the two random genetic effects to obtain the predicted values ​​of the shared genome effect and the predicted values ​​of the subgroup-specific deviation effect for each pig to be evaluated. The sum of the predicted value of the shared genome effect and the predicted value of the subgroup-specific deviation effect for the same pig to be evaluated is determined as the estimated genomic breeding value of that pig.

[0087] Furthermore, the unsupervised genome stratification based on genotype data includes: Principal component analysis was performed on the genotype data to obtain the principal component characteristics of the pigs to be evaluated; The principal component features are nonlinearly reduced in dimensionality using a uniform manifold approximation and projection method to obtain low-dimensional genome features. The low-dimensional genomic features were clustered using a density-based hierarchical clustering method to obtain multiple genetically homogeneous subgroups.

[0088] Furthermore, regarding the first [unsupervised genome stratification obtained] The genotype matrix of the genetically homogeneous subgroups is constructed according to the following formula:

[0089] in, It is the first Pigs in a genetically homogeneous subgroup At the site Allele dosage on It is the first loci in genetically homogeneous subpopulations The frequency of encoded alleles.

[0090] Furthermore, the first genomic kinship matrix is ​​constructed according to the following steps: The centralized genotype matrix is ​​processed according to the following formula. Standardize:

[0091] in, It is the first Sample size of each genetically homogeneous subpopulation Represents the trace of a matrix. Indicates matrix transpose; Standardized genotype matrix corresponding to each genetically homogeneous subpopulation By vertically piecing together data according to the individual dimension, a combined standardized genotype matrix is ​​obtained:

[0092] The first genome phylogenetic matrix is ​​constructed according to the following formula:

[0093] Among them, the first The blocks are:

[0094] in, It is the first The and the first Sample size of each genetically homogeneous subpopulation , The first genomic kinship matrix represents the total number of the genetically homogeneous subpopulations. It includes genomic kinship between individuals within the same genetically homogeneous subgroup and between individuals from different genetically homogeneous subgroups, and is used to construct the variance-covariance structure of the shared genomic effect.

[0095] Furthermore, the second genomic kinship matrix is ​​constructed according to the following steps:

[0096] Among them, the Each diagonal piece is calculated using the following formula:

[0097] The second genomic kinship matrix Genomic kinship among pigs to be evaluated within the same genetically homogeneous subgroup is preserved, while subgroup-specific genomic kinship among individuals from different genetically homogeneous subgroups is set to zero, in order to construct the variance-covariance structure of the subgroup-specific deviation effect.

[0098] Furthermore, the linear mixture model of the two random genetic effects is expressed as:

[0099] in, This represents a phenotypic record vector from all genetically homogeneous subpopulations. Represents the fixed effects vector. Represents the fixed effects design matrix. This indicates the shared genome effect. This indicates the subgroup-specific deviation effect. These represent the association matrices that link phenotypic records to shared genome effects and subgroup-specific deviation effects, respectively. This represents a random residual vector.

[0100] Furthermore, the shared genome effect, subgroup-specific deviation effect, and random residual vector respectively follow the following rules:

[0101] in, Indicates the variance of shared genome effects. This indicates the variance of the subgroup-specific deviation effect. Represents the residual variance. The unit matrix is ​​represented; the shared genome effect and the subgroup-specific deviation effect are independent of each other, and their covariance is zero.

[0102] Furthermore, the first is determined according to the following formula. Genomic estimated breeding value of the pigs to be evaluated:

[0103] in, Indicates the first Predicted values ​​of shared genome effects for 10 pigs to be evaluated. Indicates the first Predicted values ​​of subgroup-specific deviation effects for each pig to be evaluated.

[0104] Furthermore, the multiple source groups are pig herds distinguished by at least one of breed, strain, breeding farm, breeding enterprise, and geographical origin; the phenotypic data include at least one of growth traits, reproductive traits, feed efficiency traits, carcass traits, and meat quality traits.

[0105] The beneficial effects of this invention include: 1. The method proposed in this application for joint genetic assessment across populations is based on unsupervised genome stratification using genotype data of the pigs to be assessed. It does not require pre-classification of populations according to source tags such as breed, strain, or breeding company. It can identify potential genetically homogeneous subpopulations in mixed populations, thereby reducing the impact of missing or incorrect source tags or inconsistencies between source tags and the actual genetic structure on population structure analysis.

[0106] 2. This application proposes a method for cross-population joint genetic assessment that constructs two complementary genomic kinship matrices based on unsupervised genomic stratification results. The first genomic kinship matrix includes genomic kinship relationships between individuals within the same genetically homogeneous subgroup and between individuals from different genetically homogeneous subgroups, used to construct a variance-covariance structure for shared genomic effects. The second genomic kinship matrix adopts a block-diagonal structure, retaining specific genomic kinship relationships between individuals within the same genetically homogeneous subgroup and setting subgroup-specific genomic kinship relationships between individuals from different genetically homogeneous subgroups to zero, used to construct a variance-covariance structure for subgroup-specific deviation effects. Thus, it is possible to describe the shared genetic components between different genetically homogeneous subgroups and the specific genetic differences within each genetically homogeneous subgroup separately within the same genetic assessment model.

[0107] 3. The method for cross-population joint genetic assessment proposed in this application utilizes a linear mixture model of two random genetic effects to simultaneously leverage shared genetic information across subpopulations and subpopulation-specific genetic information. Compared to methods that perform genetic assessments based on pre-defined labels, this invention increases the available genetic information and expands the population size. Compared to methods that directly merge data from multiple populations, this invention can further describe specific genetic deviations within different genetically homogeneous subpopulations, thereby improving the model's ability to analyze complex population structures.

[0108] 4. This application proposes a method for cross-population joint genetic assessment that uses the sum of predicted values ​​of shared genome effects and subpopulation-specific deviation effects for the pigs to be assessed as the genome-estimated breeding value. This ensures that the obtained genome-estimated breeding value simultaneously includes cross-subpopulation shared genetic information and subpopulation-specific genetic information, thereby improving the accuracy of cross-population joint genetic assessment and providing technical support for the joint genetic assessment of important economic traits in pig herds of different breeds, strains, breeding companies, or geographical origins. In the validation dataset of this application, compared with single-population genetic assessment methods, the prediction accuracy of this invention in correcting for 100 kg backfat thickness, feed conversion ratio, and total litter size traits is improved by 6.51%, 6.14%, and 7.34%, respectively, indicating that this invention can improve the accuracy of genetic assessment of important economic traits in pigs under complex population structures.

[0109] The raw materials and reagents used in the method for cross-population joint genetic assessment provided by this invention are all commercially available.

[0110] The present invention will be further illustrated below with reference to the embodiments: Application Overview like Figure 1 As shown, the method in this application embodiment takes genotype and phenotypic data of pigs to be evaluated from multiple source groups as input, infers the potential population structure in the mixed population based on the genotype data, and determines the genetically homogeneous subgroup to which each pig to be evaluated belongs.

[0111] After obtaining the results of the genetically homogeneous subgroup classification, a first genomic kinship matrix and a second genomic kinship matrix were constructed. The first genomic kinship matrix includes genomic kinship relationships between pigs within the same genetically homogeneous subgroup and between pigs from different genetically homogeneous subgroups, and is used to construct the variance-covariance structure of shared genomic effects. The second genomic kinship matrix adopts a block diagonal structure, retaining subgroup-specific genomic kinship relationships between pigs within the same genetically homogeneous subgroup, and setting the subgroup-specific genomic kinship relationships between pigs from different genetically homogeneous subgroups to zero, and is used to construct the variance-covariance structure of subgroup-specific deviation effects.

[0112] Two genomic kinship matrices were simultaneously incorporated into a linear mixture model of two random genetic effects to predict the shared genome effect and subgroup-specific deviation effect of each pig to be evaluated, and the sum of the two predicted values ​​was determined as the estimated genomic breeding value of the pig to be evaluated.

[0113] Therefore, the method in this application embodiment does not require the use of preset varieties, strains, or breeding company labels to determine subpopulation affiliation, and can simultaneously utilize the genetic information shared between subpopulations and the specific genetic information within subpopulations in the same model.

[0114] After introducing the basic principles of this application, various non-limiting embodiments of this application will be described in detail below with reference to the accompanying drawings.

[0115] Exemplary methods A method for cross-population joint genetic assessment according to an embodiment of this application includes the following steps: S101: Acquire and process genotype and phenotypic data.

[0116] Genotypic and phenotypic data of pigs to be evaluated from multiple source populations were obtained, and a correspondence between genotypic and phenotypic data was established based on individual identification numbers. Genotypic data included allele doses at multiple genetic marker loci for each pig to be evaluated, with allele doses coded as 0, 1, or 2. Phenotypic data included observed values ​​of the economic traits to be evaluated or phenotypic values ​​corrected for known environmental effects.

[0117] In one specific embodiment, data were collected from three pig breeding companies in China, covering three pig breeds, with target traits including corrected 100 kg backfat thickness, feed conversion ratio, and total number of piglets born.

[0118] S102: Unsupervised genomic stratification of pigs to be evaluated.

[0119] A genomic kinship matrix was constructed using genotype data and eigenvalue decomposition was performed to obtain the principal component features of each pig to be evaluated. The principal component feature vectors were then input into a uniform manifold approximation and projection algorithm, using the nearest neighbor number and minimum distance as dimensionality reduction parameters to map the principal component features to a two-dimensional low-dimensional genomic feature space. Subsequently, a hierarchical noise density space clustering algorithm was applied to cluster the two-dimensional low-dimensional genomic features. This clustering algorithm constructed a minimum spanning tree based on the cross-distance between individuals and determined stable clusters based on the clustering stability in the hierarchical clustering tree. Each stable cluster was considered a genetically homogeneous subpopulation, and the total number of subpopulations was obtained. And the subgroup affiliation of each pig to be evaluated.

[0120] In this embodiment, the first 10 principal components are selected for unsupervised genomic clustering analysis to calculate the total number of subgroups. And the subgroup affiliation of each pig to be evaluated.

[0121] S103: Construct a centralized and standardized genotype matrix for each genetically homogeneous subpopulation.

[0122] For the A centralized genotype matrix is ​​constructed from three genetically homogeneous subgroups according to the following formula:

[0123] in, It is the first Pigs in a genetically homogeneous subgroup At the site Allele dosage on It is the first loci in genetically homogeneous subpopulations The coding allele frequencies. The genotype matrix corresponding to each genetically homogeneous subpopulation contains the same genetic marker loci, which are arranged in the same order and use the same allele coding direction.

[0124] The centralized genotype matrix is ​​standardized according to the following formula:

[0125] in, It is the first Sample size of each genetically homogeneous subpopulation Represents the trace of a matrix. This represents the matrix transpose. The standardization makes the average diagonal elements of the genomic kinship matrix within each subgroup equal to 1, thereby placing the genomic kinship of different genetically homogeneous subgroups on a comparable scale.

[0126] S104: Construct the first genome kinship matrix.

[0127] Will The standardized genotype matrix of each genetically homogeneous subpopulation is assembled vertically along the individual dimension:

[0128] The first genome kinship matrix was constructed based on the combinatorial normalized genotype matrix:

[0129] The first genome kinship matrix Each block is represented as:

[0130] in, It is the first The and the first Sample size of each genetically homogeneous subpopulation , The first genomic kinship matrix represents the total number of the genetically homogeneous subpopulations. It includes genomic kinship between individuals within the same genetically homogeneous subgroup and between individuals from different genetically homogeneous subgroups, and is used to construct the variance-covariance structure of the shared genomic effect.

[0131] S105: Construct a second genome kinship matrix.

[0132] Construct a block diagonal matrix consisting of the genomic kinship matrix within each genetically homogeneous subpopulation according to the following formula:

[0133] Among them, the Each diagonal piece is calculated using the following formula:

[0134] The second genomic kinship matrix preserves the genomic kinship among pigs to be evaluated within the same genetically homogeneous subgroup, and sets the subgroup-specific genomic kinship between pigs from different genetically homogeneous subgroups to zero, in order to construct the variance-covariance structure of subgroup-specific deviation effects.

[0135] S106: Construct and solve a linear mixture model of two random genetic effects.

[0136] The first and second genomic kinship matrices are simultaneously incorporated into the following linear mixture model of two random genetic effects:

[0137] in, This represents a phenotypic record vector from all genetically homogeneous subpopulations. Represents the fixed effects vector. Represents the fixed effects design matrix. This indicates the shared genome effect. This indicates the subgroup-specific deviation effect. These represent the association matrices that link phenotypic records to shared genome effects and subgroup-specific deviation effects, respectively. This represents a random residual vector.

[0138] Shared genome effects, subgroup-specific deviation effects, and random residual vectors each follow the following rules:

[0139] in, Indicates the variance of shared genome effects. This indicates the variance of the subgroup-specific deviation effect. Represents the residual variance. The unit matrix is ​​represented; the shared genome effect and the subgroup-specific deviation effect are independent of each other, and their covariance is zero.

[0140] In this embodiment, the average information-restricted maximum likelihood algorithm is used to estimate the variance components. When the absolute value of the difference between the log-likelihood values ​​of two adjacent iterations is less than a preset convergence threshold, the model is considered to have converged. The predicted values ​​of shared genome effects and subgroup-specific deviation effects are obtained using the converged variance components. Other solution methods capable of estimating the above variance components and predicting random genetic effects can also be used to implement this application.

[0141] S107: Determine the estimated breeding value based on the genome.

[0142] According to the following formula, the first The sum of the predicted values ​​of shared genome effects and subgroup-specific deviation effects for each pig to be evaluated:

[0143] in, Indicates the first Predicted values ​​of shared genome effects for 10 pigs to be evaluated. Indicates the first The predicted subgroup-specific deviation effects for each pig to be evaluated are calculated. Output includes individual number, inferred genetically homogeneous subgroup, predicted shared genome effect, predicted subgroup-specific deviation effect, and estimated genomic breeding value. Candidate breeding pigs can be ranked and screened based on the estimated genomic breeding value.

[0144] S108: Validate the predictive performance of the method.

[0145] In this embodiment's dataset, a single-population GBLUP constructed according to preset source populations is used as model A, and the method of this application embodiment is used as model B. Repeated Monte Carlo cross-validation is employed: 80% of individuals within each initial source population are randomly selected as the training set, and the remaining 20% ​​are used as the test set; phenotypic records of the test individuals are masked during model fitting, and both models use the same data partitioning, with independent repetitions 20 times. The predictive ability is evaluated using the Pearson correlation coefficient between the estimated breeding values ​​and corrected phenotypic values ​​in the test set.

[0146] like Figure 2 As shown, compared with model A, model B improved the average predictive ability of 100 kg backfat thickness (BF100), feed conversion ratio (FCR), and total number of piglets born (TNB) by 6.51%, 6.14%, and 7.34%, respectively. This result indicates that identifying genetically homogeneous subpopulations through unsupervised genomic stratification and constructing a double-random genetic effect model using two complementary genomic kinship matrices can improve the accuracy of cross-population joint genetic assessment of important economic traits in pigs under complex population structures.

[0147] The basic principles of this application have been described above with reference to specific embodiments. However, it should be noted that the advantages, benefits, and effects mentioned in this application are merely examples and not limitations, and should not be considered as essential features of each embodiment of this application. Furthermore, the specific details disclosed above are for illustrative and facilitative purposes only, and are not limitations. These details do not restrict the application from being implemented using the specific details described above.

[0148] The above description of the disclosed aspects is provided to enable any person skilled in the art to make or use this application. Various modifications to these aspects will be readily apparent to those skilled in the art, and the general principles defined herein can be applied to other aspects without departing from the scope of this application. Therefore, this application is not intended to be limited to the aspects shown herein, but rather to be accorded the widest scope consistent with the principles and novel features disclosed herein.

[0149] The above description has been given for purposes of illustration and description. Furthermore, this description is not intended to limit the embodiments of this application to the forms disclosed herein. Although numerous exemplary aspects and embodiments have been discussed above, those skilled in the art will recognize certain variations, modifications, alterations, additions, and sub-combinations thereof.

Claims

1. A method suitable for joint genetic assessment across populations, characterized in that, include: Obtain genotypic and phenotypic data of the pigs to be evaluated from multiple source groups; Unsupervised genome stratification was performed based on the genotype data to determine the genetically homogeneous subgroup to which each pig to be evaluated belonged; Based on the classification results of the genetically homogeneous subgroups and the genotype data, a first genomic kinship matrix is ​​constructed to describe the shared genomic kinship between different genetically homogeneous subgroups and within each genetically homogeneous subgroup, and a second genomic kinship matrix is ​​constructed to describe the specific genomic kinship of each genetically homogeneous subgroup. A linear mixture model of two random genetic effects is constructed based on the first genomic kinship matrix and the second genomic kinship matrix. The first random genetic effect is used to characterize the genomic effect shared among different genetically homogeneous subgroups, and the second random genetic effect is used to characterize the subgroup-specific deviation effect of each genetically homogeneous subgroup relative to the shared genomic effect. The phenotypic data were used to solve the linear mixed model of the two random genetic effects to obtain the predicted values ​​of the shared genome effect and the predicted values ​​of the subgroup-specific deviation effect for each pig to be evaluated. The sum of the predicted value of the shared genome effect and the predicted value of the subgroup-specific deviation effect for the same pig to be evaluated is determined as the estimated genomic breeding value of that pig.

2. The method according to claim 1, characterized in that, The unsupervised genome stratification based on the genotype data includes: Principal component analysis was performed on the genotype data to obtain the principal component characteristics of the pigs to be evaluated; The principal component features are nonlinearly reduced in dimensionality using a uniform manifold approximation and projection method to obtain low-dimensional genome features. The low-dimensional genomic features were clustered using a density-based hierarchical clustering method to obtain multiple genetically homogeneous subgroups.

3. The method according to claim 1, characterized in that, The first obtained from unsupervised genome stratification The genotype matrix of each genetically homogeneous subgroup is constructed according to the following formula: in, It is the first Pigs in a genetically homogeneous subgroup At the site Allele dosage on It is the first loci in genetically homogeneous subpopulations The frequency of encoded alleles.

4. The method according to claim 3, characterized in that, The first genome kinship matrix is ​​constructed according to the following steps: The centralized genotype matrix is ​​processed according to the following formula. Standardize: in, It is the first Sample size of each genetically homogeneous subpopulation Represents the trace of a matrix. Indicates matrix transpose; Standardized genotype matrix corresponding to each genetic homogeneous subpopulation By vertically piecing together data according to the individual dimension, a combined standardized genotype matrix is ​​obtained: The first genome phylogenetic matrix is ​​constructed according to the following formula: Among them, the first The blocks are: in, It is the first The and the first Sample size of each genetically homogeneous subpopulation , The first genomic kinship matrix represents the total number of the genetically homogeneous subpopulations. It includes genomic kinship between individuals within the same genetically homogeneous subgroup and between individuals from different genetically homogeneous subgroups, and is used to construct the variance-covariance structure of the shared genomic effect.

5. The method according to claim 4, characterized in that, The second genome kinship matrix is ​​constructed according to the following steps: Among them, the Each diagonal piece is calculated using the following formula: The second genomic kinship matrix Genomic kinship among pigs to be evaluated within the same genetically homogeneous subgroup is preserved, while subgroup-specific genomic kinship among individuals from different genetically homogeneous subgroups is set to zero, in order to construct the variance-covariance structure of the subgroup-specific deviation effect.

6. The method according to claim 5, characterized in that, The linear mixture model of the two random genetic effects is expressed as follows: in, This represents a phenotypic record vector from all genetically homogeneous subpopulations. Represents the fixed effects vector. Represents the fixed effects design matrix. This indicates the shared genome effect. This indicates the subgroup-specific deviation effect. These represent the association matrices that link phenotypic records to shared genome effects and subgroup-specific deviation effects, respectively. This represents a random residual vector.

7. The method according to claim 6, characterized in that, The shared genome effect, subgroup-specific deviation effect, and random residual vector respectively follow the following: in, Indicates the variance of shared genome effects. This indicates the variance of the subgroup-specific deviation effect. Represents the residual variance. The unit matrix is ​​represented; the shared genome effect and the subgroup-specific deviation effect are independent of each other, and their covariance is zero.

8. The method according to claim 7, characterized in that, The number is determined according to the following formula. Genomic estimated breeding value of the pigs to be evaluated: in, Indicates the first Predicted values ​​of shared genome effects for 10 pigs to be evaluated. Indicates the first Predicted values ​​of subgroup-specific deviation effects for each pig to be evaluated.

9. The method according to claim 1, characterized in that, The multiple source groups are pig herds that are distinguished according to at least one of the following: breed, strain, breeding farm, breeding enterprise, or geographical origin; The phenotypic data include at least one of growth traits, reproductive traits, feed efficiency traits, carcass traits, or meat quality traits.

10. The method according to claim 1, characterized in that, The step of solving the linear mixture model of two random genetic effects using the phenotypic data includes: The variance components of the linear mixture model of the two random genetic effects were estimated using the average information-restricted maximum likelihood algorithm. The model is considered converged when the absolute value of the difference between the log-likelihood values ​​of two consecutive iterations is less than the preset convergence threshold. The predicted values ​​of the shared genome effect and the predicted values ​​of the subgroup-specific deviation effect are obtained by using the converged variance components.

11. The method according to claim 2, characterized in that, The principal component analysis of the genotype data to obtain the principal component characteristics of the pigs to be evaluated includes: A genome kinship matrix was constructed using the genotype data and its features were decomposed. The top 10 principal components were selected as the principal component features. The nonlinear dimensionality reduction of the principal component features using the uniform manifold approximation and projection method includes: using the number of nearest neighbor individuals and the minimum distance as dimensionality reduction parameters to map the principal component features to a two-dimensional low-dimensional genome feature space. The method of using density-based hierarchical clustering to cluster the low-dimensional genomic features includes: constructing a minimum spanning tree based on the distance between individuals, determining stable clusters based on the clustering stability in the hierarchical clustering tree, treating each stable cluster as a genetically homogeneous subgroup, and obtaining the total number of subgroups K and the subgroup affiliation of each pig to be evaluated.

12. The method according to claim 1, characterized in that, It also includes verifying the method: The single-population GBLUP method constructed according to the preset source groups is used as a control model, and the method described in claim 1 is used as the model to be verified. Repeated Monte Carlo cross-validation was used. 80% of the individuals in each initial source population were randomly selected as the training set, and the remaining 20% ​​of the individuals were selected as the test set. The phenotypic records of the test individuals were masked during model fitting. The two models used the same data partitioning and were independently repeated multiple times. The predictive ability was evaluated by the Pearson correlation coefficient between the estimated breeding values ​​and corrected phenotypic values ​​of the test set genome.

13. The method according to claim 1, characterized in that, After determining the sum of the predicted value of the shared genome effect and the predicted value of the subgroup-specific deviation effect for the same pig to be evaluated as the estimated genomic breeding value of the pig to be evaluated, the method further includes: Output individual ID, inferred genetically homogeneous subpopulation, predicted value of shared genome effect, predicted value of subpopulation-specific deviation effect, and estimated breeding value of genome; Candidate breeding pigs are ranked and screened based on the estimated breeding values ​​from the genome.

14. A system for cross-population joint genetic assessment based on the method of any one of claims 1 to 13, characterized in that, include: The data acquisition module is used to acquire genotypic and phenotypic data of the pigs to be evaluated from multiple source groups; An unsupervised stratification module is used to perform unsupervised genome stratification based on the genotype data to determine the genetically homogeneous subgroup to which each pig to be evaluated belongs; The kinship matrix construction module is used to construct, based on the division results of the genetically homogeneous subgroups and the genotype data, a first genomic kinship matrix describing the shared genomic kinship between different genetically homogeneous subgroups and within each genetically homogeneous subgroup, and a second genomic kinship matrix describing the specific genomic kinship of each genetically homogeneous subgroup; The model building module is used to construct a linear mixture model of double random genetic effects based on the first genomic kinship matrix and the second genomic kinship matrix, wherein the first random genetic effect is used to characterize the genomic effect shared among different genetically homogeneous subgroups, and the second random genetic effect is used to characterize the subgroup-specific deviation effect of each genetically homogeneous subgroup relative to the shared genomic effect. The model solving module is used to solve the linear mixed model of double random genetic effects using the phenotypic data to obtain the predicted values ​​of the shared genome effect and the predicted values ​​of the subgroup-specific deviation effect for each pig to be evaluated. The breeding value determination module is used to determine the sum of the predicted value of the shared genome effect and the predicted value of the subgroup-specific deviation effect for the same pig to be evaluated as the estimated genomic breeding value of the pig to be evaluated.

15. The system according to claim 14, characterized in that, The unsupervised hierarchical module includes: Principal component analysis unit, used to perform principal component analysis on the genotype data to obtain the principal component characteristics of the pigs to be evaluated; The nonlinear dimensionality reduction unit is used to perform nonlinear dimensionality reduction on the principal component features using a uniform manifold approximation and projection method to obtain low-dimensional genome features. Clustering units are used to cluster the low-dimensional genomic features using a density-based hierarchical clustering method to obtain multiple genetically homogeneous subgroups.

16. The system according to claim 14, characterized in that, The kinship matrix construction module includes: The centralized unit is used to construct a centralized genotype matrix for the k-th genetically homogeneous subpopulation obtained from unsupervised genome stratification, according to the following formula: in, It is the first Pigs in a genetically homogeneous subgroup At the site Allele dosage on It is the first loci in genetically homogeneous subpopulations The frequency of encoding alleles; Standardized units are used to normalize the centralized genotype matrix according to the following formula. Standardize: in, It is the first Sample size of each genetically homogeneous subpopulation Represents the trace of a matrix. Indicates matrix transpose; The splicing unit is used to assemble the standardized genotype matrix corresponding to each genetically homogeneous subpopulation. By vertically piecing together data according to the individual dimension, a combined standardized genotype matrix is ​​obtained: The first matrix construction unit is used to construct the first genomic kinship matrix according to the following formula: Among them, the first The blocks are: in, It is the first The and the first Sample size of each genetically homogeneous subpopulation , The first genomic kinship matrix represents the total number of the genetically homogeneous subpopulations. It includes genomic kinship between individuals within the same genetically homogeneous subgroup and between individuals from different genetically homogeneous subgroups, and is used to construct the variance-covariance structure of the shared genomic effect; The second matrix construction unit is used to construct the second genome kinship matrix according to the following formula: Among them, the Each diagonal piece is calculated using the following formula: The second genomic kinship matrix Genomic kinship among pigs to be evaluated within the same genetically homogeneous subgroup is preserved, while subgroup-specific genomic kinship among individuals from different genetically homogeneous subgroups is set to zero, in order to construct the variance-covariance structure of the subgroup-specific deviation effect.

17. The system according to claim 14, characterized in that, The linear mixture model of two random genetic effects constructed by the model building module is expressed as follows: in, This represents a phenotypic record vector from all genetically homogeneous subpopulations. Represents the fixed effects vector. Represents the fixed effects design matrix. This indicates the shared genome effect. This indicates the subgroup-specific deviation effect. These represent the association matrices that link phenotypic records to shared genome effects and subgroup-specific deviation effects, respectively. Represents a random residual vector; The shared genome effect, subgroup-specific deviation effect, and random residual vector respectively follow the following: in, Indicates the variance of shared genome effects. This indicates the variance of the subgroup-specific deviation effect. Represents the residual variance. The unit matrix is ​​represented; the shared genome effect and the subgroup-specific deviation effect are independent of each other, and their covariance is zero.

18. The system according to claim 14, characterized in that, The breeding value determination module determines the estimated genomic breeding value of the i-th pig to be evaluated according to the following formula: in, Indicates the first Predicted values ​​of shared genome effects for 10 pigs to be evaluated. Indicates the first Predicted values ​​of subgroup-specific deviation effects for each pig to be evaluated.

19. A device suitable for joint genetic assessment across populations, characterized in that, include: Memory, used to store computer programs; A processor for executing the computer program to implement the steps of the method as described in any one of claims 1 to 13.

20. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by a processor, it implements the steps of the method as described in any one of claims 1 to 13.

21. A computer program product, comprising a computer program, characterized in that, When the computer program is executed by a processor, it implements the steps of the method as described in any one of claims 1 to 13.

Images