Breeding selection optimization method, system, equipment and medium based on BLUP model

Abstract

The present invention discloses a breeding selection optimization method, system, equipment and medium based on a BLUP model, and relates to the technical field of animal breeding. The method comprises: constructing a BLUP model based on animal pedigree information; optimizing and solving the BLUP model using a block method to obtain an unbiased optimal prediction value of the animal's genetic potential; the unbiased optimal prediction value is used for breeding selection; the optimization solution process of the BLUP model includes two matrix inversion operations, one is the inversion operation of the kinship matrix, and the other is the inversion operation of a larger matrix containing fixed effects. The present invention can significantly reduce the memory load in the BLUP model solution process by adopting an iterative block calculation method, while improving the calculation efficiency and maintaining the accuracy of breeding value estimation.

Description

Technical Field

[0001] The present invention relates to the technical field of animal breeding, and in particular to a breeding selection optimization method, system, equipment and medium based on a BLUP model. Background Art

[0002] Animal breeding emphasizes the improvement and enhancement of genetic structure. The scientific theory of modern breeding can be traced back to the rediscovery of Mendel's laws of inheritance in 1900, which enabled breeding work to be carried out in accordance with scientific genetic laws. Systematic breeding of domestic animals made a huge leap forward after World War II. The BLUP (Best Linear Unbiased Prediction) developed by American statistician CR Henderson in the 1960s made a decisive contribution to this in terms of computing technology. The BLUP model optimizes breeding selection by estimating the unbiased best prediction value (EBV). Using BLUP to evaluate the genetic potential of animals is a standard method commonly adopted by countries around the world.

[0003] As the cost of sequencing technology decreases, the amount of data obtained from gene chip sequencing is increasing, and the application of the GBLUP model in animal breeding has become more extensive. However, this increase in data volume also brings new challenges: when the number of animals becomes extremely large, traditional computing methods face huge computing pressure and resource limitations. The processing of large-scale data sets requires a lot of computing resources, especially when it involves complex genetic evaluations and large matrix operations. Therefore, studying how to improve the computational efficiency of the BLUP model on large-scale data sets while maintaining its accuracy has become an urgent problem that needs to be solved.

[0004] The traditional BLUP model has a problem of low computational efficiency when solving EBV on large-scale data sets, mainly because it involves two matrix inversion operations, one of which is the inverse matrix of the kinship matrix and the other is the inverse matrix of a larger matrix containing fixed effects. Matrix inversion is a computationally expensive step, so optimizing this process is crucial to improving the efficiency of BLUP model solutions.

[0005] In ABLUP, the first matrix inversion is the inversion of the A matrix. In 1975, Henderson proposed that A can be constructed directly from the pedigree. -1 A simple method, iterative method directly constructs A -1 The first matrix inversion is avoided, which improves the computational efficiency of ABLUP to a certain extent. In GBLUP, the Apy algorithm proposes to divide the population into core population (c) and non-core population (n), and can directly construct G based on the loss of certain data. -1, there is a core population matrix inversion in the construction process. Reducing the number of core populations will reduce the accuracy of breeding value estimation. In large populations, it is still difficult to invert the core matrix. The existing Apy algorithm and ssGBLUP algorithm cannot avoid the inversion of large matrices, and the accuracy of breeding value estimation is lost. Summary of the invention

[0006] The purpose of the present invention is to provide a breeding selection optimization method, system, device and medium based on the BLUP model, which can significantly reduce the memory load in the BLUP model solving process by adopting an iterative block calculation method, while improving the calculation efficiency and maintaining the accuracy of breeding value estimation.

[0007] To achieve the above object, the present invention provides the following solutions:

[0008] A breeding selection optimization method based on a BLUP model, comprising:

[0009] Construct BLUP model based on animal pedigree information;

[0010] The BLUP model is optimized and solved using a block method to obtain an unbiased optimal prediction value of the animal's genetic potential; the unbiased optimal prediction value is used for breeding selection; the optimization solution process of the BLUP model includes two matrix inversion operations, one is the inversion operation of the kinship matrix, and the other is the inversion operation of a larger matrix containing fixed effects.

[0011] Optionally, the basic equation of the BLUP model is:

[0012]

[0013] Where X is the fixed effect matrix; X' is the transpose of X; Z is the random effect matrix, and Z' is the transpose of Z; y is the observation matrix; G is the kinship matrix, with a size of n*n; n is the number of animal individuals; G -1 is the inverse matrix of G; k represents the heritability of the phenotype, which is a constant;

[0014] The inverse operation of the kinship matrix is ​​the inverse operation of the G matrix. -1 Calculation of

[0015] The inversion of the larger matrix containing the fixed effects is:

[0016]

[0017] in, is the fixed effect value to be solved, is the random effect value to be solved.

[0018] Optionally, before optimizing and solving the BLUP model by using the block method, the method further includes: setting the G matrix in the BLUP model to (n*n), and dividing the G matrix into four blocks:

[0019]

[0020] Among them, G1 represents the first sub-matrix; G2 represents the second sub-matrix; G3 represents the third sub-matrix; G4 represents the fourth sub-matrix.

[0021] Optionally, optimizing and solving the BLUP model by using a block method includes single inversion and iterative multiple inversion.

[0022] Optionally, the single inversion is calculated according to the following formula:

[0023]

[0024] The overall inversion of the G matrix is ​​simplified to the G1 matrix and The computational complexity of inverting the two small matrices of the matrix G is n 3 , the computational complexity of the two small matrices is 2*(n / 2) 3 ;

[0025] The iterative inversion is performed multiple times, and the leftmost mixing matrix of the basic equation is calculated according to the following formula:

[0026]

[0027]

[0028] Among them, A represents the first sub-matrix of the mixing matrix; B represents the second sub-matrix of the mixing matrix; C represents the third sub-matrix of the mixing matrix; and D represents the fourth sub-matrix of the mixing matrix.

[0029] The present invention also provides a breeding selection optimization system based on the BLUP model, comprising:

[0030] Model building module, used to build BLUP models based on animal pedigree information;

[0031] The model solving module is used to optimize and solve the BLUP model using a block method to obtain an unbiased optimal prediction value of the animal's genetic potential; the unbiased optimal prediction value is used for breeding selection; the optimization solving process of the BLUP model includes two matrix inversion operations, one is the inversion operation of the kinship matrix, and the other is the inversion operation of a larger matrix containing fixed effects.

[0032] The present invention also provides an electronic device, comprising a memory and a processor, wherein the memory is used to store a computer program, and the processor runs the computer program to enable the electronic device to execute the breeding selection optimization method based on the BLUP model.

[0033] The present invention also provides a computer-readable storage medium storing a computer program, wherein the computer program, when executed by a processor, implements the breeding selection optimization method based on the BLUP model as described above.

[0034] According to the specific embodiments provided by the present invention, the present invention discloses the following technical effects:

[0035] The present invention discloses a breeding selection optimization method, system, device and medium based on a BLUP model, the method comprising constructing a BLUP model based on animal pedigree information; optimizing and solving the BLUP model using a block method to obtain an unbiased optimal prediction value of the animal's genetic potential; the unbiased optimal prediction value is used for breeding selection; the optimization solution process of the BLUP model includes two matrix inversion operations, one is an inversion operation of a kinship matrix, and the other is an inversion operation of a larger matrix containing fixed effects. The present invention can significantly reduce the memory load in the BLUP model solution process by adopting an iterative block calculation method, while improving the calculation efficiency and maintaining the accuracy of breeding value estimation intact. BRIEF DESCRIPTION OF THE DRAWINGS

[0036] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative labor.

[0037] Figure 1 The figure is a schematic diagram of the process of breeding selection optimization based on the BLUP model of the present invention. DETAILED DESCRIPTION

[0038] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0039] The purpose of the present invention is to provide a breeding selection optimization method, system, device and medium based on the BLUP model, which can significantly reduce the memory load in the BLUP model solving process by adopting an iterative block calculation method, while improving the calculation efficiency and maintaining the accuracy of breeding value estimation.

[0040] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, the present invention is further described in detail below with reference to the accompanying drawings and specific embodiments.

[0041] The BLUP method was first applied in animal breeding. The traditional animal model is based on the kinship matrix (also known as the A matrix) constructed based on pedigree information to solve the mixed model equations (MME), so it is called ABLUP. The ABLUP model uses animal pedigree information to estimate EBV. Pedigree information has traditionally been used to estimate genetic parameters in the field of animal breeding, but it can only provide limited genetic information because it mainly relies on family relationships rather than specific genetic markers.

[0042]

[0043] With the rapid development of biotechnology, especially the advancement of chip sequencing technology, the GBLUP (Genomic BLUP) model came into being. The GBLUP model uses dense genetic marker data obtained from gene chips. Compared with ABLUP, it can estimate genetic potential more accurately because gene chips provide more detailed genetic information.

[0044]

[0045] However, although the information of chip sequencing is dense, the cost of measurement is high, so the number of measurements is limited. In order to obtain genetic estimates for the entire population, the ssGBLUP (Single-Step GBLUP) model combining traditional pedigree information and gene sequencing data was proposed. The ssGBLUP model takes into account both the genotype and pedigree information of the animal, providing a comprehensive genetic assessment.

[0046]

[0047]

[0048] Based on the above content, in order to solve the specific problems described in the background technology, the present invention provides a Figure 1 The breeding selection optimization method based on the BLUP model shown includes:

[0049] Step 100: constructing a BLUP model based on animal pedigree information;

[0050] Step 200: Optimizing and solving the BLUP model using a block method to obtain an unbiased optimal prediction value of the animal's genetic potential; the unbiased optimal prediction value is used for breeding selection; the optimization and solving process of the BLUP model includes two matrix inversion operations, one is the inversion operation of the kinship matrix, and the other is the inversion operation of a larger matrix containing fixed effects.

[0051] There are two matrix inversion operations in the BLUP model. One is the inverse matrix of the kinship matrix. Taking GBLUP as an example, that is, solving G -1 matrix.

[0052]

[0053] The other time is the inverse of a larger matrix containing the fixed effects, that is, the inverse of the rightmost mixing matrix in the equation below.

[0054]

[0055] The technical principles of this implementation scheme are as follows:

[0056] Take the G matrix (n*n) as an example and divide it into four blocks:

[0057]

[0058] (1) Single inversion

[0059] The single inversion is calculated according to the following formula:

[0060]

[0061] This formula simplifies the overall inversion of the G matrix to the G1 matrix and The computational complexity of inverting the two small matrices of the matrix G is n 3 , the computational complexity of the two small matrices is 2*(n / 2) 3 .

[0062] (2) Iterate multiple times to find the inverse

[0063] On a block basis, if the G1 matrix and If the matrix is ​​still too large, it can be solved by further matrix block solving. The second inversion is a composite matrix inversion.

[0064] The inversion of the larger matrix containing the fixed effects is:

[0065]

[0066] That is, the following part of the operation:

[0067]

[0068] The matrix can be considered as a whole matrix, which can be split into ABCD and then inverted in blocks:

[0069]

[0070]

[0071] Based on the above specific process, this embodiment has the following beneficial effects:

[0072] The number of animal individuals is n, the fixed effect one-hot encoding classification quantity is m, and in the direct inversion calculation method in the traditional BLUP model, the computational complexity of the first kinship matrix inversion is n 3 , the computational complexity of the second mixing matrix inversion is (n+m) 3 , perform a matrix partition, the computational complexity is 2*(n / 2) 3 and 2*[(n+m) / 2] 3 ,As the number of block iterations increases, the single inversion computation complexity is lower, thus the time required for the entire calculation process is greatly shortened.,Compared to the direct inversion calculation method in the traditional BLUP model, the block inversion ,technology used in the present invention greatly shortens the calculation time.

[0073] When processing large-scale data sets, this embodiment can significantly improve computing efficiency and provide a more efficient and energy-saving solution for complex genetic evaluation and breeding value estimation. In addition, by reducing computing time, this method also helps to reduce energy consumption, further reducing the demand for high-performance computing resources, and is conducive to achieving more environmentally friendly and cost-effective scientific computing.

[0074] In addition, the present invention also provides a breeding selection optimization system based on the BLUP model, comprising:

[0075] Model building module, used to build BLUP models based on animal pedigree information;

[0076] The model solving module is used to optimize and solve the BLUP model using a block method to obtain an unbiased optimal prediction value of the animal's genetic potential; the unbiased optimal prediction value is used for breeding selection; the optimization solving process of the BLUP model includes two matrix inversion operations, one is the inversion operation of the kinship matrix, and the other is the inversion operation of a larger matrix containing fixed effects.

[0077] The present invention also provides an electronic device, comprising a memory and a processor, wherein the memory is used to store a computer program, and the processor runs the computer program to enable the electronic device to execute the breeding selection optimization method based on the BLUP model.

[0078] The present invention also provides a computer-readable storage medium storing a computer program, wherein the computer program, when executed by a processor, implements the breeding selection optimization method based on the BLUP model as described above.

[0079] The various embodiments in this specification are described in a progressive manner, and each embodiment focuses on the differences from other embodiments. The same or similar parts between the various embodiments can be referenced to each other.

[0080] This article uses specific examples to illustrate the principles and implementation methods of the present invention. The above examples are only used to help understand the core idea of ​​the present invention. At the same time, for those skilled in the art, according to the idea of ​​the present invention, there will be changes in the specific implementation methods and application scope. In summary, the content of this specification should not be understood as limiting the present invention.

Claims

1. A breeding selection optimization method based on BLUP model, characterized in that: include: Construct BLUP model based on animal pedigree information; The BLUP model is optimized and solved by using a block method to obtain an unbiased optimal prediction value of the animal's genetic potential; the unbiased optimal prediction value is used for breeding selection; the optimization solution process of the BLUP model includes two matrix inversion operations, one is an inversion operation of a kinship matrix, and the other is an inversion operation of a larger matrix containing fixed effects; The basic equation of the BLUP model is: Where X is the fixed effect matrix; X' is the transpose of X; Z is the random effect matrix, and Z' is the transpose of Z; y is the observation matrix; G is the kinship matrix, with a size of n*n; n is the number of animal individuals; G -1 is the inverse matrix of G; k represents the heritability of the phenotype, which is a constant; The inverse operation of the kinship matrix is ​​the inverse operation of the G matrix. -1 Calculation of The inversion of the larger matrix containing the fixed effects is: in, is the fixed effect value to be solved, is the random effect value to be solved; Before optimizing and solving the BLUP model by using the block method, the method further includes: setting the G matrix in the BLUP model to (n*n), and dividing the G matrix into four blocks: Among them, G1 represents the first sub-matrix; G2 represents the second sub-matrix; G3 represents the third sub-matrix; G4 represents the fourth sub-matrix; The optimization and solution of the BLUP model by using the block method includes single inversion and iterative multiple inversion; The single inversion is calculated according to the following formula: The overall inversion of the G matrix is ​​simplified to the G1 matrix and The computational complexity of inverting the two small matrices of the matrix G is n 3 , the computational complexity of the two small matrices is 2*(n / 2) 3 ; The iterative inversion is performed multiple times, and the leftmost mixing matrix of the basic equation is calculated according to the following formula: Among them, A represents the first block matrix of the mixing matrix; B represents the second sub-matrix of the mixing matrix; C represents the third sub-matrix of the mixing matrix; and D represents the fourth sub-matrix of the mixing matrix.

2. A breeding selection optimization system based on a BLUP model, applied to the method described in claim 1, characterized in that: include: Model building module, used to build BLUP models based on animal pedigree information; The model solving module is used to optimize and solve the BLUP model using a block method to obtain an unbiased optimal prediction value of the animal's genetic potential; the unbiased optimal prediction value is used for breeding selection; the optimization solving process of the BLUP model includes two matrix inversion operations, one is the inversion operation of the kinship matrix, and the other is the inversion operation of a larger matrix containing fixed effects.

3. An electronic device, characterized in that: The electronic device comprises a memory and a processor, wherein the memory is used to store a computer program, and the processor runs the computer program to enable the electronic device to execute the method according to claim 1.

4. A computer-readable storage medium, characterized in that: The computer program is stored therein, and when the computer program is executed by a processor, the method as claimed in claim 1 is implemented.

Images