Genome selection method based on deep learning reaGP

Abstract

The invention relates to the technical field of genome selection, and particularly discloses a genome selection method based on deep learning reaGP, which comprises the following steps: constructing a genome reference group, and measuring economic characters of the group; the determined economic traits comprise three economic traits of fattening traits, correcting phenotypes of the economic traits of the selected population, and performing quality control on genome data; performing coding combination on the genome data and the frequency to form a 3D data structure for being input into a convolutional neural network; constructing a convolutional neural network with a residual unit and an attention mechanism to stabilize the gradient and enhance the feature extraction capability; the genome breeding value of the population is selected by using the convolutional neural network, and the deep learning model is constructed, so that compared with a linear model, an existing machine learning model and an existing deep learning model, the three characters are improved by 0.5%-18% on the whole; and the complexity prediction accuracy is improved.

Description

Technical Field

[0001] The present invention relates to the technical field of genome selection, and in particular to a genome selection method based on deep learning reaGP. Background Art

[0002] As a nonlinear model, deep learning can get better results by simply fine-tuning parameters, and has achieved good results in image processing tasks. In the early stages of biological breeding, when using linear models to predict phenotypes, there is no need to establish a model formula like a linear model and then solve gebv. When facing nonlinear problems, linear models cannot better fit equations. In addition, linear models cannot better solve the problem of p>n. It is easy to produce overfitting, as well as some non-additive effects (dominance, epistasis, etc.) controlled by complex phenotypes. Deep learning models can better capture these properties, which brings more advantages to animal and plant breeding. For some biological populations, economic traits are complex phenotypes controlled by some non-additive effects. Deep learning methods may bring unlimited possibilities for improving the accuracy of biological breeding value estimation. Summary of the invention

[0003] The purpose of the present invention is to provide a genome selection method based on deep learning reaGP to solve the problems in the above background.

[0004] The purpose of the present invention can be achieved through the following technical solutions:

[0005] A genome selection method based on deep learning reaGP, comprising the following steps:

[0006] Step 1: Construct a reference population and measure the economic traits of the population;

[0007] Step 2: Collect blood from the reference group, extract DNA, perform genotyping using biochips, and process and quality control the genotyping data;

[0008] Step 3: Convert the genome data (0, 1, 2) to 0, 0, 0, 1, 1, 1; calculate the frequency information of the three allele groups; combine the genome data and the frequency information to convert them into three-channel data, which is used to provide richer feature information for the neural network and serve as the input data of the neural network;

[0009] Step 4: Build a convolutional neural network model. The network structure of the convolutional neural network model is two layers of convolution and two fully connected layers. The number of nodes is 64 and 1. The size of each convolution kernel is 7*7 and 5*5 respectively, and the number of channels in the convolution layer is 64.

[0010] Step 5: Use cross-validation to divide the training population into training set and test set. Use the trained model to evaluate the test set data. The predicted value (y) in the test set data is used as the genomic breeding value estimate (GEBV) predicted by deep learning. Use the y value and the true phenotype for Pearson correlation, and use the correlation coefficient as the accuracy evaluation criterion.

[0011] As a further solution of the present invention: the economic traits of the population include: three forms of fattening traits, namely, ketone body weight WWT, weaning weight CWT, and daily weight gain FDG.

[0012] As a further solution of the present invention: in step 2, the quality control is to eliminate unqualified individuals and SNPs;

[0013] The biochip is an Illumina BovineHD high-density chip.

[0014] As a further solution of the present invention: the data format of the neural network input data in step three is: a: 0, 0, P (aa), a: 1, 0, P (Aa), a: 1, 1, P (AA).

[0015] As a further solution of the present invention: in the step 3, calculating the allele frequency information is to calculate the frequency of occurrence of the allele corresponding to each marker in all samples; wherein the allele frequency information is used to provide biological explanation for the three-channel input information of the neural network;

[0016] Specifically include:

[0017] Assume that for a particular genetic marker, genotypes Aa and AA appear D times and H times respectively; if the population size is N, the total number of alleles will be 2N, then the frequency P of allele A is:

[0018] The frequency of allele a is q=1p, and after calculating the allele frequencies of the genetic markers, the genotype frequencies of genotypes AA, Aa and aa of the specific marker are obtained from p2, 2pq and q2.

[0019] As a further solution of the present invention: in step 3, the three-channel data set is used to expand the common one-dimensional genome data into a feature dimension of richness, and the input format of each sample is changed from (1, n) to (3, M, M);

[0020] The calculation formula is as follows: where n is the number of markers for the genomic data and M is the side length for the three-way data, the calculation formula is as follows:

[0021] P L , P RAs the number of symmetrically filled 0s, the calculation formula is as follows:

[0022] P R =P L +1.

[0023] As a further solution of the present invention: in the step 4, a residual module is added between the fully connected layer and the second convolutional layer; the residual module is used to prevent the gradient disappearance or explosion problem during the deep network training process.

[0024] As a further solution of the present invention: the structure of the residual module is two layers of 3*3 convolution kernels, and a 1*1 convolution layer is added to make the number of input channels correspond to the number of output channels, so as to realize information transmission;

[0025] The residual module has two stride=1 convolutional layers, on top of which an identity mapping layer with stride=1 is added to ensure that the input dimension and output dimension are consistent;

[0026] The SE attention mechanism is added between the two convolutional layers and after the residual module; it is used to enhance the recognition ability of important traits associated with specific genotypes.

[0027] As a further solution of the present invention: in the step 5, the predicted result y value is obtained through the network model according to the phenotype of each trait and the three-channel data, wherein y WWT Estimated breeding value for weaning weight genomics, y FDG is the genomic estimated breeding value of daily gain during the fattening period, y CWT Estimated breeding values ​​for carcass weight genomics.

[0028] Beneficial effects of the present invention:

[0029] By building a deep learning model (reaGP), the overall improvement in the three traits was 0.5% to 18% compared to linear models, existing machine learning models, and deep learning models; compared to traditional genomic selection methods: improved the accuracy of complex trait prediction;

[0030] Traditional methods have limitations in dealing with complex characteristics and nonlinear problems, but deep learning models can significantly improve these problems and enhance the accuracy and stability of predictions through the advantages of multi-layer neural networks.

[0031] By fine-tuning the model parameters, the prediction performance can be further improved to ensure better results under different traits and conditions.

[0032] This method not only utilizes genomic and phenotypic data, but also combines other biological information as auxiliary variables, which significantly improves the estimation accuracy of genomic selection breeding values ​​and breaks through the limitations of relying solely on genomic and phenotypic data.

[0033] Compared with traditional breeding methods, deep learning models can screen and predict excellent individuals more quickly, shorten the generation interval, improve breeding efficiency, and ultimately accelerate genetic progress. BRIEF DESCRIPTION OF THE DRAWINGS

[0034] The present invention will be further described below in conjunction with the accompanying drawings.

[0035] Figure 1 It is a schematic flow chart of a genome selection method of deep learning reaGP of the present invention;

[0036] Figure 2 It is a schematic diagram of the three-channel form of reaGP data in the present invention;

[0037] Figure 3 is a schematic diagram of the model structure of reaGP in the present invention;

[0038] Figure 4 is a schematic diagram of the residual structure of reaGP in the present invention;

[0039] Figure 5 It is a comparative schematic diagram of the genome prediction accuracy of three traits of Chinese Western cattle of the present invention. DETAILED DESCRIPTION

[0040] 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.

[0041] Embodiment 1

[0042] A genome selection method based on deep learning reaGP, comprising the following steps:

[0043] Step 1: Construct a genomic reference population and measure the economic traits of the population; the measured economic traits include three economic traits of fattening traits, specifically: daily weight gain FDG, carcass weight CWT, and weaning weight WWT;

[0044] Step 2: Correct the phenotypes of economic traits of the selected population and perform quality control on the genomic data;

[0045] Step 3: Combine the genomic data with the frequency encoding to form a 3D data structure for input into the convolutional neural network;

[0046] Step 4: Build a convolutional neural network with residual units and attention mechanism to stabilize the gradient and enhance feature extraction capabilities;

[0047] Step 5: Use the above convolutional neural network to select the genomic breeding value of the population.

[0048] Embodiment 2

[0049] See also Figure 1 - Figure 4 As shown, the present invention is a genome selection method based on deep learning reaGP, comprising the following steps:

[0050] Step 1: Construct a reference population and measure the economic traits of the population, including three fattening traits, namely, keto body weight WWT, weaning weight CWT, and daily weight gain FDG;

[0051] Step 2: Collect blood from the reference group, extract DNA, perform genotyping using biochips, and process and quality control the genotyping data;

[0052] Among them, quality control involves the elimination of unqualified individuals and SNPs, and the biochip is the Illumina BovineHD high-density chip;

[0053] Step 3: Convert the genomic data (0, 1, 2) to 0, 0, 0, 1, 1, 1; calculate the frequency information of the three allele groups; combine the genomic data and frequency information to convert them into three-channel data, which is used to provide richer feature information for the neural network and serve as the input of the neural network; the data format is: a: 0, 0, P(aa), a: 1, 0, P(Aa), a: 1, 1, P(AA);

[0054] Calculating the allele frequency information is to calculate the frequency of the corresponding allele in each marker in all samples; the allele frequency information is used to provide biological explanation for the three-channel input information of the neural network; the calculation formula is as follows:

[0055] We apply Hardy-Weinberg equilibrium and assume that for a particular genetic marker, genotypes AA and Aa appear D and H times respectively; if the population size is N, the total number of alleles will be 2N, then the frequency P of allele A is:

[0056] The frequency of allele a is q = 1-p. After calculating the allele frequencies of genetic markers, the genotype frequencies of genotypes AA, Aa, and aa of a specific marker are obtained from p2, 2pq, and q2;

[0057] The three-channel data set is used to expand the ordinary one-dimensional genomic data into the feature dimension of richness. The input format of each sample is changed from (1, N) to (3, M, M);

[0058] The calculation formula is as follows: where N is the number of markers for the genomic data and M is the side length for the three-way data, the calculation formula is as follows: That is, assuming that the dimension N of the genomic data can be directly taken as a complete square root, then M is used as the side length of the three-channel data; otherwise, 0 is used to perform symmetrical padding so that the dimension of the genomic data becomes N+P L +P R , satisfying the condition of taking a complete square root, then M is used as the side length of the three-channel data.

[0059] P L , P R As the number of symmetrically filled 0s, the calculation formula is as follows:

[0060] P R =P L +1, where P L As the first position of the one-dimensional genome data N is filled with 0, P R The terminal position of the one-dimensional genome data N is filled with 0.

[0061] Step 4: Build a convolutional neural network model. The network structure of the convolutional neural network model consists of two convolutional layers with stride = 2 and two fully connected layers, where the nodes are 64 and 1, the convolution kernel size of each layer is 7*7 and 5*5 respectively, and the number of convolutional layer channels is 64; among them, a BN layer, a Droput layer, and a ReLU activation function layer are added after each convolutional layer.

[0062] To alleviate the gradient explosion or gradient vanishing problem, a residual module is added between the fully connected layer and the second convolutional layer; the residual module is used to prevent the gradient vanishing or exploding problem during the deep network training process;

[0063] Among them, the structure of the residual module is two layers of 3*3 convolution kernels, in which a BN layer, a Droput layer, and a ReLU activation function layer are added after each convolution layer.

[0064] And add a 1*1 convolution layer to make the number of input channels correspond to the number of output channels. After the convolution layer, the BN layer, the Droput layer, and the ReLU activation function layer are added to realize the transmission of information.

[0065] In order to enhance the feature information mining capability, the SE attention mechanism is added between the two convolutional layers and after the residual module; the attention mechanism is used to enhance the recognition of important traits associated with a specific genotype;

[0066] The residual module has two stride=1 convolutional layers, on top of which an identity mapping layer with stride=1 is added to ensure that the input dimension and output dimension are consistent;

[0067] The specific formula of the residual is:

[0068] H(X)=F(X)+0;

[0069] The output of the residual network H(X) consists of two parts, including the two residual layers F(X) built in step 4 and the identity mapping 0. When back propagating and performing gradient derivation, This may be due to the activation functions such as sigmoid or Tanh in the network layer. As the number of network layers increases, the gradient of the entire network is continuously derived. This causes the gradient to disappear, and prevents the gradient from disappearing by introducing the identity mapping x.

[0070] The network gradient at this time is expressed as

[0071]

[0072] Where X in F(X) is the output of the convolutional neural network when the residual layer is added, θ is the continuously updated weight parameter, and α is a fixed learning rate.

[0073] Step 5: Use cross-validation to divide the training population into a training set and a test set, and use the trained model to evaluate the test set data. The specific formula is y = f(x), where f(*) is the optimal weight for model training, x is the input data, and the predicted value of y is accumulated by x and the internal weight. The predicted value (y) in the test set data is used as the genomic breeding value estimate (GEBV) predicted by deep learning; use the y value and the true phenotype for Pearson correlation, and the correlation coefficient is used as the accuracy evaluation standard;

[0074] The predicted value is equivalent to the genomic breeding value estimate (GEBV) obtained based on the solution of the linear model;

[0075] The predicted result y value is obtained through the network model based on the phenotype and three-channel data of each trait, where y WWT Estimated breeding value for weaning weight genomics, y FDG is the genomic estimated breeding value of daily gain during the fattening period, y CWT Estimated breeding values ​​for carcass weight genomics.

[0076] By building a deep learning model (reaGP), the overall improvement in the three traits was 0.5% to 18% compared to linear models, existing machine learning models, and deep learning models; compared to traditional genomic selection methods: improved the accuracy of complex trait prediction;

[0077] Traditional methods have limitations in dealing with complex characteristics and nonlinear problems, but deep learning models can significantly improve these problems and enhance the accuracy and stability of predictions through the advantages of multi-layer neural networks.

[0078] By fine-tuning the model parameters, the prediction performance can be further improved to ensure better results under different traits and conditions.

[0079] This method not only utilizes genomic and phenotypic data, but also combines other biological information as auxiliary variables, which significantly improves the estimation accuracy of genomic selection breeding values ​​and breaks through the limitations of relying solely on genomic and phenotypic data.

[0080] Compared with traditional breeding methods, deep learning models can screen and predict excellent individuals more quickly, shorten the generation interval, improve breeding efficiency, and ultimately accelerate genetic progress.

[0081] Embodiment 3

[0082] Reference Figure 5 As shown, based on the above technical solution, this embodiment uses the Huaxi cattle population to perform genome prediction of three traits, specifically:

[0083] Step 1: Determine the traits of Huaxi cattle, including three fattening traits, namely daily weight gain FDG, carcass weight CWT, and weaning weight WWT;

[0084] Huaxi cattle are weaned at 6 months of age. After weaning, they are fattened under consistent feeding and management conditions, and the growth traits of the animals are measured every six or twelve months until slaughter; live weight is recorded after a 24-hour fasting period; Huaxi cattle are slaughtered at 18 to 24 months of age;

[0085] Step 2: Blood was collected from each cow in the reference group, DNA was extracted, and genotyping was performed using the Illumina BovineHD (770k) high-density chip. The data after genotyping were processed and quality controlled using plinkV1.90 software. The genomic data after quality control was 550K, and the numbers of the three types were 1299, 1362, and 1303 respectively;

[0086] Step 3: Convert the genomic data and frequency information into a three-channel data format; specifically, first calculate the frequencies of AA, Aa, and aa alleles in 550K. Secondly, encode the three alleles: AA: 1, 1, P(AA), Aa: 1, 0, P(Aa), aa: 0, 0, P(aa). Next, determine whether 550K can be square rooted, 574547≈757. (It cannot be completely square rooted), at this time, the length of the one-dimensional genome data is (757+1)*(757+1)=574564, and the three-channel side length M=757. The dimension is 17 more than before, and at this time, the original one-dimensional genome is symmetrically padded with 0s. The left end of the genome is padded with 0s P L =17 / 2 rounded down to 8; right zero padding P at the end of the genome R =17-8=9; Finally, we get three-channel data of shape (3, M, M) as the input X of the neural network. We build a neural network based on residual and channel attention mechanism. And we use the three-channel data and phenotype y as the input of the deep learning model for model training;

[0087] Step 4: Use ten cross-validations to divide the training data set into a training set and a test set, and use the trained model to predict the test set data; calculate the predicted value GEBV and the actual phenotype y of the test set * The Pearson correlation coefficient is:

[0088]

[0089] Step 5: Repeat step 4 ten times and take the average of the results;

[0090] Step 6: Use the genomic data and phenotypic data as input data for the other five models, namely DNNGP, SVR, RKHS, BayesB, and GBLUP; repeat steps 4 and 5 to calculate the Pearson correlation coefficients of the five methods respectively.

[0091] Figure 4 This is a comparison of the Pearson correlation coefficients of the six methods under ten cross-validations. From the box plot, we can see that reaGP has the highest mean value among the three traits and a lower standard deviation, and the model is more stable.

[0092] An efficient and high-quality Huaxi cattle breeding method based on deep learning genomic selection technology has improved prediction accuracy compared to other linear models. It provides a new path for Huaxi cattle in genomic selection prediction accuracy;

[0093] Through the application of this embodiment, the breeding efficiency and accuracy of Huaxi cattle can be significantly improved, and the improvement and optimization of Huaxi cattle population can be promoted, which has broad application prospects and economic value. This method can quickly estimate the genome breeding value of Huaxi cattle in my country, which has great application value and promotion prospects for beef cattle breeding in my country.

[0094] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modification, equivalent substitution, improvement, etc. made within the spirit and principle of the present invention shall be included in the protection scope of the present invention. Any reference numeral in a claim shall not be regarded as limiting the claim involved.

[0095] In addition, it should be understood that although the present specification is described according to implementation modes, not every implementation mode contains only one independent technical solution. This description of the specification is only for the sake of clarity. Those skilled in the art should regard the specification as a whole. The technical solutions in each embodiment may also be appropriately combined to form other implementation modes that can be understood by those skilled in the art.

[0096] The above is a detailed description of an embodiment of the present invention, but the content is only a preferred embodiment of the present invention and cannot be considered to limit the scope of implementation of the present invention. All equivalent changes and improvements made within the scope of the present invention should still fall within the scope of the patent coverage of the present invention.

Claims

1. A genome selection method based on deep learning reaGP, characterized in that: The following steps are involved: Step 1: Construct a reference population and measure the economic traits of the population; Step 2: Collect blood from the reference group, extract DNA, perform genotyping using biochips, and process and quality control the genotyping data; Step 3: Convert the genome data (0, 1, 2) to 0, 0, 0, 1, 1, 1; calculate the frequency information of the three allele groups; combine the genome data and the frequency information to convert them into three-channel data, which is used to provide richer feature information for the neural network and serve as the input data of the neural network; Step 4: Build a convolutional neural network model. The network structure of the convolutional neural network model is two layers of convolution and two fully connected layers. The number of nodes is 64 and 1. The size of each convolution kernel is 7*7 and 5*5 respectively, and the number of channels in the convolution layer is 64. Step 5: Use cross-validation to divide the training population into training set and test set. Use the trained model to evaluate the test set data. The predicted value (y) in the test set data is used as the genomic breeding value estimate (GEBV) predicted by deep learning. Use the y value and the true phenotype for Pearson correlation, and use the correlation coefficient as the accuracy evaluation criterion.

2. A genome selection method based on deep learning reaGP according to claim 1, characterized in that: The economic traits of the group include: three types of fattening traits, namely, keto body weight WWT, weaning weight CWT, and daily weight gain FDG.

3. A genome selection method based on deep learning reaGP according to claim 1, characterized in that: In step 2, quality control is to eliminate unqualified individuals and SNPs; The biochip is an Illumina BovineHD high-density chip.

4. A genome selection method based on deep learning reaGP according to claim 1, characterized in that: The data format of the neural network input data in step three is: a: 0, 0, P (aa), a: 1, 0, P (Aa), a: 1, 1, P (AA).

5. The method for genome selection based on deep learning reaGP according to claim 1, characterized in that: In the step 3, calculating the allele frequency information is to calculate the frequency of the allele corresponding to each marker in all samples; wherein the allele frequency information is used to provide biological explanation for the three-channel input information of the neural network; Specifically include: Assume that for a particular genetic marker, genotypes Aa and AA appear D times and H times respectively; if the population size is N, the total number of alleles will be 2N, then the frequency P of allele A is: The frequency of allele a is q=1p, and after calculating the allele frequencies of the genetic markers, the genotype frequencies of genotypes AA, Aa and aa of the specific marker are obtained from p2, 2pq and q2.

6. A genome selection method based on deep learning reaGP according to claim 1, characterized in that: In step 3, the three-channel data set is used to expand the common one-dimensional genome data into the feature dimension of richness, and the input format of each sample is changed from (1, n) to (3, M, M); The calculation formula is as follows: where n is the number of markers in the genomic data and M is the side length of the three-way data, the calculation formula is as follows: P L , P R As the number of symmetrically filled 0s, the calculation formula is as follows: P R =P L +1。 7. A genome selection method based on deep learning reaGP according to claim 1, characterized in that: In the step 4, a residual module is added between the fully connected layer and the second convolutional layer; the residual module is used to prevent the gradient from disappearing or exploding during the deep network training process.

8. A genome selection method based on deep learning reaGP according to claim 7, characterized in that: The structure of the residual module is two layers of 3*3 convolution kernels, and a 1*1 convolution layer is added to make the number of input channels correspond to the number of output channels to achieve information transmission; The residual module has two stride=1 convolutional layers, on top of which an identity mapping layer with stride=1 is added to ensure that the input dimension and output dimension are consistent; The SE attention mechanism is added between the two convolutional layers and after the residual module; Used to enhance the identification of important traits associated with specific genotypes.

9. The method for genome selection based on deep learning reaGP according to claim 1, characterized in that: In step 5, the predicted result y value is obtained through the network model based on the phenotype of each trait and the three-channel data, where y WWT Estimated breeding value for weaning weight genomics, y FDG is the genomic estimated breeding value of daily gain during the fattening period, y CWT Estimated breeding values ​​for carcass weight genomics.

Images