Functional locus combinations for predicting lifetime farrowing traits in sows and their applications

Abstract

This invention discloses a functional locus combination for predicting sow lifetime litter size and its application. The functional locus combination comprises 15,138 SNP loci, determined based on pig genome version 11.1. Based on multi-breed sow genotype filling data, pedigree and phenotypic data, this invention combines GWAS and machine learning feature extraction models to screen SNP loci that significantly contribute to sow lifetime litter size, ultimately obtaining SNP loci with outstanding contributions to sow lifetime litter size. This provides a new method for predicting sow lifetime litter size and improves the accuracy of breeding for these traits. This invention can improve the accuracy of genomic breeding for sow lifetime litter size, accelerate the genetic progress of genomic selection for sow lifetime litter size, and has significant economic benefits, making it highly valuable in pig molecular breeding.

Description

Technical Field

[0001] This invention relates to the fields of pig molecular marker screening technology and genome breeding, specifically to a combination of functional loci for predicting lifetime litter size traits in sows and its application. Background Technology

[0002] As a typical complex trait, lifetime litter size in pigs is regulated by a multi-gene network and environmental factors. The accurate identification of key SNP loci is a core challenge in modern molecular breeding. Screening for functional SNPs associated with lifetime litter size can not only help develop cost-effective, low-density breeding chips to reduce costs, but also significantly improve the accuracy of genome prediction, thereby driving the transition of breeding strategies from marker-assisted selection to genome selection.

[0003] However, with the widespread adoption of gene sequencing technology, the massive amounts of high-density genotyping data have placed enormous pressure on computation and storage. In studies of lifetime litter size, the genetic contribution of many SNP loci is negligible, instead introducing noise and increasing the analytical burden. Therefore, there is an urgent need to utilize advanced bioinformatics methods and predictive models to efficiently reduce the dimensionality of whole-genome data and screen for the core SNP set associated with lifetime litter size, in order to significantly improve breeding efficiency and reduce costs while ensuring predictive efficacy.

[0004] Genome-wide association studies (GWAS), by systematically comparing genetic variations among individuals with different phenotypes, have become an effective tool for identifying trait-related loci (Hirschhorn & Daly, 2005). However, traditional GWAS often relies on linear models, making it difficult to analyze the complex interactions and nonlinear genetic effects among genetic variations. Against this backdrop, machine learning methods have demonstrated unique advantages, enabling them to extract biologically significant key SNPs from massive genetic data through efficient feature extraction and pattern recognition, providing a new path to improve the accuracy and efficiency of breeding selection.

[0005] Therefore, it is essential to use machine learning feature extraction models combined with genome-wide association analysis to identify functional loci associated with sows' lifetime farrowing traits. Summary of the Invention

[0006] The purpose of this invention is to overcome the shortcomings of the prior art and provide a functional locus combination for predicting the lifetime littering traits of sows and its application. This invention combines the quantification of the genetic contribution of SNPs to the lifetime littering traits of pigs with the results of genome-wide association analysis of lifetime littering traits to screen out functional loci that are meaningful for lifetime littering traits.

[0007] This invention collects ear tissue from pigs at a pig farm in southern China. Using gene chip sequencing and whole-genome resequencing data, a large number of SNP markers are obtained through genotype imputation technology. Combined with genome-wide association analysis and machine learning feature extraction models, functional markers related to sows' lifetime litter size are screened. Noise sites in the imputation data are removed, ultimately forming a 15K gene locus set applicable to sows' lifetime litter size. This invention identifies SNP molecular markers related to sows' lifetime litter size, providing a new low-density molecular marker set for predicting and genetically improving pig-related lifetime litter size traits, which is of great significance for improving pig breeding and reproductive efficiency.

[0008] To achieve the above objectives, the technical solution designed by the present invention is as follows: This invention provides a combination of functional loci for sows’ lifetime farrowing traits, which includes 15,138 SNP loci, and their locus information is shown in Table 2.

[0009] The above-mentioned functional site combination includes the top 10,000 sites selected by genome-wide association analysis in ascending order of P-value, and 7,210 sites with an Importance Score greater than 0 selected by a machine learning feature extraction model. After merging and deduplicating the two parts, a total of 15,138 SNP sites are obtained. The functional site combination is determined based on porcine genome version 11.1.

[0010] The present invention also provides an application of the above-mentioned functional site combination in the screening of superior lifetime littering traits, the improvement of lifetime littering traits in pigs, and the improvement of breeding and reproductive efficiency in pigs.

[0011] The present invention also provides a probe assembly for detecting the above-mentioned combination of functional sites, the probe assembly comprising probes for detecting the above-mentioned 15,138 SNP sites.

[0012] The present invention also provides a 15K liquid phase chip for predicting and screening pig lifetime litter traits, the 15K liquid phase chip comprising the above-described probe combination.

[0013] The present invention also provides a system for detecting the above-mentioned functional site combinations. The detection method of the system is to detect the above-mentioned 15,138 SNP sites in the genome of the variety to be tested by sequencing.

[0014] The present invention also provides an application of the above-described 15K liquid phase chip or the above-described system in the screening of superior lifetime breeding traits.

[0015] The present invention also provides an application of the above-mentioned 15K liquid phase chip or the above-mentioned system in improving the lifetime farrowing traits of pigs and improving the breeding and reproductive efficiency of pigs.

[0016] The beneficial effects of this invention are: 1. This invention uses multi-breed sow genotype filling data, pedigree and phenotypic data, combined with GWAS and machine learning feature extraction models, to screen SNP loci that make a significant contribution to the lifetime littering traits of sows. Finally, it obtains SNP loci that make an outstanding contribution to the lifetime littering traits of sows, providing a new method for predicting the lifetime littering traits of sows and improving the breeding accuracy of lifetime littering traits.

[0017] 2. The SNP markers screened in this invention can be applied to genome-wide association analysis and genome selection related to lifelong littering traits in pigs, providing new molecular marker resources for marker-assisted selection of lifelong littering traits in pigs.

[0018] 3. This invention can improve the accuracy of genomic breeding of sows' lifetime littering traits, accelerate the genetic progress of genomic selection of sows' lifetime littering traits, and has significant economic benefits. It is of great value in pig molecular breeding. Attached Figure Description

[0019] Figure 1 This is a schematic diagram of the screening process for functional loci of sow lifetime farrowing traits according to the present invention. Detailed Implementation

[0020] The present invention will now be described in further detail with reference to specific embodiments, so that those skilled in the art can understand it.

[0021] Example 1 like Figure 1 The screening method for functional loci of lifetime farrowing traits in sows, as shown, includes the following steps: Step 1: Ear tissue samples were collected from sows and sequenced using an 80K microarray. Individuals with extreme phenotypes were selected for whole-genome resequencing to obtain a series of sequencing data. Using the whole-genome resequencing data as a reference panel, the microarray data was filled to the whole-genome level. Quality control was performed on the filling results, retaining the original genome sequence (DR). 2 SNP sites with a maf value > 0.8 and a maf value > 0.01.

[0022] Step 2: Record phenotypic data such as total number of piglets born and number of live piglets for all sequencing sows, and remove individuals with a total number of piglets born less than 3 and a birth weight less than 0.3 kg.

[0023] Step 3: The litter size trait is a repeating trait. An A matrix needs to be constructed based on pedigree information. The EBV is calculated using HIBLUP software and used as the phenotypic criterion for subsequent analysis. HIBLUP identifies single-trait models using the `--single-trait` option when fitting repeating trait models. Note that the repeating phenotypic file contains phenotypic values ​​from multiple measurements of the trait, as well as the recorded environmental factors, and includes column names. The main parameters are as follows: "--pheno-pos": This specifies the position of the trait to be analyzed in the column of the phenotype file; "--rand": When fitting a repeated record model, you need to specify the ID in the first column of the phenotypic file as the random environmental effect to decompose the variance of the permanent environmental effect.

[0024] Step 4: Genome-wide association analysis (GWAS) was performed using the genome filling data and corrected EBV values ​​to screen loci significantly associated with lifetime litter size traits. In this step, the GWAS employed a mixed linear model using rMVP software. Fixed effects included farm, calving season, and breed, with the first three principal components used as covariates. Results were sorted by p-value from low to high, and the top 10K loci were selected.

[0025] Step 5: Based on the machine learning feature extraction model, dimensionality reduction is performed on the genome filling data to screen for loci significantly associated with lifetime reproductive traits. First, 80% of the data is used as the training set and 20% as the validation set. The prediction accuracy on the validation set is used as the evaluation criterion for model performance. Then, a grid search is used to select the optimal parameter combination for each model. The optimal parameters for the models are as follows: Model 1 parameters are "max_depth=3; eta=0.01; gamma=0.0001; min_child_weight=45; subsample=0.8; colsample=0.7; n_estimators=200", and other parameters are default.

[0026] Model 2 parameters are "hidden_layer_size=(100,50); activation='relu'; solver='adam'; alpha=0.0001; batch_size=32", and other parameters are default.

[0027] Model 3 parameters are "alpha = 0.01; L1_ratio = 0.1", and other parameters are default.

[0028] Finally, the accuracy of the estimated breeding values ​​of the markers selected by each model is calculated using Hiblup software, and the optimal SNP set is selected.

[0029] Step 6: Integrate the GWAS and machine learning feature extraction results to obtain the final 15K set of functional loci for sow lifetime farrowing traits.

[0030] The reference genome and genome-wide association analysis results in this invention are based on pig genome version 11.1 (https: / / ftp.ensembl.org / pub / release-105 / fasta / sus_scrofa).

[0031] Based on the above screening method, ear tissue samples were collected from different breeds of sows to screen for lifetime farrowing trait loci, as detailed below: Approximately 9,000 sows were collected (80% of the total population, 7,200 sows, were selected for sequencing, and another 20% (1,800 sows) were selected as a validation population). Ear tissue samples from the 80% sows were then subjected to 80K liquid-phase microarray capture sequencing. Ear tissue samples from 140 Large White, 140 Landrace, and 40 Duroc pigs were selected for 30x resequencing. Using the resequencing data as a reference panel, Beagle 5.4 software (http: / / faculty.washington.edu / browning / beagle / beagle.html) was used to fill the microarray sequencing data to the whole genome level by breed. The Beagle software outputs the squared correlation between the filled genotype and the true genotype (DR) in the filling results. 2 This reflects the accuracy of the fill result. Therefore, the DR in the fill result is retained. 2 SNPs with a value greater than 0.8 were used for subsequent analysis. The plink software was used to further filter the imputed data, removing sites with a maf value less than 0.05.

[0032] The litter size trait is a repeatable trait, which requires constructing an A matrix based on pedigree information, calculating EBV using HIBLUP software, and using it as a phenotype for subsequent analysis. HIBLUP identifies single-trait models by using the --single-trait when fitting repeatable models. It should be noted that the repeatable phenotype file contains phenotypic values ​​of the trait measured multiple times, as well as the recorded environmental factors, and includes column names.

[0033] Genome-wide association analysis (GWAS) was performed on the traits of total litter size and live litter size. A mixed linear model using rMVP was employed, with farm, calving season, and breed as fixed effects, and the first three principal components as covariates. The top 10,000 loci were selected as candidate loci. Using Python, the accuracy of breeding value estimations calculated from loci selected by different models was compared, and 7,210 loci selected by Model 3 were ultimately chosen as candidate loci. The two results were combined, resulting in 15,138 SNP loci as functional locus combinations. The breeding accuracy calculations for each model are shown in Table 1. The specific genomic physical location information of the functional locus combinations is shown in Table 2.

[0034] Table 1. Number of screening sites and accuracy of estimated breeding values ​​for each model Table 2. Locatory location information of functional locus combinations Note: The number before "-" is the chromosome number, and the number after "-" is the locus information.

[0035] Example 2: 15K liquid phase chip for prediction and screening of pig lifetime litter size traits Based on the SNP locus information obtained in Example 1, a 15K liquid phase chip was developed for predicting and screening pig lifetime litter size traits, as detailed below: The 15K liquid phase chip includes probes that detect the 15,138 SNP sites mentioned above.

[0036] Example 3: Application of functional loci of sow lifetime farrowing traits screened by the above-mentioned 15K liquid-phase chip or sequencing system in genome breeding. To ensure the validity of the functional site set obtained by the above 15K liquid-phase chip or sequencing system, two commercial SNP chip sites were used as controls.

[0037] The detection method of the sequencing system in this embodiment uses the Illumina sequencing platform and Q-T7B sequencing instrument to detect the 15,138 SNP sites mentioned above in the genome of the sow breed to be tested.

[0038] Twenty percent of the population from Example 1 was selected as the validation population, comprising 1800 individuals. The GBLUP method was used to estimate the lifetime total litter size breeding value of the validation population, and the Pearson correlation coefficient between the estimated breeding value and the true phenotype was calculated using five-fold cross-validation to determine the accuracy of the estimated breeding value.

[0039] The results are shown in Table 3. The accuracy of the estimated breeding value for lifetime total litter size based on the 15K locus was 0.938. This result indicates that the predicted sow litter size performance is highly correlated with its EBV and can accurately reflect the lifetime litter size performance of the population. Therefore, selecting sows with high lifetime litter size performance as candidate individuals can effectively improve the sow litter size performance of the population and accelerate genetic progress. Compared with the Illumina Porcine SNP 60K chip and the Neocate 50K chip, this method shows significant improvement, demonstrating the effectiveness of the functional loci screened by the above-mentioned 15K liquid phase chip or sequencing system. This allows it to be used for screening for superior lifetime litter size traits, improving lifetime litter size traits in pigs, and enhancing pig breeding and reproductive efficiency. It accelerates the genetic progress of genomic selection for lifetime litter size traits in sows, has significant economic benefits, and is of great value in pig molecular breeding.

[0040] Table 3. Comparison of the accuracy of SNP sets from different sources in estimating breeding values ​​for total litter size. All other parts not described in detail are existing technologies. Although the above embodiments have provided a detailed description of the present invention, they are only some embodiments of the present invention, not all embodiments. People can obtain other embodiments based on these embodiments without creative effort, and these embodiments all fall within the protection scope of the present invention.

Claims

1. A combination of functional loci for a sow's lifetime farrowing trait, characterized in that: The functional locus combination contains 15,138 SNP loci, which were determined based on porcine genome version 11.1; their locus information is as follows: Note: The number before "-" is the chromosome number, and the number after "-" is the locus information.

2. The application of the functional site combination of claim 1 in the screening of superior pig lifetime littering traits, the improvement of pig lifetime littering traits, and the enhancement of pig breeding and reproductive efficiency.

3. A probe assembly for detecting the functional site combination of claim 1, characterized in that: The probe assembly contains probes for detecting the 15,138 SNP sites as described in claim 1.

4. A 15K liquid phase chip for predicting and screening pig lifetime litter traits, characterized in that: The 15K liquid phase chip includes the probe assembly as described in claim 5.

5. A system for detecting the combination of functional sites as described in claim 1, characterized in that: The detection method of the system is to detect the 15,138 SNP sites described in claim 1 in the genome of the variety to be tested by sequencing.

6. The application of the 15K liquid phase chip of claim 4 or the system of claim 5 in the screening of superior lifetime littering traits in pigs.

7. The application of the 15K liquid phase chip of claim 4 or the system of claim 5 in improving the lifetime farrowing traits of pigs and enhancing the breeding and reproductive efficiency of pigs.

Images