Genetic evaluation and association study of candidate gene using SNP and pedigree data in domestic animal

Abstract

This doctoral dissertation consists of four studies related to genetic evaluation and association study using SNP and pedigree data in livestock. Genetic evaluation of economic traits is essential to calculate estimated breeding value (EBV), heritability, genetic correlation and reliability of breeding value. And genomic selection (or prediction) is applied to calculate genomic estimated breeding value (GEBV) using SNP data. I also examined genetic association related to major economic traits of swine using SNP marker driven by NGS analysis. The chapter 1 introduced the basic background and necessity of the series of works in this doctoral dissertation. The genetic evaluation is important to assess the genetic ability of individual animal and heritability of specific traits in animal populations. In chapter 2, I proposed genetic evaluation to analyze correlated phenotypes applying principle component analysis (PCA). In this study, principal components analysis (PCA) was adapted to generate a new index as a measure of racing performance of horses. This method allows us to reduce the number of variables considered in the evaluation of the horses' racing performance, which may facilitate modeling genetic programs. The resulted racing time, earning prize and rank were selected for generating new indices as the representation of racing performance of the horses. Three indices used in this study were: 1) PCA1 generated from the modified values of racing time, earning prize and rank, 2) PCA2 generated from the modified racing time and rank, and 3) the adjusted racing time. Heritability and repeatability were 0.324 (± 0.026) and 0.334 (± 0.034) for adjusted racing time, 0.319 (± 0.014) and 0.326 (± 0.018) for PCA1, and 0.324 (± 0.010) and 0.332 (± 0.012) for PCA2 respectively. Estimated heritability and repeatability for three indices showed similar values for domestic racing records. However, models using PCA showed better fitting for data than model using racing time as a performance index. The proposed methodology is efficient to evaluate the total variance in this group of correlated traits, allowing reduction in the number of variables for genetic evaluation and construction of better fitting model. After studying traditional genetic evaluation, I investigated genetic variant related to economic traits of domesticated animal. In chapter 3, my study was intended to find new SNP marker for growth traits in pig population in Korea. Using newly-identified 192 SNPs from NGS analysis, I conducted candidate association study to search for genetic variants associated with the average daily gain (ADG), backfat thickness (BF) and days to 90kg (T90). I performed an association analysis using the variance component model based on linear mixed model. Interestingly, these limited number of SNPs explained 0.69, 0.26 and 0.06 of phenotypic variance for ADG, BF and T90, respectively. With Hardy-Weinberg threshold, I also found 20 significant SNPs related to 6 genes with high LD. Without Hardy-Weinberg threshold, two SNPs commonly associated with ADG and BF. To my knowledge, these SNPs are newly reported for the genetic association with growth traits of pigs. It complements a recent pig association study of growth traits that identified other SNPs as the most significant variants. Thoroughbred, a relatively recent horse breed, is best known for its use in horse racing. Although myostatin (MSTN) variants have been reported to be highly associated with horse racing performance, the trait is more likely to be polygenic in nature. In chapter 4, I applied genomic selection to simulated data mimicking current Thoroughbred situation in Korea. Genomic prediction has become the promising paradigm in animal breeding programs. Present study was compared the accuracies of genomic prediction in terms of selection strategy and candidate phenotypes using ridge regression BLUP (rrBLUP). And I also tried to evaluate the efficiency of EBV as response variable in genomic selection. Simulated data mimicking horse genome was used for genomic prediction. The accuracy of traditional EBV using BLUP showed a little difference between phenotypic and EBV selection strategy. The accuracy of genomic prediction for 480 individuals was calculated by 5-fold cross validation. In terms of selection strategy, accuracies of EBV-selection were 0.56 and 0.73 for phenotype and EBV as candidate phenotype. On the other hand, accuracies of phenotypic selection were 0.55 and 0.68 for phenotype and EBV. As reported in this research, genomic prediction showed higher (0.13-0.17) accuracy than traditional BLUP method. In this study, EBV selection strategy using EBV as response variable showed highest accuracy of prediction in genomic selection. Finally, chapter 5 compared prediction accuracy of single step BLUP (ssBLUP) method for different ratio of training set (60, 80 and 90%). Predictions performed using real SNP data with pedigree information for racing related traits which are racing time and total earnings of Thoroughbred. Even, two phenotypes showed highly negative correlation (r2 = -0.85), interestingly, predictions were most accurate in 60% training for racing time EBV (0.30 ± 0.069) and in 90% training for total earning EBV (0.58 ± 0.120). Finally, GWAS using ssBLUP detected SNPs related to racing time and total earnings. This study provides useful information for assessing the accuracy of genomic selection on horse traits using SNP data at the population level.