Abstract
Simple SummaryWe estimated genetic parameters and conducted genomic prediction for five types of sperm morphology abnormalities in a large Duroc boar population. Results reflect that the studied traits showed moderate to low heritability and moderate to high repeatability. Both genomic best linear unbiased prediction (GBLUP) and single-step GBLUP (ssGBLUP) showed comparative predictive abilities, and ssGBLUP outperformed GBLUP under many circumstances. To our knowledge, this is the first time that the genetic parameters and genomic predictive ability of sperm morphology abnormalities were reported in such a large Duroc boar population.Artificial insemination (AI) has been used globally as a routine technology in the swine production industry. However, genetic parameters and genomic prediction accuracy of semen traits have seldom been reported. In this study, we estimated genetic parameters and conducted genomic prediction for five types of sperm morphology abnormalities in a large Duroc boar population. The estimated heritability of the studied traits ranged from 0.029 to 0.295. In the random cross-validation scenario, the predictive ability ranged from 0.212 to 0.417 for genomic best linear unbiased prediction (GBLUP) and from 0.249 to 0.565 for single-step GBLUP (ssGBLUP). In the forward prediction scenario, the predictive ability ranged from 0.069 to 0.389 for GBLUP and from 0.085 to 0.483 for ssGBLUP. In conclusion, the studied sperm morphology abnormalities showed moderate to low heritability. Both GBLUP and ssGBLUP showed comparative predictive abilities of breeding values, and ssGBLUP outperformed GBLUP under many circumstances in respect to predictive ability. To our knowledge, this is the first time that the genetic parameters and genomic predictive ability of these traits were reported in such a large Duroc boar population.
Highlights
Genomic prediction (GP) [1], termed genomic selection in the fields of animal and plant breeding, has advantages in precisely predicting genetic values, phenotypic values, and disease risks by integrating dense genetic marker panels into prediction models
The basic idea of GP is to trace the effects of all quantitative trait loci (QTL) using dense genomic markers that are in linkage disequilibrium
Total genetic values can be obtained by summing the effects of all markers [1] or solving mixed model equations (MMEs) with a marker-derived relationship matrix as a covariance matrix [2]
Summary
Genomic prediction (GP) [1], termed genomic selection in the fields of animal and plant breeding, has advantages in precisely predicting genetic values, phenotypic values, and disease risks by integrating dense genetic marker panels into prediction models. Total genetic values can be obtained by summing the effects of all markers [1] or solving mixed model equations (MMEs) with a marker-derived relationship matrix as a covariance matrix [2]. Many methods for either marker effect estimation [1,3,4] or direct genetic value prediction [2,5,6,7] have been established. Among the previously proposed GP approaches, single-step genomic best linear unbiased prediction (ssGBLUP) [5,6] is preferable for genetic evaluation in populations with large numbers of phenotyped individuals but few genotyped individuals, especially in the breeding programs of domesticated animals. The estimated breeding values (EBVs) are calculated by solving MMEs with H (a blend of both pedigree- and genetic-marker-derived relationship matrices) as a relatedness matrix. Previous studies have shown that ssGBLUP is perceived to be better or at least comparable in performance to conventional BLUP and the standard
Sign-in/Register to view highlights & summary 
Published Version (
Free)
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call