Accuracies of Genomic Prediction of Traits Associated with Lactation and Reproduction in Yorkshire and Landrace Sows

and Implications Genomic prediction involves statistical methods to estimate the genetic merit of selection candidates based on genetic markers spaced across the genome. The benefit of genomic prediction depends on the accuracies with which we can predict the genomic estimated breeding values (GEBV) of selection candidates based on their marker genotypes. The objective of this study was to estimate the accuracies of GEBV for traits associated with lactation and reproduction in Yorkshire and Landrace sows. Across both breeds, genomic predictions had greater accuracy than pedigree-based predictions. This result suggests that accuracy of selection can be improved by genomic prediction and, thereby, increase selection response compared to pedigree based genetic evaluation. Introduction Genetic improvement involves selection of superior animals as parents of future generations. Traditionally, animals are selected based on estimated breeding values (EBV) based on phenotypic records and pedigree. This method has been very successful for easy to measure production traits, but not for traits that are difficult to measure or for traits with low heritability. Many of these difficult to measure traits (e.g. feed efficiency, disease resistance, etc.) are of high economic importance. Genomic prediction is a method for estimation of breeding values based on marker information (GEBV) from high-density SNP chips and can be used to overcome these limitations. Most traits associated with lactation and reproduction in pigs are either less heritable, or appear later in life, or are difficult to measure on a routine basis. So for these traits, genomic prediction can be an attractive alternative to traditional pedigree based methods. Before practical implementation of genomic prediction methods, accuracies