The nature and scope of some whole genome analyses in US beef cattle

The production of EPDs from statistical analysis of pedigree and performance records is a tried and true method for characterizing the genetic merit of animals. However, response to selection on such EPDs is suboptimal because there are no EPDs for some economically-relevant traits and for those that are available the accuracies are typically low at the ideal selection ages around puberty, especially for traits with low heritabilities, or those that can be measured in only one sex or late in life. In theory, more accurate genomic enhanced EPDs could be obtained in target populations of selection candidates from prior knowledge of EPD associated with various chromosome fragments. This prior knowledge is typically obtained from Bayesian analyses in a so-called training or discovery population. Three critical issues arise in relation to this process. First, what is the upper limit for accuracy or predictive ability using a high-density 50k genomic panel? Second, how is predictive ability influenced by genetic distance between the training and target population? A close distance would involve training in widely used sires for prediction in their offspring, whereas a long distanc