Use of phenotypes from research herds to develop genomic selection for scarcely recorded traits like feed efficiency

Introduction There has been long running interest in how feed intake and feed efficiency should be taken into account in breeding decisions (for review Veerkamp, 1998). Initially the interest in feed intake was based on trying to reduce the amount of feed required per unit of production, i.e. improving feed efficiency. However, in the past two decades, interest has shifted towards the role of feed intake and its relationship with energy balance (EB), health, and fertility. There is, as yet, no direct selection practiced for feed efficiency or EB using actual feed intake observations. This is primarily because the large resource demand of measuring, particularly, individual feed intake in dairy cows. This makes routine selection in breeding programs too difficult. Similar arguments hold for detailed fertility measures using progesterone (van der Lende et al., 2004), methane (Wall et al., 2010) and several disease traits. An alternative might be to combine existing datasets from, for example, research herds in different countries and use these as a reference herd for calibrating a SNP key. In the RobustMilk database we combined data from research herds in four countries to generate sufficient data to achieve the research objectives of the project. Research was undertaken on how to combine the data (with different recording systems, feeding systems, and genetic groups), QTL detection, and statistical models. In this study the objective was to test the accuracy of the genomic breeding values from the RobustMilk database in predicting the breeding values based on progeny information in the UK, The Netherlands and Ireland.