A Single Step SNP Model Applied to Test-Day Data of Dairy Cows

Abstract Text: A single step SNP model was applied to test-day data of German Holstein cows to evaluate genotyped and phenotyped animals jointly and account for the impact of genomic pre-selection.  A total of 343 million test-day records from 19.2 million cows were analyzed with a multiple lactation random regression test-day model, and >100,000 genotyped animals were considered with a genomic model assuming all SNP markers explaining equal genetic variance. In addition, international phenotypes of c.a. 113,000 Holstein bulls were added in order to use a multiple-country genomic bull reference population. A residual polygenic effect was fitted to the single step SNP model to reduce the overestimation problem of genomic prediction. Computing strategies for estimating effects of the model, including SNP effects, were developed. Technical issues for a routine implementation were addressed using the data of German Holsteins.

Keywords: single step genomic model, marker effects, test-day data, dairy cows