Comparison of test-day and lactation models for genetic evaluations of Icelandic dairy cows for production traits and somatic cell score

Predicted genetic progress in production traits was compared using three different models for genetic evaluation of Icelandic dairy cows. The models were: a random regression (RR) test-day model, model using lactation yields until day 305 from calving (LAC1), and the model currently used for the national evaluation, based on lactation yield from calving to the end of lactation regardless of the length of the lactation (LAC2). Additionally, genetic evaluation for somatic cell score with RR and LAC1 were compared. Predicted genetic progress for protein yield was highest when using RR, or 0.170 σ_a/yr, compared with 0.167 σ_a/yr and 0.158 σ_a/yr for LAC1 and LAC2, respectively. Results for other production traits were similar. The main reason could be the shorter generation interval when records can be utilized before the end of lactation and reliable estimated breeding values thus obtained earlier. Application of an RR model will be beneficial for genetic evaluation for production traits and somatic cell score.