Accounting for heterogeneous variance in a test-day model for joint genetic evaluation of Austrian and German Simmental cattle.

M. Lidauer, R. Emmerling and E.A. Mäntysaari 1 Agrifood Research Finland MTT, Animal Production Research, 31600 Jokioinen, Finland 2 Bavarian Institute of Animal Production, Postfach 1180, 85580 Poing-Grub, Germany INTRODUCTION A joint genetic evaluation of Simmental cattle would improve the ongoing close co-operation in breeding work of Austria, Bavaria and Baden-Württemberg. The countries comprise very heterogeneous herd environments due to large differences in climate and topography as well as in herd management. Hence, variances in the joint test-day (TD) data are heterogeneous, which should be accounted in the breeding value estimation. In this work we present an adaptation of the multiplicative mixed model (MMM) of Meuwisssen et al. (1996) for a multiple-trait random regression (RR) TD model. MATERIAL AND METHODS Data. The joint TD data from Austria, Bavaria and Baden-Württemberg included 88 million milk yield observations from all lactations of 4.2 million Simmental cows, which calved the first time from 1990 onwards. The joint pedigree included 6.2 million animals. Multiplicative random regression model. TD observations, which belonged to the same production year × month × region × parity class and to the same herd × TD (HTD) class were assumed to be homogeneous and represented a stratum i. Heterogeneity was assumed across strata. The TD milk yields (y) of the first (F), second (S), and all later (L) lactations were modelled by a multiplicative multiple-trait reduced rank RR TD model: