Epsilon Archive for Student Projects

Abstract

Progress in dairy cattle breeding is facilitated by good methods to calculate estimated breeding value (EBV) from several countries. It is now possible to exchange genetic material across countries. In 1983 Interbull was formed with the intention to increase the transparency of national genetic evaluation systems in the member countries to facilitate the across-country evaluations. Today, Interbull publishes results of routine evaluations for production, conformation traits, udder health longevity and calving traits.

The objective of this study was to extend the method of conversion of breeding values from single trait to multiple traits. Bulls with EBVs in both Sweden and Denmark were used for the conversions.

A data set containing the information on bulls with evaluations in both Sweden and Denmark was used for this study. The Swedish information included EBVs for bulls, their effective number of daughters and the heritabilities used for different traits. The Danish information included EBVs for bulls together with their corresponding reliabilities. The bulls were born 1986 or later. The common bulls originated from seven countries: Sweden, Denmark, the Netherlands, Germany, France, the United States and Canada.

To find the Danish traits best suited for conversion the traits with correlation of EBVs significantly different from zero were used.

Three different methods of converting Danish traits that showed high genetic correlation with corresponding Swedish EBV were performed and a converted total merit index (C_TMI) were calculated for each of the methods.

The first method was the Wilmink method (Wilmink et al., 1986). One Danish trait was used for calculation of a converted Swedish EBV which in turn was multiplied by the economic weights used in calculation of the Swedish total merit index (TMI) to calculate a converted TMI (C_TMIWilmink).

One alternative to the Wilmink method was to use two or more Danish traits for calculation of converted breeding values for each of the Swedish traits and then combining these into the Swedish TMI (C_TMIindirect reg)

Another alternative to the Wilmink method was to use two or more Danish traits for direct computation of a converted Swedish TMI (C_TMIdirect reg).

The bulls were ranked in three different Top 20 lists. Correlations between the three different methods were calculated, both for all bulls in the material and for the Top 20 bulls. For the Top 20 bulls the best correlation was observed for C_TMIWilmink and C_TMIindirect reg. The same procedure performed for all bulls in the material gave the best correlation for C_TMIdirect reg. and C_TMIindirect reg.