Holstein Conversion Equations Based on Population Variances and a Full-Brother Model

A full-brother method was used to develop conversion equations that predict US Holstein ETA for milk, fat, and protein yields from French EBV. Regression coefficients were ratios of genetic standard deviations of sires for the two populations adjusted for the genetic correlation between countries. Based on data from full brothers, the regression coefficients were used to determine the base differences of countries (intercepts), scaled to the variation in the importing country. Full-brother equations also were developed to convert evaluations from Canada and The Netherlands to a US equivalent so that the full-brother method could be compared with traditional methods. Genetic correlation with the US was assumed to be .9 for French data and 1.0 for data from Canada and The Netherlands. Standard deviations for the populations were computed from nearly 5900 US bulls and about 2000 bulls from' each of the other countries. The US had 178 full-brother families in common with France, 123 in common with Canada, and 46 in common with The Netherlands. For conversion of Canadian evaluations to a US equivalent, regression coefficients from the fullbrother method were 20% lower than from the Goddard method and 25% lower than from the Wilmink method. (