3 METHODS TO VALIDATE THE ESTIMATION OF GENETIC TREND FOR DAIRY-CATTLE

Three methods are proposed to validate the estimation of genetic trend for dairy cattle. With the first method, official evaluations, which are generally derived from a repeatability animal model applied to lactation records from several parities, are compared with evaluations based only on first parity to determine whether estimates of genetic trend are similar. With the second method, daughter yield deviations are analyzed within sire by calving year to determine whether these deviations remain stable over time. With the third method, variations of successive official evaluations are analyzed by regression to determine any systematic trend associated with information from additional daughters. Access to raw data is needed for the first two methods, but only published evaluations are needed for the third method. The three methods were applied to French Holstein evaluations calculated with 1990 and 1993 animal models and validated estimates of genetic trend based on information from the 1993 animal model.