Sire-dam pedigree in MACE - results from a full-scale pilot study

Earlier studies (e.g. De Jong, 2003) have indicated that international genetic evaluations are sensitive to definition of genetic groups. De Jong (2003) suggested using sire-dam pedigree instead of sire-MGS pedigree in order to get the phantom parent groups further away from the animals with data. Based on this hypothesis Van der Linde et al. (2005) collected sire-dam pedigree, made appropriate changes to the HAUSA MACE software (Klei, 1998; Klei & Weigel, 1998) and conducted a pilot study for seven countries participating in the evaluation for the Holstein breed for protein yield. Van der Linde et al. (2005) estimated genetic correlations and predicted breeding values using both the sire-dam pedigree and the sireMGS pedigree. The genetic correlations were rather similar independent of the pedigree structure (average decrease of 0.017) while the CPU time increased with a factor 28. For breeding value prediction the computing time changed with a factor 9 when changing the pedigree structure. Van der Linde et al. (2005) also looked at predictability of proofs and found an increase in predictive ability when using the sire-dam pedigree compared to the sire-MGS pedigree.