Accuracy of genetic prediction obtained using genetic relationships based on pedigree or marker information

Maiwashe, A. and Garrick, D.J. 2006. Accuracy of genetic prediction obtained using genetic relationships based on pedigree or marker information. In: Rege, J.E.O.; Nyamu, A.M.; Sendalo, D. (eds.). 2006. The role of biotechnology in animal agriculture to address poverty in Africa: Opportunities and challenges. Proceedings of the 4th All Africa Conference on Animal Agriculture and the 31st annual meeting of Tanzania Society for Animal Production, Arusha, Tanzania, 20–24 September 2005. Dar es Salaam, Tanzania: TSAP and Nairobi, Kenya: ILRI.

The accuracy of genetic prediction when pedigree ( P A ) or pedigree and marker-based ( PM A ) relationships were used in mixed model equations was assessed. Daughter yield deviations (DYD) records for milk, fat and protein yields on Holstein sires (n = 1,811) with DNA marker information were analysed. Three sets of estimated breeding values (EBV) were obtained: (1) EBV-ALL (considered all sources of information and the inverse of), (2) EBV-PED (similar to EBV-ALL but the sire’s own DYD was excluded), and (3) EBV-MRK (similar to EBV-PED except that the inverse of was used). Linear and rank correlations were computed between EBV-ALL and EBVPED or EBV-MRK. The accuracy increased by 4.3% for milk yield but did not change for fat and protein yields when was replaced by for sires without sons. These results suggest that genetic markers may be used to enhance accuracy in genetic evaluation particularly for young animals.