Skip to main content
ARS Home » Research » Publications at this Location » Publication #196609
Title: PRIOR GENETIC CORRELATIONS AND NON-MEASURED TRAITS

Author
item MARK, T - ROYAL VET AND AGRIC UNIV
item SULLIVAN, P - CDN
item FIKSE, W - INBERBULL CENTRE
item Vanraden, Paul

Submitted to: Interbull Annual Meeting Proceedings
Publication Type: Proceedings
Publication Acceptance Date: 6/4/2006
Publication Date: 7/1/2006
Citation: Mark, T., Sullivan, P.G., Fikse, W.F., Van Raden, P.M. 2006. Prior genetic correlations and non-measured traits. Interbull Annual Meeting Proceedings. Interbull Bull. 35:72-75.

Interpretive Summary: Current international genetic evaluations are based on how related country populations are genetically (across-country genetic correlations). Those correlations may be influenced strongly by prior expectations that were not based on sound scientific principles. Objective methods to predict prior correlations were investigated, and their possible use in international evaluations of measured and non-measured traits was illustrated with national genetic evaluations for Holstein milk yield. Possible sources of variation associated with estimated genetic correlations were identified and grouped: 1) climatic variables, 2) production system indicators, and 3) national genetic evaluation descriptors. Variation was explained best by a statistical model that included country ratios for milk yield, grazing status, wind speed, temperature, heritability, parities included in evaluations, and the number of bulls in common between the two countries. This model could be improved by adding other variables (for example, herd size, average months of grazing, or peak milk yield). Eliminating sources of variation in the model that may not be related to a true interaction of genotype and environment (for example, heritability, parities, and number of bulls in common) may be desirable when generating prior genetic correlations. Such prior correlations can be used to improve post-processing of estimated genetic correlations among measured traits and to predict international breeding values for nonmeasured traits.

Technical Abstract: Current international genetic evaluations are based on how related country populations are genetically (across-country genetic correlations). Those correlations may be influenced strongly by prior expectations that were not based on sound scientific principles. Objective methods to predict prior correlations were investigated, and their possible use in international evaluations of measured and non-measured traits was illustrated with national genetic evaluations for Holstein milk yield. Possible sources of variation associated with estimated genetic correlations were identified and grouped: 1) climatic variables, 2) production system indicators, and 3) national genetic evaluation descriptors. Variation was explained best by a statistical model that included country ratios for milk yield, grazing status, wind speed, temperature, heritability, parities included in evaluations, and the number of bulls in common between the two countries. This model could be improved by adding other variables (for example, herd size, average months of grazing, or peak milk yield). Eliminating sources of variation in the model that may not be related to a true interaction of genotype and environment (for example, heritability, parities, and number of bulls in common) may be desirable when generating prior genetic correlations. Such prior correlations can be used to improve post-processing of estimated genetic correlations among measured traits and to predict international breeding values for nonmeasured traits.