Quantitative milk genomics: estimation of variance components and prediction of fatty acids in bovine milk

TY - GEN

T1 - Quantitative milk genomics: estimation of variance components and prediction of fatty acids in bovine milk

AU - Krag, Kristian

PY - 2012/11/2

Y1 - 2012/11/2

N2 - The composition of bovine milk fat, used for human consumption, is far from the recommendations for human fat nutrition. The aim of this PhD was to describe the variance components and prediction probabilities of individual fatty acids (FA) in bovine milk, and to evaluate the possibilities of altering the FA compistion. Variance components were estimated by only using information from SNP markers. The prediction of FA was based on mid-infrared spectra's, and different strategies for variable selection and modeling approaches were tested. Results showed that variance components could be estimated from SNP markers, with a performance similar to traditional pedigree approaches. The heritability and correlation estimates indicate, that the composition of saturated FA and unsaturated FA can be altered independently, though selection and regulations in feeding rgimes. For the prediction FA, vaiable selctions and modeling approaches were found only to have a minor impact on the prediction

AB - The composition of bovine milk fat, used for human consumption, is far from the recommendations for human fat nutrition. The aim of this PhD was to describe the variance components and prediction probabilities of individual fatty acids (FA) in bovine milk, and to evaluate the possibilities of altering the FA compistion. Variance components were estimated by only using information from SNP markers. The prediction of FA was based on mid-infrared spectra's, and different strategies for variable selection and modeling approaches were tested. Results showed that variance components could be estimated from SNP markers, with a performance similar to traditional pedigree approaches. The heritability and correlation estimates indicate, that the composition of saturated FA and unsaturated FA can be altered independently, though selection and regulations in feeding rgimes. For the prediction FA, vaiable selctions and modeling approaches were found only to have a minor impact on the prediction

M3 - PhD thesis

SN - 978-87-92936-24-0

PB - Aarhus University, Faculty of Science and Technology

ER -