Title: Application of the a posteriori granddaughter design to the Holstein genome Authors
Submitted to: Animal
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: December 20, 2013
Publication Date: April 1, 2014
Repository URL: http://handle.nal.usda.gov/10113/58858
Citation: Weller, J.I., Cole, J.B., Van Raden, P.M., Wiggans, G.R. 2014. Application of the a posteriori granddaughter design to the Holstein genome. Animal. 8(4):511-519. Interpretive Summary: A posteriori granddaughter design was applied to the entire genome to determine haplotype effects for 33 traits of Holsteins. For all traits, the haplotype effect was important in at least 1 chromosomal region. Confidence intervals ranged from 1 to 8 haplotype segments of ~75 markers each. If 50 grandsire families were analyzed and at least 5 grandsires were segregating for a QTL, the hypothesis of obtaining concordance by chance for even a single nucleotide polymorphism could be rejected with high power.
Technical Abstract: An a posteriori granddaughter design was applied to determine haplotype effects for the Holstein genome. A total of 52 grandsire families, each with >=100 genotyped sons with genetic evaluations based on progeny tests, were analyzed for 33 traits (milk, fat, and protein yields; fat and protein percentages; somatic cell score; productive life; daughter pregnancy rate; heifer and cow conception rates; service-sire and daughter calving ease; service-sire and daughter stillbirth rate; 18 conformation traits; and net merit). Of 617 haplotype segments spanning the entire bovine genome and each including ~5 × 10^6 bp, 5 cMorgans, and 50 genes, 608 autosomal segments were analyzed. For all traits, there was at least one chromosomal region in which the nominal probability for the haplotype effect was < 10^-8, which corresponds to genomewide significance of <10^-4. Net merit had 7 chromosomes with nominal probabilities of <10^-8. For each of those putative quantitative trait loci (QTL), at least 1 grandsire family had a within-family contrast with a t-value of >3. Confidence intervals were estimated by the nonparametric bootstrap for the largest effect for each of 9 traits. The bootstrap distribution generated by 100 samples was bimodal only for net merit, which had the widest 90% confidence interval (8 haplotype segments). For all other chromosomes, the confidence interval spanned less than a third of the chromosome. The narrowest confidence interval (a single haplotype segment) was found for somatic cell score. If 50 grandsire families were analyzed and >=5 grandsires were segregating for a QTL, the hypothesis of obtaining concordance by chance for even a single nucleotide polymorphism could be rejected with high power. Determination of the actual polymorphisms responsible for observed genetic variation should increase the accuracy of genomic evaluations and rates of genetic gain.