Genome-Wide Association Studies For Economically Important Traits In Dairy Cattle

Most economically important traits in dairy cattle including milk yield, milk fat yield and milk protein yield, are quantitative traits influenced by a large number of genes as well as many environmental factors. The identification of the genetic factors contributing to economically important traits will not only provide some genetic clues for further investigations on the genetic mechanism underlying these traits, but also help us understanding the impact of the long-term directional artificial selection on the genetic properties of the reference populations. Furthermore, such findings could guide the establishment and modification of the breeding objectives in the current dairy cattle breeding programs.Genome-wide association study proposed by human geneticists could achieve the whole genome scan for the detection of causal genetic variants by exploiting linkage disequilibrium at the population level. As the development of high-density SNP panels for many species, this approach had been gradually applied to the genetic dissection for the quantitative traits in a wide variety of domesticated animals, plants and model organisms.The international estimated breeding values of Brown Swiss progeny proven bulls were collected routinely at Interbull Center (Uppsala, Sweden), as well as the genome-wide SNP data typed by using the Illumina Bovine SNP50Beadchip (including54001SNPs). In the present study, a single-marker linear mixed model was employed to carry out GWAS analyses for identification of the genetic loci affecting nine economically important traits:milk yield (n=5043), milk fat yield (n=5043), milk protein yield (n=5043), lactating cow’s ability to recycle after calving (n=3807), body depth (n=4402), stature (n=4610), angularity (n=2061), somatic cell score (n=4803), and milking speeding (n=4411). In summary, the main results were as follow:1. The analyses of confounding factors in the sampled populationBoth the principal component analysis and the simple linear regression were used to investigate what was the primary confounding factor. The result of the principal component analysis indicated the risk of population stratification was very weak in this sampled population and could be neglected in the subsequent association analysis. In contrast, the multiple familial relationships were the major confounding factor. The genome-wide association results from simple linear regression showed the huge inflation existed between the expected and the observed P-values of genotyped SNPs, and the correction for such false positives based on such the model with genomic control was not suitable for the current analyses.2. Genome-wide association analyses for nine traits based on linear mixed modelAs shown in the quantile-quantile (Q-Q) plots for nine traits, the inflation of the association P-values from linear mixed model was low, but the λ values were still significantly larger than1. However, the λ1000values were close to1.05, which is considered generally benign. Based on linear mixed model and genomic control, a total of74genome-wide significant SNPs distributed across12chromosomes were identified for these traits, of which24SNPs were significant for multiple traits and the number of SNPs only significantly associated with a single trait were50. In particular, a total of four associated genomic regions distributed on chromosome6,11,24and25were very remarkable.3. The conditional genome-wide analyses for milk production and body size traitsThe conditional genome-wide association analyses were performed, in which most significant SNP and de-regressed EBVs of stature was included in the linear mixed model as a covariate, for the further investigation on the mechanism underlying the association region on BTA25. Such results suggested that only one QTL with pleiotropic effects contributing these traits harbored at this physical region.4. Partitioning of total genetic variance for milk production traits and statureBoth the total genetic variance and the genetic variance explained by each chromosome separately were estimated based on the method of genomic partitioning for total genetic variance. The results showed that the common SNPs on all the autosomes could explain a very large proportion of the phenotypic variance for these five traits. In the meantime, the results from linear regression for the proportion of genetic variance explained by each chromosome against the physical length of each chromosome suggested the weakly positive linear relationship between them (P<0.05). Overall, the result demonstrated that the polygenic hypothesis war true for these traits.5. The functional annotation for the distinct association regionsThe functional annotation for the distinct association peaks distributed on the four chromosomes was conducted by using AnnotQTL that is web-based software. Together with the previous reports,1GFALS was very likely to be a plausible candidate gene influencing milk production traits, stature and other relevant traits, and1L-8gene could be as a candidate gene contributing to milking speed. The present study demonstrated that genome-wide association study is indeed an efficient approach to identification of genetic factors influencing complex traits in domesticated animals. Here a total of74SNPs distributed on12chromosomes were detected to be genome-wide associated with the nine traits, in particular the four remarkable genomic regions on chromosome6,11,24and25, repectively. Several genes harbored at these regions (e.g. IGFALS, IL-8) could be the plausible candidate genes which was deserving of further validation in other populations and molecular dissection to explore the potential economic impact and the genetic mechanisms underlying these production traits in cattle. The partitioning of total genetic variance implied the milk production traits and stature conform to the polygenic inheritance.