Estimation Of Genetic Parameters And Genomic Prediction For Female Fertility Traits In Dairy Cattle

The intensive selection for milk production has led to the decline of female fertility in dairy cattle.Poor female fertility not only caused economic losses,but also aggregated environmental pressures.This thesis carried out an extensive investigation for female fertility in dairy cattle,including estimating genetic parameters,genome-wide association studies,genotype imputation,traditional genetic evaluation and genomic evaluation,and exploring genotype by environment interaction.The findings of this PhD project could help us better understand the genetic architecture underlying female fertility and genetically improve female fertility of current dairy cattle populations.First,(co)variance components for female fertility traits across parities were investigated in Chinese Holsteins.Heritabilities for female fertility traits were generally below 5%.The same performance of a trait in heifers and cows were genetically different,but different parities of cows were genetically similar.For the Chinese Holstein population,interval from calving to first insemination,interval from first to last insemination and non-return rate within 56 days after first insemination were suggested to be included in the selection index to improve female fertility.Second,quantitative trait loci(QTL)for female fertility traits were detected in Chinese and Nordic Holsteins,using single-trait association analysis,validation approach and meta-analysis.Variants associated with female fertility traits in Chinese Holsteins were validated with Nordic Holsteins.The power and the resolution of QTL detection were improved using validation approach and meta-analysis.Genes IL6R,SLC39A12,CACNB2,ZEB1,ZMIZ1 and FAM213A were suggested as candidate genes for female fertility traits.Third,imputation for single nucleotide polymorphism(SNP)selected from whole genome sequencing(WGS)data,in a customized LD chip,was investigated in Danish Jersey.These WGS SNP were selected in Nordic(NOR)dairy breeds from assiociation analyses and in French(FRA)dairy breeds from both association analyses and bioinformatics analyses.The correlations between imputed and observed genotypes and their concordance rate were used to measure the imputation accuracy.Imputation correlations were 91.6%for NOR SNP and 87.6%for FRA SNP,while concordance rates were 96.6%for NOR SNP and 93.5%for FRA SNP.Then,these imputed WGS SNP were investigated for genomic prediction in Danish Jersey.Compared with using 54K SNP alone,adding selected WGS SNP led to improvements of reliabilities for milk and protein,small improvements for fat and mastitis,and no improvement for female fertility.A Bayesian four-distribution model outperformed a GBLUP model for milk and protein,but extra gains in reliabilities from using the selected WGS SNP were smaller for a Bayesian four-distribution model than a GBLUP model.No significant difference was observed between considering 54K and the selected WGS SNP as a single genetic component or as two separate genetic components.Finally,genotype by environment interaction(GxE)under conventional and organic production systems were investigated for female fertility traits in Danish Holsteins,using a bivariate model and a reaction norm model.Heterogeneities in variance components were observed between conventional and organic production systems.Furthermore,significant GxE were observed for number of inseminations and non-return rate within 56 days after first insemination in heifers,and interval from calving to first insemination in the analysis using the bivariate model,and for interval from calving to first insemination and number of inseminations in cows in the analysis using the reaction norm model.This thesis performed estimation of genetic parameters,genome-wide association study,genetic evaluation and genomic prediction,and detection of QTL in several numerically small populations,including Chinese Holsteins,Danish Jersey,and Danish Holsteins from the organic production system.The results from this thesis provided reference for selection of low heritable traits in a numerically small population.