Genomic Selection And Genome-wide Association Study Of Wool Quality And Erythrocyte Traits In Alpine Merino Sheep

Alpine Merino sheep is a first-class new breed of Merino sheep with wool fiber diameter of 19.1～21.5μm bred in the alpine cold and drought ecological area,they live in high-altitude cold and drought areas all the year round,it has the characteristics of high production performance,excellent wool quality,and strong adaptability to high altitude and low oxygen.However,the study about genomic selection and genome-wide association study of important traits such as wool quality and resistance to high altitude hypoxia related to its outstanding characteristics is still in infancy stage.How to improve the accuracy of breeding value estimation to shorten the generation interval,accelerate genetic progress,and discover important traits associated candidate genes have become an urgent problem to be solved in the improvement of Alpine Merino sheep breeding and the construction of the complete structure of the breed.In this study,adopting the method of genomic selection and genome-wide association study,taking Alpine Merino sheep as the research object,combining SNP microarray data of different densities,based on GBLUP and Bayes-Alphabet models,to study the influence of different factors on the accuracy of GEBV estimation;And adopting genome-wide association study to map QTLs associated with wool quality and high altitude hypoxia adaptability to search for crucial regional information and candidate genes.The results are as follows.1.The influence of additive and dominance genetic effects on the accuracy of genome predictionCombined with Affymetrix HD 630K microarray data,based on the GBLUP model,the model(MAG)with only additive genetic effects and the model(MADG)with both additive and dominance genetic effects were used to analyze the 9 wool quality traits of 498 Alpine Merino sheep genomic prediction of erythrocyte traits related to high altitude hypoxia adaptability,the estimation results of the genetic variance components showed that the dominant variances of the three traits of fleece extension rate,red blood cell count and hematocrit accounted for 73%,28%and 25%of the total phenotypic variance,respectively.For traits with a higher proportion of dominant variance,the genomic breeding value(GEBV)obtained by the MAG model is more accurate;The results of multiple cross-validation show that the prediction accuracy of the two models is higher than that of MADG except for the staple strength(R~2=0.25)and the mean corpuscular hemoglobin concentration(R~2=0.12).The above results indicate that the MAG model is more suitable for genomic prediction of wool quality and erythrocyte traits of Alpine Merino sheep.2.The influence of marker density,statistical model and heritability on the accuracy of genomic predictionAdopting the microarray data of two different densities of 50K and 630K,based on the GBLUP and Bayes-Alphabet model,821 Alpine Merino sheep with different heritability levels of 6 wool quality traits were used for genomic prediction and analysis.The heritability estimation results show that the heritability of fleece extension rate and staple strength are 0.29 and 0.35,respectively,which are the traits of medium heritability level;the heritability of staple length,fiber diameter,coefficient of variation of fiber diameter,and clean fleece weight rate were 0.68,0.44,0.55 and 0.46,respectively,which are the traits of high heritability level;When the marking density is increased from 50K to 630K,the prediction accuracy of fleece extension rate and staple strength are increased by 11%(Bayes A)and 13%(Bayesion LASSO),respectively,the clean fleece weight rate and the coefficient of variation of wool diameter are only increased by 1%(Bayes B),while the staple length decreased by 6%(Bayesion LASSO);In addition,the prediction accuracy of GBLUP model for traits with medium heritability level is higher than that of Bayes-Alphabet model,while models such as Bayes B and Bayesion LASSO have higher accuracy for traits with high heritability level.The above results show that increasing the marker density could effectively improve the accuracy of GEBV estimation for medium heritability traits such as fleece extension rate,but it has a weak effect on high heritability traits such as staple length,and even decreases the accuracy;The GBLUP model is more suitable for genomic prediction of traits with medium heritability,and the Bayes-Alphabet model is more suitable for genomic prediction of traits with high heritability.3.Genome-wide association study of wool quality and erythrocyte traitsAdopting single markers and haplotypes,based on the GLM model,perform genome-wide association study on erythrocyte traits.Through gene mapping and functional annotation,six genes DHCR24,EFNB2,SH2B3,PLCB1,SPATA9 and FLI1were selected as potential candidate genes for high altitude hypoxia adaptation,especially PLCB1 and FLI1,are closely associated with erythropoiesis,normal hematopoiesis and oxygen-carrying function,respectively;Adopting single markers,based on the Farm CPU model,perform genome-wide association study on wool quality traits,and four genes PBX1,TRPC3,SLITRK5 and PVRL1 were selected as potential candidate genes for wool quality traits,especially PBX1 and SLITRK5,which are associated with the senescence delay of hair follicle mesenchymal stem cells,the differentiation of the epidermis,and epidermal homeostasis.The above results could be used as a significant effect region for the prediction of Alpine Merino Sheep genome,and also provide valuable reference for the searching of functional genes of Alpine livestock.The study optimized the GBLUP model by including additive and dominant genetic effects,and found that traits with dominant variance such as FER accounted for a larger proportion of the total phenotypic variance(more than 25%),the MAG model with higher accuracy for estimated in GEBVs,it is more suitable for genomic prediction of wool quality and erythrocyte traits of Alpine Merino sheep.Increasing the marker density could effectively improve the estimation accuracy of GEBVs for medium heritability level traits such as FER,but has little effect on high heritability level traits such as SL;by comparing the GP accuracy of the two types of models,GBLUP is more suitable for genomic prediction of traits with medium heritability level,Bayes B and Bayes LASSO models are more suitable for traits with high heritability level.Based on different GWAS models and marker types,four candidate genes(PBX1,TRPC3,SLITRK5 and PVRL1)associated with wool quality traits and six candidate genes(DHCR24,EFNB2,SH2B3,PLCB1,SPATA9 and FLI1)associated with high altitude hypoxia adaptability were selected,especially PLCB1 and FLI1 is closely related to erythropoiesis,normal hematopoiesis and oxygen-carrying function respectively;PBX1and SLITRK5 are closely related to hair follicle mesenchymal stem cell senescence delay,epidermal differentiation and epidermal homeostasis,respectively.It provides suitable GEBVs estimation models and marker information with important effects for the genomic selection of Alpine Merino sheep,and provides valuable references for the selecting of functional genes of plateau livestock.