The Prior Information From GWAS And EQTL Increase The Accuracy Of Genomic Selection In Several Sheep Meat Traits

Revealing the genetic architecture underlying the meat trait variation and increasing the accuracy of genomic selection(GS)are two of important contents in the field of sheep genetic and breeding research because genetic gain is positively correlated with the accuracy of GS.However,higher genotyping costs and lower economic returns limited the revealing genetic architecture and the application of the GS in sheep breeding.Therefore,in the current study,we aimed to uncover the possible mechanism that single nucleotide polymorphism(SNP)regulates meat traits by lowcost whole-genome genotyping and RNA-sequencing.This study accurately quantified three types of molecular phenotypes(including gene expression,exon expression and splicing)in the liver and muscle and then detected expression quantitative trait loci(eQTL;including gene expression QTL,geQTL;exon expression QTL,eeQTL and splicing QTL,sQTL).Further,the current study implemented association analysis between molecular phenotypes and 41 meat traits.Finally,this study integrated the prior information from genome-wide association study(GWAS)and eQTL on the GS to investigate the accuracy change in several meat traits of muti-breed sheep population(most animals were either purebred Merino or crossbreds of Border Leicester,Poll Dorset or White Sufolk rams mated to Merino dams or Border Leicester ×Merino dams).These results may provide a low cost,high versatility and high accuracy GS strategy for meat sheep breeding.The results are as follows:1.Analysis of transcriptional molecular phenotypes in sheep liver and muscleA total of 298 RNA samples of muscle and liver from 149 sheep were sequenced.Three types of molecular phenotypes(gene expression,exon expression,and intron excision events)in two tissues were quantified and the heritability was estimated.The results suggested that in sheep liver,the median heritability of 3780 heritable genes(27.94%),51975 exons(81.37%)and 51940(56.64%)intron excision events were 0.17,0.26,and 0.13,respectively.The average heritability were 0.24,0.30 and 0.19.In muscle,the median heritability of 5143 heritable genes(39.60%),52247 exons(86.75%),and 53928 intron excision events(61.80%)were 0.22,0.36,and 0.16,respectively.The average heritability were 0.30,0.41,and 0.22.Many molecular phenotypes had low heritability indicating considerable environmental effects on the molecular phenotypes.Both in the liver and muscle,the heritability of exon and the ratio of heritable exons are higher than gene and splicing,indicating that the exon is an informative molecular phenotype.2.eQTL analysis in sheep liver and muscleThis study used the molecular phenotypes from result 1 and 149 imputed whole genome SNPs to implement eQTL detection.Association analysis between molecular phenotypes and nearby SNP(± 1Mb)identified many significant eQTL(P < 1e-5).In liver,441859 significant geQTL,478640 eeQTL and 794248 were identified.In muscle,287471 significant geQTL,327006 significant eeQTL and 523104 significant sQTL were identified.The median distance of eQTL from the transcription start sites(TSS)ranged from 85.92 Kb to 128.35 Kb across three types of eQTL.The number of overlapping SNPs across geQTL,eeQTL and sQTL within each tissue,and the number of overlapping SNPs between liver and muscle within each eQTL type were all significantly(P < 0.05)greater than that expected by chance.Identified eQTL were significantly(P < 0.05)enriched in GWAS hits associated with 56 carcass traits and fatty acid profiles.These results indicate that the SNP regulate sheep meat phenotype may by regulating the transcriptional molecular phenotypes.For example,hundreds of sQTL in the liver and muscle were mapped to CAST(calpastatin)and which was reported to be associated with shear force.3.Association analysis between molecular phenotypes and meat traitsAssociation analysis between meat traits collected from 149 individuals and three types of molecular phenotypes from result 1 were implemented.Majority of 41 meat traits are low-or median-heritability traits.The variance explained by molecular phenotype in muscle was higher that in liver.Association analysis between molecular phenotypes and 41 meat traits identified many significant molecular phenotypes,e.g.,in the liver,the abundance of GYS2 was significantly(P < 0.05)related to Ratio 630/580 day2(RCR2).In the liver,an intron excision event V76667(15:49393786-49506051,anchored to STIM1)was significantly(P < 0.05)related to carcass eye muscle area(CEMA).In muscle,the abundance of PSMB1 was significantly(P < 0.05)related to carcass eye muscle width(CEMW).4.The prior information from GWAS and eQTL on GS in several meat traitsAbout 13 K core SNPs were obtained by integrating the results of eQTL analyses from result 2 and GWAS results from previously published work.Core SNPs were mainly located on Chromosomes 1,2,3,6 and 11.These core SNPs were added to 50 K SNP chip.GBLUP,BayesR and BayesRC methods were used to evaluate the accuracy of GS in six meat traits in a large multi-breed population(9025 individuals with Fat Depth C;19071 individuals with Carcass Eye Muscle Depth;11991 individuals with Post Weaning Eye Muscle Depth;7611 individuals with Intra-muscular Fat;14638 individuals with Post Weaning Weight and 8734 individuals with Shear Force Day 5).The result shows that when the core SNPs were add to 50 K SNP panel,the accuracy of GBLUP increased 0.041,0.055 across six meat traits in the pure Merino sheep population when fitting one genomic relationship matrix(G)and two G matrices compare to 50 K SNP panel.In cross breed validation population,accuracy of GBLUP increased 0.0355,0.041 when fitting one G matrix and two G matrix,respectively.When the core SNPs were add to 50 K SNP panel,the accuracy of BayesR and BayesRC increased 0.0609,0.0613 in pure Merino sheep population.In cross breed validation population,the accuracy of BayesR and BayesRC increased 0.0655,0.0716,respectively.On average,integrating the prior information from GWAS and eQTL on GS can increase the accuracy of meat traits in sheep.