Genome Selection Of Meat Rabbits Based On Low-coverage Sequencing Data

The meat rabbit is a grain-type herbivore,which is high in production.Rabbit meat is rich in protein,minerals.And the fat and cholesterol content are low.Rabbit meat is known as "beauty meat","puzzle meat",which is favored by people.China is the largest country raising rabbits in the world,and the breeding of improved varieties of meat rabbits is the cornerstone of the development of meat rabbit industry.In the past ten years,conventional breeding methods have achieved remarkable results in the selection of meat rabbits.However,conventional breeding methods have the disadvantages of long cycle,high cost and low efficiency,which lead to deviations in the estimation of breeding values and affect the breeding of meat rabbits.In recent years,Genomic selection（GS）has been made up of conventional breeding mode.It has been widely recognized and applied in genetic improvement in livestock.The estimation accuracy of generated species value is affected by factors such as reference groups,marking density.With the application of low-coverage sequencing technology and genotype imputation method,it is possible to obtain large-scale whole genome sequence data by low cost.This will advance the development of genomic selection.This study is based on genome low-coverage sequencing techniques to assess factors such as different sequencing coverage,group sizes,allele frequency and imputation strategies and the effect of these factors on genotyping accuracy.And use accurately imputed genotype information for genomic selection of meat rabbit.Obtaining estimated effects of genomic breeding values under different models.Provide scientific basis and efficient means for genetic breeding of meat rabbits.The main findings are as follows.（1）In this study,Kangda V is a research object,and the meat rabbits genome selection reference group was constructed.Measure and organize phenotypic data of 16 growth and slaughter traits including feed conversion ratio,35-day-old weight,full evisceration weight,half eviscerated weight and carcass weight.Whole-genome sequencing of 1548 meat rabbits in the Kangda meat rabbit reference group was performed using low-coverage sequencing technology,with an average sequencing coverage of 2.66×.（2）Using three genotype filling strategies（Base Var+STITCH,Bcftools+Beagle4 and GATK+Beagle5）,five sequencing depths（0.2×,0.4×,0.6×,0.8× and 1×）,six population sizes（200,400,600,800,1000 and 1200）,nine Minor allele frequency（MAF）intervals（[0～0.05],[0.05～0.1],[0.1～0.15],[0.15～0.2],[0.2～0.25],[0.25～0.3],[0.3～0.35],[0.35～0.4],[0.4～0.45],and [0.45～0.5]）to impute the low-coverage sequencing data respectively,and the high-coverage sequencing typing results of the same individual are used as the standard,to assess the consistency and accuracy of genotype imputation.The results showed that Base Var+STITCH strategy has higher genotyping accuracy and is less affected by MAF.When the sequencing depth is 1× and the population size is ≥ 400,the genotype accuracy can reach 0.94,and the genotype consistency can reach 97.75%.When the sequencing coverage is lower（≥ 0.2×）,the population size needs to reach more than 600,or when the sequencing coverage is ≥ 0.4×,the population size needs to be more than 400,and high-accuracy genotypes can also be obtained,which can be used as ultra-low coverage sequencing recommended depth and group size.（3）The single-trait genomic best linear unbiased prediction（GBLUP）model and the single-trait best linear unbiased prediction（BLUP）model were used to estimate individual breeding values.The random cross-validation results showed that genome selection accuracy is overall better than conventional BLUP.Further comparison of the Genomic breeding value（GEBV）estimation accuracy of growth traits and slaughter traits under the single-trait GBLUP model and different multi-trait GBLUP models.The random cross-validation results showed that for most growth traits,the accuracy of GEBV estimation obtained by the multi-trait GBLUP model was slightly improved or basically the same as that obtained by the single-trait GBLUP model,but the accuracy of GEBV prediction of body length was significantly lower.Among them,for the time-consuming,labor-intensive,and difficult to measure feed conversion ratio traits,the GEBV prediction accuracy of the two-trait GBLUP model using body length and feed conversion ratio is significantly higher than that of the single-trait GBLUP model.The multi-trait GBLUP model for the combination of various growth traits and slaughter traits has a significantly higher GEBV estimation accuracy for most slaughter traits than the single-trait GBLUP model,and the GEBV estimation accuracy for half eviscerated weight is the most obvious.In conclusion,whole-genome low-coverage sequencing combined with genotype imputation can provide accurate genotype data and provide reliable and comprehensive genomic information for subsequent genetic analysis,such as genome selection.Compared with conventional breeding selection methods,genomic selection has a greater advantage in meat rabbit breeding,and the multi-trait GBLUP model is generally beneficial to improve the prediction accuracy of genome selection,providing an important scientific basis for promoting the precise breeding of important economic traits in meat rabbits.