The Imputation Accuracy Comparison Of Low-density SNP Sets And Validation Of Genome Prediction Accuracy

Using genotype imputation to impute the low-density SNP chip to the high-density level for genomic selection,which provides a new idea to solve the cost problem caused by using high-density markers for genomic selection in large populations.This study using four marker screening methods to select representative sets of high-density SNP chips with different density gradients in the reference panel of Huaxi Cattle.And then using genotype imputation to impute the low-density SNP data to high-density ones with the same imputation parameters for subsequent genomic studies.The study compared the differences in imputation concordance among SNP panels and demonstrated the effects of several factors on imputation results,that is,marker screening method,marker density,minor allele frequency,reference population size and composition of reference population.Summarize the influence of different factors on imputation effect,to provide a reference for the design of low-density SNP imputation chips in Huaxi Cattle.The remaining 1431 Huaxi cattle after quality control screening was randomly divided into a reference population and a validation population.Four marker screening methods,Equidistance(EQ),High MAF(HM),High Linkage disequilibrium(HL)and High MAF-Linkage disequilibrium(HML)were used to screen 17 SNP sets of different densities from the Illumina Bovine HD chip locus set in the reference panel.Each low-density set was then imputed to the 770K density level in the validation panel using Beagle(v5.1)to calculate the imputation concordance.The study found;the higher SNP densities resulted the increase of the imputation concordance;the effect of selecting SNP loci containing linkage disequilibrium information for genotype imputation is better;the HML method had the highest imputation concordance at a density of 20K(EQ:0.9533,HM:0.9533,HL:0.9642,HML:0.9644);the imputation concordance increased with the increase of the reference population size,when the reference population size is less than 200,the increase in imputation concordance is the largest,and as the size of the reference group increases,the increase in imputation concordance progressively decreases(ICmin=0.9104,ICmax=0.9614).The composition of the reference population has an impact on the imputation results,and a reference population composed of individuals with similar age as the population can help improve the imputation results.As the marker’s own MAF progressively increases,the mean imputation concordance decreases gradually(ICmaf0.05=0.9817,ICmaf0.5=0.9010).The genotype imputation results will affect the accuracy of the GEBV prediction,and the higher the imputation concordance,the smaller the gap between the true GEBV.