Effects Of SNPS With Different Densities And Low-Density Imputation Panels On Genomic Prediction Accuracy Of Harvest Weight Trait In Litopenaeus Vannamei

Litopenaeus vannamei is one of the three excellent shrimp species in the world because of its fast growth and strong stress resistance.The market share of domestic new varieties of Litopenaeus vannamei is small,and most provenances depend on foreign countries.In the genetic evaluation of breeding base population,the additive genetic variance may be overestimated because the genetic background of most founder populations is not clear and their genetic performance differences are ignored.Therefore,it is necessary to find an accurate model to evaluate the basic group.In traditional genetic evaluation methods,breeding value is usually estimated based on pedigree information,and the accuracy of SIB selection and inter family selection is not high.Genome selection can further improve the accuracy of selection by using high-density molecular genetic markers covering the whole genome.However,in practical application,due to the low economic value of Penaeus monodon,the relative cost of high-density genotyping is still very high,which greatly limits its application space.Therefore,we need to explore ways to reduce the cost of genome selection.The main objectives of this study are as follows:(1)by constructing pBLUP,pBLUP-GG,GBLUP,ssGBLUP and ssGBLUP-MF model,we can provide breeding analysis models and evaluation methods for accurately evaluating the breeding base population of Litopenaeus vannamei(2)GBLUP model and ssGBLUP model were constructed,and SNP panels with different densities were used to predict body harvest weight,in order to find the low density threshold for genome selection(3)GBLUP model and ssGBLUP model were constructed,and different filling scenarios are set according to the target panel density and reference population size.The filled SNP panel is used for genome prediction to further reduce the cost of genome selection.The main contents and results of this study are as follows:1.Genetic evaluation of harvest weight traits of Litopenaeus vannamei based on re sequencing SNP informationIn this study,we collected three genetic resource populations of Litopenaeus vannamei with different growth rate and survival rate as founder populations.Incomplete diallel hybridization was used to produce families,and basic breeding population was established.The whole genome of some individuals in the families was re sequenced,and4653383 SNPs were obtained.By constructing the pBLUP model,the pBLUP-GG model with genetic groups,the GBLUP model,the ssGBLUP model and the ssGBLUP-MF model with metafounders,the accuracy of breeding value estimation for body weight traits of breeding base population of Litopenaeus vannamei with different models was compared.The results showed that:(1)compared with pBLUP model,based on the whole genome re sequencing SNP data,the prediction accuracy of ssGBLUP model for body weight traits of Litopenaeus vannamei increased by 10.53%,and the prediction accuracy of GBLUP model decreased by 38.24%(2)Compared with pBLUP model,the heritability of pBLUP-GG model,ssGBLUP model and ssGBLUP-MF model decreased by 35.16%-68.13%,and the accuracy of genome prediction increased by 9.33%-10.53%.The results showed that the one-step genome BLUP(ssGBLUP)model combined with whole genome re sequencing SNP and pedigree information could effectively deal with the effect of unknown parent group,The predictive power(0.76)was higher than that of the genome BLUP(GBLUP)model(0.42)and the traditional pedigree BLUP(pBLUP)model(0.68).2.Genetic evaluation of harvest weight traits of Litopenaeus vannamei based on SNP information of different densitiesIn this study,GBLUP model and ssGBLUP model were constructed to explore the effect of two SNP panel construction methods on genome prediction of body weight traits of Litopenaeus vannamei using SNP panels with different densities.22 SNP panel densities were set,ranging from 0.1k to 500 K.At the same time,two methods of panel construction are set up,the first is to randomly select SNP sites in the whole genome,and the second is to randomly select SNP sites in the window.The results showed that when SNP marker density was reduced to 8k-10 k,there was no significant difference between the prediction ability of ssGBLUP and GBLUP models and the prediction results based on re sequencing data.Using this density can effectively evaluate the genome of harvested body weight and reduce the cost of genotype identification.3.Genetic evaluation of harvest weight traits of Litopenaeus vannamei based on genotype filling SNP informationIn this study,GBLUP model and ssGBLUP model were constructed to explore the effect of genome prediction of body weight traits of Litopenaeus vannamei by filling different density target panels to 10 K with genotype filling technology under different reference population sizes.There are two kinds of genetic relationships between reference population and target population.The first is that the genetic relationship between the reference population and the target population is random,and four filling reference population sizes are set at the same time,which are 25%,50%,75% and 90% respectively;The second is that the genetic relationship between the reference population and the target population is full sib,and the size of the reference population is 90%.At the same time,the density of 9 filling target individuals is set,which is 1k-9k,with an interval of 1K.The results showed that increasing the size of reference population and increasing the density of target SNP panel could increase the accuracy of genotype filling(0.76-0.79),but the prediction ability of GBLUP model based on SNP panel filling(0.15-0.34)was lower than that based on 10 K real genotype panel(0.44-0.72),and the prediction accuracy of ssGBLUP model was basically consistent with that based on 10 K real genotype panel.It is necessary to further design the low-density target panel and optimize the filling scheme to improve the accuracy of genome prediction.