Genome-Wide Association Study And Genomic Selection For Reproductive Traits In Large White Pigs

Reproductive traits directly affect the production efficiency and economic benefits of pig farms.Studying and selecting sows with excellent reproductive traits is an important way to improve the competitiveness of the pig industry.The genetic improvement of important traits of pigs by traditional methods has achieved remarkable results in the past.However,traditional breeding methods have the problems of long cycle and high cost.Therefore,the use of new breeding theories and methods is of great significance for accelerating genetic improvement.In this study,genome-wide association analysis and genome selection were introduced into the genetic breeding of breeding pigs,in order to provide technical support for genetic improvement of breeding pigs.The main results obtained in this study are as follows:(1)The ss GWAS analysis of TNB,NBA and NHP phenotypes in Large White pigs was performed using a 50 K SNP chip,and 55 genomic regions were identified,of which the 95.54-97.31 Mb region on chromosome 3 explained the largest genetic variation.Through gene annotation and HPO analysis,28 candidate genes related to reproduction were found.Among them,some genes related to litter size were found,including CAMKMT,FXR1,PTPN21,ROBO2,ADARB1 and so on.This provides a reference for genetic mechanism analysis and breeding of pigs.(2)The accuracy of genome selection for three reproductive traits of Large White pigs was evaluated by using traditional GS methods(BLUP and Bayes)and ML methods(SVR and CNN).Compared with the traditional GS method,the ML method has more advantages in the accuracy of genome selection,especially the radial kernel method of SVR.However,due to the lack of hyperparameter tuning,the unbiasedness of model fitting is lacking.Combining accuracy and unbiasedness,the performance of GBLUP method also has certain advantages.(3)By adding additive,dominance and epistatic effects to the GBLUP repeatability model,the genetic parameters,model fitting degree and model prediction accuracy of reproductive traits of Large White pigs were compared and analyzed.It can be seen that the narrow heritability range of the three traits is 0.03 ~ 0.08,and the broad heritability range is 0.06 ~ 0.17.After adding non-additive effects to the model,the generalized heritability was improved in all three traits.Secondly,the model fitting degree considering nonadditive effect is better.Finally,considering the additive-nonadditive effect can improve the accuracy of genome selection,the most improved model is the A + D + AA model.In this study,genomic regions significantly associated with reproductive traits and important candidate genes affecting reproductive traits were obtained.Secondly,using the superiority of ML in analyzing the nonlinear relationship of complex traits,it is found that the GS method using ML strategy is more accurate than the traditional GS method.At the same time,it is found that the comprehensive consideration of additive-nonadditive effects in the GBLUP model can also improve the accuracy of genome selection.The results provide a new molecular marker and breeding method for the genetic improvement of reproductive traits in Large White pigs.It enriched the genetic research content of reproductive traits.