The Cenomic Selection Analysis And Re-analysis Of Quantitative Traitloci Of Hybrid Yield Of Rapeseed

Rapeseed(Brassica napus)is one of the important oil crop widely grown in the world.How to take the advantage of the heterosis for rapeseed production and efficiently carry out hybrid breeding is always the topic research of great interest in rapeseed.In this study,a double haploid population denoted as TN DH population,derived from the cross between the cultivar of B.napus Ningyou7(Chinese semi-winter type)and Tapidor(European winter type),and its reconstructed F2 population(RC-F2)were used as plant materials.Based on the phenotypes of multiple traits and the genotypes of TN DH population by the chip of Brassica Illumina 60k collected by previously researchers,we detected the QTLs associated with rapeseed hybrid seed yield.We constructed statistical models link the phenotypes and genotypes,estimated the prediction accuracy on hybrid seed yield by marker-assisted selection and genomic selection models.We analyzed the effects of different genetic factors contributing to heterosis and the impact on the prediction accuracy of genomic selection.The main results are as follows:1.Reanalyzed the seed yield of DH and RC-F2 populations across three environments,for its genetic architecture of heterosis.The results show a strong heterosis of the hybrid seed yield.The dominance effect has more importance than epistasis effect on hybrid seed yield.2.Compared the prediction accuracies between marker-assisted selection and genomic selection.The results show that the accuracies of genome selection(GBLUP?BayesCn)were much higher than marker-assisted selection for hybrid seed yield.3.Compared the influences of different genetic factors on prediction accuracies.The results show that the dominance model(considered additive and dominance effects)improved 33%prediction accuracy compared with additive model,while the epistasis model(considered additive,dominance and epistasis effects)improved 9%prediction accuracy compared with dominance model.The dominance effect has more importance than epistasis effect on genomic selection of hybrid seed yield.4.Estimated the seed yield performance of 16110 potential hybrids and planted proving population for the future analysis.Selected the top 1%highest seed yield performance of the potential hybrids and its average predict value was 16%higher than the 318 hybrids.Simulation predictions of genomic selection can detect more excellent hybridized combinations.5.Analyzed the prediction accuracies of genomic selection with different agronomic traits of TN DH population.The results show the accuracies of genomic selection were positively correlated with the heritability of different agronomic traits(except for oil content).We compared the prediction accuracy between marker-assisted selection and genomic selection,the influences of different genetic factors on the prediction accuracy.We expect to establish a reasonable and effective genomic selection model that can effectively predict yield and other complex traits in rapeseed.That may be able to provide a reference for genomic selection of rapeseed.