Genetic Architecture Analysis Of Restorer Recurrent Selection Population In Brassica Napus

Rapeseed(Brassica napus.L)is one of the most important oil crop wordwide.The production and breeding of rapeseed is great significant for edible oil supply in our country.China is one of the original centers of Brassica rapa,and has rich genetic resources.However,China has a short cultivar history of only about 100 years of Brassica napus.So the breeding resource is very limited.Heterosis is an effective way to increase crop yield,and it is depended on the genetic distance between the restorers and maintainers of male sterile line.These negative factors greatly limited the untilization of heterosis in China.Population improvement,such as recurrent selection,can not only improve the combining ability level of breeding population,but also domesticate exotic germplasm.It can breed new genotypes or resources for breeding through recombinant and selection.In this study,we have constructed a new recurrent selection population named R,which genetic background contained European winter-type rapeseed,artificial synthetic Brassica napus,subgenomic materials,Pol CMS restorer lines and yellow seed-coat materials.The Ogura CMS restorer gene Rfo was employed as the random pollination system.We used Brassica SNP60 K Bead Chip to genotype a group of 175 individuals from the population,and evaluate the GCA level and heterosis performance of these individuals by using three testers in three different environments,and put the huayouza62 as the control.We investigated eight traits in the field trails,including Single plant Yield(SPY),number of branches(NB),main inflorescence length(LMI),silique number of main inflorescence(SNMI),total silique numberper plant(TSN),length of silique(LS),seeds number per silique(SNS),thousand seeds weight(TSW).Genetic structure and selective sweeps of the R population were also analyzed,and the main results are as follows:1.Phenotypic variation and correlation coefficients of eight traitsAnalysis of phenotype of the 175 individuals in 3 environments showed that the R population had large variation with the maximum coefficient of variation was 86.23% for the SPY in Yichang,and the minimum was 10.11% for LS in Wuhan.The averages SPY in the three environments were 13.93±0.43 g,7.41±0.22 g and 8.08±0.53 g,respectively;the averages NB were 6.56±0.11,6.29±0.07 and 5.02±0.09 respectively;the averages LMI were 62.59±0.73 cm,54.181±0.45 cm and 52.02±0.57 cm,respectively;the averages SNMI were 69.83±0.97,62.47±0.77 and 56.41±0.84,respectively;the averages TSN were 180.11±5.34,145.23±3.2 and 140.78±7.18,respectively;the averages LS were 5.18±0.4 cm,5.77±0.47 cm and 5.63±0.53 cm,respectively;the averages SNL were 15.42±0.22,15.54±0.25 and 14.02±0.22,respectively;the averages TSW were 3.58±0.03 g,3.23±0.03 g and 4.03±0.04 g,respectively.Correlation analysis showed the SPY has the maxmum correlations with TSN of 0.89 at the highly significant level.2.GCA analysis for the R populationGCA was estimated with the 525 hybrids in three environments.The averages GCA of the eight traits were:SPY,1.34;NB,0.19;LMI,0.74;SNMI,2.65;TSN,15.06;LS,0.14;SNS,0.38;TSW,0.04.All averages GCA for the eight taits were positive,indicating that additive effects were present to some extent.However,these traits NB,SNMI,LS,SNS and TSW performed weak positive in GCA,suggesting that the genetic gains for these traits might limited.3.Heterosis analysisAll the hybrids between the testers and individuals from R population were used to analyze heterosis for each trait in three environments.The results of the middle parental heterosis(MPH)showed that all traits of the combinations T1×R had positive advantages in three environments;only the LMI of the combinations T2×R had negative heterosis in Yichang;all the traits of the combinations T3×R had almost no heterosis in three environments.The trend of the over-standard heterosis(OSH)performance was similar to MPH.This mean that the hybrids crossed between T1 and R or T2 and R would exhibit high and strong heterosis.These results provide a good reference for the subsequent application of breeding.4.Genetic structure and diversity analysis of the R populationBrassica 60 K SNP chip was used to genotype these 175 individuals.We have analyzed genetic diversity,population structure,kinship and linkage disequilibrium(LD)of the R population.We found the average polymorphism information content(PIC)value of all SNPs was 0.292,and 76% of them have a value over 0.25,indicating that the the R population presented a good level of genetic diversity.We also found that the R population had been divided into two sub-populations,and there was no significant relationship between any two individuals.A large LD degree were detected in the R population,reaching about 2.4 Mb for the whole genome,0.8 Mb for the A genome and 4.8 Mb for the C genome,respectively.5.Selective sweeps analysisBased on the SPY related GCA performance,we have selected top 20% high GCA lines from the population to construct the next cycle population.Selective sweeps analysis was conducted between the populations on the genomic level.We identified 376 genomic regions as the selective sweeps.These genome regions covered 3.31%(21.26 Mb)of the assembled rapeseed genome.Many known QTLs were mapped in some of these selected regions.Finally,we found 96.05% of these selected regions were distributed on the C genome,indicating that the C genome had sustained stronger selection pressure in the breeding project than the A genome.