Genome Wide Association Study Of Combining Ability And Identification Of QTLs For Important Agronomic Traits In Rice

Rice is one of the most important food crops and feeds more than half of the population in the world.High yield and good grain quality are still the most important goals in rice basic research and breeding.Application of hybrid rice is an important and reliable way,but in the practical production,many breeders still need a lot of experience to breed elite hybrid rice based on the phenotype of parental lines.The quote of combining ability makes it possible to study the potential of parental yield,which greatly accelerates the research and utilization of molecular mechanism of heterosis.On the other hand,grain shape is the intuitive performance of appearance of rice and has a direct effect on grain weight,which is one of the three main components of grain yield.In addition,rice grain shape is also a highly important quality trait that influences its market value.Therefore,it is of great significance and the hot field to explore the genetic basis of grain shape and the exploration of functional network.Over the past few years,a large number of QTL controlling grain shape have been detected by using various mapping populations but most of the cloned genes are major genes,a large number of minor QTL have been identified but still need to be further fine mapping and cloned.In this study,we constructed a NCII population and performed GWAS to yield related traits and explored the genetic basis of combining ability and suggest a way of using heterosis in breeding.We performed GWAS to GCA and SC A of 12 yield related traits and identified a number of loci potentially important for hybrid rice breeding and improvement.Besides,we performed the effect verification and fine localization of a new identified locus of grain width in the near-isogenic population.Furthermore,we Finally,we selected two F2 populations derived by Pusa and H2613S,Pusa and C815S,genotyped by RICE6K single nucleotide polymorphism array to construct high-density linkage maps in order to dissect the genetic basis of this complex trait.The backcross populations of each QTL were constructed to validate the genetic effects.The main results of the research are as follows:1.We planted a worldwide collection of 529 cultivated rice accessions and the 96 accessions with ideal plant type and medium heading date(80-120 days)were crossed with each of the 4 female parents to construct a NCII population.2.All traits were measured utilizing a high-throughput phenotyping facility(HRPF)and combined with artificial investigation.Statistical analysis showed that there were high correlations among phenotype,SCA,GCA and heterotic advantage for all traits;phenotype was always positively correlated with GCA and SCA,GCA and SCA always positively correlated with heterotic advantage,and SCA contribute more to the heterosis.GCA always negatively or not correlated with SCA,the correlations with heterosis and parent genetic distance(PGD)and observed heterozygosity(OH)were not particularly evident in our data.3.We filtered out SNPs with MAF<0.05 and missing data rates>10%,leaving 1,663,267 SNPs for further analysis.We used P?2.39 × 10-7 as the genome-wide significance threshold for parent lines and NCII population after Bonferroni correction.To analyze the genetic architecture of complex agronomic traits and their combining ability,we performed genome-wide association scans using a Q+K model and finally identified 34 significant sites.4.By comparing associations summarized in Manhattan plots,we found the peaks of parental GCA were similar to hybrid yield traits,but quite different from the parental yield traits.Using an additive model,we identified many loci in parental genome additively on hybrid yield traits,moreover,the loci of the phenotypic identification with the general combining ability had higher significance.This indicates GCA and the yield of inbred lines were genetically controlled by different sets of loci on the rice genome that are transmitted into offspring.The genetic essence of GCA was the additive effects of the parental genomes on hybrid phenotype.5.GCA could well reflect the inherited additive effects,among them,three loci contributing most to grain yield were Ghd8,GS3 and a seed setting rate QTL that we named qSSR4.Comparing the top 10 and bottom 10 parents for GCA of grain yield,we found that these superior loci were accumulated in parents with high GCA,in parents with low GCA the contrary is the case.6.An additive model is not suitable for identifying associations of SCA,in order to dissect the genetic architecture of SCA,we changed the genotype of hybrid SNPs to fit a pseudo non-additive model.Using the new genotypes,we detected many significant associations of hybrid SCA,which differed from associations of hybrid yield traits.Most SCA associations showed over-dominance and their effects on yield traits are complex.Hd3a and Hdl are known genes associated with SCA of heading date in our GWAS.Their separate effects on heading date were small,but their interaction led to extremely late flowering.A locus we named qGL12 on chromosome 12,significantly reduced grain length when it was heterozygous.As qGL12 showed no interaction with other grain length or grain width genes,this locus may represent over-dominance.To sum up,the genetic basis of SCA involves genes with non-additive epistatic or over-dominant effects,and these loci directly contributed to heterosis advantage or disadvantage.7.Comparing the top 20 and bottom 20 hybrids for SCA of grain yield,we found only a locus on chromosome 11,which we named qFN11,accumulated in high SCA combinations.Further analysis of the effects of three genotypes of each locus,we found that homozygote disadvantage coexisted with heterozygote advantage.High SCA combinations contain more advantageous genotypes and less disadvantageous genotypes than low SCA combinations.8.A large number of QTL have been discovered by using genome-wide association studies,among them GS3 and GW5 are two major QTL which had been cloned.Besides,a novel QTL affect grain width had not been reported,namely qGW1.The qGW1 is a rare allele and significant effects on grain width and length-to-width ratio.9.NIL of qGW1 was constructed by backcrossing with the recurrent parent,HD9802S,for several times followed by a self-cross.Two populations of 192 individuals were used to confirm the effect of qGW1 in Wuhan and Hainan.The results showed that there was a significant difference in grain width and length-to-width ratio between near isogenic lines,and there was no significant change in other yield related traits.10.The male parent(Pusa)had the longest grain in a collection of 529 core germplasms.The male sterile line(H2613S and C815S)were the female parent.Two F2 segregation populations were constructed and genotyped by RICE6K array to construct genetic linkage maps to dissect the genetic basis of grain shape.We phenotyped grain length,grain width,length-to-width ratio and thousand grain weight,and a total of 35 and 30 QTLs were identified in two populations in two years,respectively.Compared with traditional SSR marker map,our high density SNP linkage map detected more significant sites and more precise positions.Among them,including the cloned major gene,GW7/GL7,can be initially positioned within a range of 136 kb.11.We focused on newly found minor QTLs,designed the KASP makers and combined the molecular marker assisted selection system to construct the backcross separation populations and validated the effects of QTLs.In population 1,we validated the effects of qGW1,qGL2.2,qGW3,qGS7,qKGW8 and qGW9.In population 2,we validated the effects of qGW2,qGL2,qGS7,qGS9,qGW10 and qGL12.These results suggested that RICE6K array is a feasible and effective method for the identification of QTL,and the detected major QTL and most minor QTL are real.By further purifying the genetic background,the effects and the genetic variation explained by each QTL will elevated,it lays a solid foundation for fine mapping and map based cloning of the QTLs.