Genetic Basis Of Interspecific Heterosis Between G Hirsutum And G Barbadense Revealed By CSILs

The genetic basis of heterosis has been discussed for nearly a century but little consensus has emerged,and the important of loci with overdominant effects(ODO)is debated.One problem has been the use of whole-genome segregating populations.To access the contribution of overdominance(ODO)to heterosis in the absence of epistasis,we made each lines in introgression lines backcross with TM-1 again to abtain a set of F1 lines(ILHs),and then we carried out quantitative genetic and phenotypic analyses on the three populations between ILHs and their two parents over past five years.The main results were summarized as follows:To maximize the genetic and phenotypic diversity,we expanded our measurements on an interspecific IL population where most of the G barbadense genome is represented in the TM-1 background.A total of 17 traits were investigated and compared in our research.Heterosis was exhibited in five traits including lint yield per plant,seed-cotton yield per plant,boll number,boll weight and lint index;the ILH means were significantly higher than those of the ILs and TM-1.Other phenotypes,such as lint percentage,fruit branches,plant height,seed index,and other traits showed no heterotic effects,where the ILHs were,on average,less than their parents.Among these traits,lint yield per plant exhibited the greatest heterotic effects.In order to study the performance of heterosis in ILs systematically,we chose lint yield per plant as the phenotype that represents yield.Then we performed a correlation analysis between lint yield per plant and all other phenotypes.According to the results,five traits correlated strongly with lint yield per plant(r>0.5)and were considered as yield-related traits.Six traits were considered as intermediates(0.5>r>0.1).The other five traits were classified as non-yield(r<0.1).To test whether heterotic phenotypes for yield traits are associated with a particular mode-of-inheritance mechanism at specific loci,the entire phenotypic database for 17 traits was subjected to a QTL analysis in which each IL and ILH was compared with TM-1.At least one of the lines had a significant difference in a trait where the IL was considered as harboring a QTL at the specific insertion region.we resolved 396 quantitative trait loci(QTL)for 11 yield and non-yield traits on average in the five years.Of 396 QTL,214 were in the yield group,with 182 for the non-yield group on average.To investigate this phenomenon thoroughly,we classify each QTL into the following mode-of-inheritance categories:recessive,additive,dominant,or ODO.After classification,we compared the differences in QTL distribution between yield and non-yield groups.Remarkably,the distribution of QTL numbers shows that the group of yield traits had many more increasing ODO QTL accompanied by more decreasing recessive QTL than in the non-yield group on average.Considering the deviation brought by pleiotropism in the experimental statistics,a correlation analysis was conducted for all traits to correct for it.The number of QTL that affected highly correlated traits(where r>0.5 or r<-0.5)with the same mode of inheritance was considered as a single IL QTL.Even with this pleiotropism-corrected approach,the mode-of-inheritance mechanisms among the nonredundant QTL also showed some significant differences between the groups.In the end,we used a more stringent statistical analysis with complete pleiotropism.In this approach,if a QTL affected several traits from the same group with the same mode of inheritance,it was considered as a single QTL irrespective of the correlations among the traits.Even under these conservative assumptions,ODO of the yield traits was still 25 times higher than the non-yield traits.To imitate each QTL's mode of inheritance accurately,we used our large data set of quantitative measurements to attach a quantitative index to the mode of inheritance of each QTL,and then we could study the distribution of QTL quantitatively.The“yield”curve has a peak in the ODO domain.In contrast,most of the QTL for the non-yield group and for the decreasing yield phenotypes resided in the additive-dominant domain.In order to provide a good reference for high yield breeding,we summarized the overdominant quantitative trait loci detected in yield-related traits in five years.Totally,13 ODO QTL in lint yield per plant(LPP),11 in seed-cotton yield per plant(SPP),17 in boll number(BN),7 in boll weight(BW),14 in lint index(LI),11 in lint percentage(LP)were detected at least in one year.Among these ODO QTL,2 ODO QTL(oLPP-9-2 andoLPP-25-1)in lint yield per plan,3 ODO QTL(oSPP-2-1,oSPP-3-1 and oSPP-14-1)in seed-cotton yield per plant,4 ODO QTL(oBN-5-1,oBN-8-1,oBN-18-1 andoBN-26-1)in boll number,6 ODO QTL(oBW-9-2,oBW-16-1,oBW-18-2,oBW-25-1,oBW-26-1 and oBW-26-2)in boll weight and 1 ODO QTL(oLP-21-2)in lint percentage could be stablely detected in all five years.