| The new strategy of crop molecular breeding, proposed by Zhikang Li, was to combine the phenotype selection with AB-QTL and find the change of allelic frequency due to directional selection with molecular markers, which had been conducted in"International Rice Molecular Breeding Program". A lot of introgressed lines (ILs) with favorable genes were selected from advanced backcross populations. The methods and techniques successfully used in rice molecular breeding were extended to main crops e.g. soybean, maize and wheat etc. However, there existed large differences in the methods of target trait screening, ILs genotyping and phenotyping used by various institutes, which would influence the effects of hitchhiking. And the relationship of selection intensity with hitchhiking effects stayed unclear. Therefore, it was necessary to initiate the theoretical research on QTL mapping with selected ILs.In this study, Shuhui 527 and Minghui 86, two elite indica restorers in China were used as recurrent parents and Milagrosa from Philippines as donors, we developed two BC2F2 populations and the derived BC2F2:3 populations.There were 85 and 96 SSR makers evenly distributed on 12 chromosomes applied to genotype the advanced backcross populations of Shuhui527/Milagrosa and Minghui86/Milagrosa respectively. The genetic composition and marker distortion of two populations with two continuous generations were analyzed. The QTL identification on grian length, grain width, the ratio between grain length and width, and thousand-grain weight were conducted with BC2F2 populations of Shuhui 527/ Milagrosa and Minghui 86/Milagrosa. The grain length was used as an example to simulate the extreme positive and negative selection under 4 selection tensities, from which comparative QTL mapping was done by one-way ANOVA and Chi-square test based on genetic hitchhiking. The main results were as follows:1. Genetic composition of BC2F2 and BC2F2:3In Shuhui 527/Milagrosa BC2F2 population, there were 65.91 loci per plant averagely had recurrent parent homozygotes among which 58.85 loci didn't change their genotypes in BC2F2:3, 4.68 loci became heterozygotes and 1.72 loci evolved into donor homozygote. There were 5.35 loci showed donor homozygotes per plant averagely in BC2F2, and 2.71 loci didn't change in BC2F2:3, 0.79 loci evolved heterozygotes, 1.79 loci became recurrent parent homozygotes. In Minghui 86/Milagrosa BC2F2 population, there were 73.02 loci per plant averagely showed recurrent parent homozygotes, and 60.50 loci didn't change their genotypes in BC2F2:3, 7.11 loci evolved into heterozygotes, 3.57 loci changed into donor homozygotes. There were 5.66 loci per plant averagely showed donor homozygotes in BC2F2 population, and 1.85 loci didn't change their genotypes in BC2F2:3, 0.97 loci changec into heterozygotes, 2.53 loci evolved into recurrent parent homozygotes.2. Marker segregation distortion of BC2F2 and BC2F2:3Chi-square test was conducted in BC2F2 and BC2F2:3 of Shuhui 527/ Milagrosa and Minghui 86/ Milagrosa. We found five hot spot regions of segregation distortion in Shuhui527/ Milagrosa BC2F2 population, located respectively on RM128～RM297 region of chromosome 1, RM438～RM341 region of chromosome 2, RM282～RM487 region near the centromere of chromosome 3, RM125～RM248 region almost covering the whole chromosome 7 and RM434～RM215 region of chromosome 9. In Minghui86/Milagrosa BC2F2 population, we identified three hot spot regions of segregation distortion, located respectively on RM9～RM5536 region of chromosome 1, RM282～RM319 near the centromer of chromosome 3 and RM536～RM21 region of chromosome 11. The segregation distortion region on chromosome 1 and 3 overloped each other in the two populations. A total of 27 among 41 segregation distortion loci detected in BC2F2 were same as those detected in BC2F2:3 of Shuhui527/ Milagrosa population. There were 25 among 63 segregation distortion loci detected in BC2F2 same as those identified in BC2F2:3 of Minghui86/Milagrosa. Hot spot regions of segregation distortion highly overlapped in two populations.3. The analysis of QTL mapping on rice grain traits in BC2F2 random populationThe QTL identification of grian length, grain width, the ratio grain length to width and thousand grain weight was conducted in Shuhui 527/Milagrosa and Minghui86/Milagrosa BC2F2 population by one-way ANOVA (for single QTL)In Shuhui 527/ Milagrosa BC2F2 population, a total of 30 QTLs for grain length, grain width, ratio of grain length to width and thousand grain weight detected by one-way ANOVA (P<0.01), among which 3 pleiotropic loci could be found. The 5 QTLs underlying the ratio of grain length to width could be identified as well in grain length due to highly close relationship between them. The results in present study indicated that QGl3.3 on chromosome 3 was a major QTL controlling grain length, ratio of grain length to width and thousand grain weight, which explained 29.37, 26.15, and 17.15% of phenotypic variation of the three traits and showed large additive effects (positive allele from Shuhui 527) and negative overdominant effects. QTgw8 was also a major QTL underlying grain width with 21.47% contribution to phenotypic variation and a minor QTL controlling thousand grain weight with 5.16% contribution to phenotypic variation. It had large additive effects (positive allele from Shuhui 527) and positive partial dominant effects in both traits. No major QTL was detected in Minghui 86/ Milagrosa BC2F2 population though a total of 31 minor QTLs were mapped.4. The simulations of extreme selections on rice grain length of BC2F2 random populations and the precision and power of QTL mapping with selected ILsThe simulated extreme selections under bi-directions and 4 selection intensities were conducted on the random BC2F2 populations according to the phenotype of grain length. Positive and negative directional selection got eight selected subpopulations with ten, twenty, thirty and fifty individuals, respectively, and four bi-directional populations with twenty, fourty, sixty and one hundred individuals were obtained by merging subpopulations throught corresponding positive and negative directional selection. With the random populations as controls, phenotypic difference, donor allele frequency from these extreme selection populations were compared. Meanwhile, the QTL detection of grian length was executed using one-way ANOVA and Chi- square test.No obvious difference of phenotypic variation range was observed in the positive and negative directional selection subpopulations from two random BC2F2 populations, except the mean which showed a bigger difference. The mean and range of divergent selection populations were close to the those of original populations, but the standard deviation of divergent selection populations were much higher.The maximum donor allelic frequency of negative extreme selection populations with ten, twenty and thirty individuals was two times as large as the original random population of Shuhui 527/Milagrosa. The maximum donor allelic frequency of positive extreme selection populations with ten and twenty individuals was one and half times as large as the original population of Minghui 86/ Milagrosa. The maximum Chi-square value decreased when increasing sample sizes.The comparative QTL detection by one-way ANOVA and Chi-square test showed that one-way ANOVA was an effective method for bilateral selection populations, the precision and power of QTL mapping increased with the sample size of selected populations. Chi-square test was more effective in QTL detection for directional selection populations, and the the precision and power of QTL mapping increased also with the sample size of selected populations, though some differences could be observed in various genetic bakcgrounds. The genetic effects of hitchhiking in different selection directions varied with genetic backgrounds as well, e.g. they were higher in negative than positive selection under Shuhui 527 background and opposite pattern was found under Minghui 86 background. |
PDF Full Text Request