Study On Population Genetic Structure And QTL Identification Using Selected Introgressed Lines In Rice

In order to solve the problems of traditional QTL mapping and crop germplasm application, Zhikang Li proposed that elite widely grown varieties as recurrent parents were used for developing advanced backcross populations to combine phenotype selection with AB-QTL, which can trace donor allele flowing by molecular markers, gain the change of allelic frequency due to directional selection and identify QTL underlying target traits. However, the genetic diversity and relationship between recurrent parents and donors from germplasm showed huge differences. The genetic structure of original backcross populations developed by various institutes usually differed. All of these may influence the selection gain and genetic effects of hitchhiking and furtherly influence the precision and power of QTL mapping of selected introgressed lines (ILs) based on the hitchhiking theory. According to the information of the genetic diversity and relationship between 55 rice varieties (accessions), Shuhui 527 and Minghui 86, widely used elite indica restorers, were applied as recurrent parents,2004, IRAT352, and Milagrose as donors for developing six random BC2F2 populations. The evaluating on target traits and genotyping by molecular markers were carried out in the developed populations. The influence of genetic structure of original populations on the precision and power of QTL identification in selected ILs could be clarified by comparative QTL mapping on target traits with both original populations and selected ILs. The main results gained were as follows:1. A total of 53 SSR markers were used to analyze the genetic diversity and relationship among 55 rice germplasms for making reference to cross combinations and improvement of parents. In present study, 267 allelic variations were detected, and the average allelic variation of 53 loci is 5.04, ranged from 4 to 7. The average polymorphism index content (PIC) of SSR markers is 0.624, ranged from 0.287 to 0.786. All germplasms could be divided into indica and japonica rices, the genetic comparability coefficient between them varied from 0.588 to 0.996. The comparability coefficient between Minghui 86 and 53 donors is from 0.655 to 0.850, it is from 0.640 to 0.873 between Shuhui 527 and the donors. The results indicated that detection of SSR polymorphisms was one of the most efficient and accurate measures to study the genetic differences among rice varieties, which was helpful to the discovery of favorable genes and their utilization in advanced-backcross introgression populations.2. A total of 14 and 16 segregation distortion hot regions (SDR) were detected in the six populations under Shuhui 527 and Minghui 86 background respectively. And two more SDRs were found on Chromosome 2 and 4 in the populations of Minghui 86 background. Most of SDRs were highly overlapped between two genetic backgrounds and deviated to either donor homozygotes or heterozygotes except those on chromosome 5 and 12. It was believed that segregation distortion usually arose with gametic selection genes or sterility genes. A novel SDR on the region between 45.6~99.4 cM of chromosome 9 was identified in all of the six BC2F2 populations, which resulted in genotypic deviation to donor homozygotes or heterozygotes. No gametic selection gene or sterility gene around this region was documented.3. According to 1000-grain weight of original populations, the best or the worst 20 plants were selected to construct 3 sets of selected introgressed subpopulations——TOP20 , LOW20 and bidirectional selective populations (40 plants). One-way ANOVA and chi-square test were used to map QTL by selected ILs which were compared with the QTLs detected by the original populations. Two expected values were used in chi-square test according to the theorical segregation ratio in BC2F2 populations and practical segregation ratio (corrected) gained in the original BC2F2 populations. The results of chi-suqare test indicated that the precision and power of QTL detection were much higher using the expected segregation ratio after correction than before correction, for QTLs detected using the uncorrected expected segregation ratio were usually caused by segregation distortion. Comparative mapping between selected ILs and original populations showed that the precision and power of QTL detections by chi-square test were much higher in ILs subpopulations derived from single directional selection (positive or negative) than ILs subpopulations derived from bi-directional selection. However, one-way ANOVA had much higher precision and power of QTL detection in ILs subpopulations derived from bi-directional selection than ILs subpopulations derived from single directional selection (positive or negative).