| As a major food and model for plant genomic research, rice plays an important role in the development of agriculture and genomics. Well bridging the genotype and phenotype is an essential task to enhance modern breeding of elite varieties suitable for sustainable agriculture. Association mapping based on linkage disequilibrium, which can simultaneously discover gene loci and favourable alleles responsible for multiple traits, is a new useful tool for crop improvement. Rice has great merits as the target crop of association mapping based on the public data of genome sequences together with a wealth of genetic diversity in natural populations of germplasm. A diverse panel of rice germplasm was used in this study, which mainly include diverse landraces or varieties from a putative mini-core collection of Chinese germplasm, along with some widely used parental lines in genetic analysis or breeding (a few from abroad). The objectives were: to analyze the genetic structure and the linkage disequilibrium pattern of the population; dissecting the phenotypic variation of a series of agronomically important and yield related traits and investigating their differentiation between subpopulations; scanning the significant marker loci associated with observed traits, and detecting the favourable alleles and haplotypes. The main results are as follows:1. The population was genotyped using 218 SSR or InDel markers all over 12 rice chromosomes. A set of 52 unlinked markers from 12 chromosomes were used to analysis the population structure. Based on morphological characterization, distance-based clustering and model-based estimation of marker data, the population showed a predominant structure with two subpopulations in correspondence with indica and japonica subspecies.2. The phenotyping of heading date (HD), plant height (PH), panicle length (PL), panicle number (PN), grain number per panicle (GNP), spikelet number per panicle (SNP), spikelet density (SD), fertility percent (FRP), grain length (GL), grain width (GW), grain length to width ratio (LWR), grain shadow area (GSA), hundred grain weight (HGW) and grain weight per plant (GWP) were carried out in different locations for two years. Nest-ANOVA was conducted to disclose the phenotypic variation from experiment, accessions in subspecies and subspecies for all investigated traits. Highly significant variation between indica and japonica subspecies was shown in PL, PN, GL, GW, LWR, HGW and GWP. The effect for population structure of all traits (i.e. the variation explained by subspecies) varied from 0.79% to 73.67%. The broad sense heritabilities for all traits varied from 33.97% to 97.99%.3. Along chromosome 7, the average LD (r~2) was estimated among markers in 2-Mb windows. The LD level in whole population had an uneven distribution with peaks on the chromosomal regions of 18～21Mb, 21～24Mb, and larger than 28Mb. Variance of LD was observed in both subpopulations along the same chromosome, but with much smaller ranges. The estimation of LD in 2Mb windows varied along chromosome 7 and showed parallel changes with inter-subspecies differentiation of marker loci (Fst).4. Based on the mixed linear model (MLM) considering population structure and family relatedness (i.e. the (Q+K) model), one to forty-six associated markers (P<0.01) per trait per experiment were scanned out on the whole genome. Most significant loci were repeated for the data from both field experiments. In most cases, markers significantly associated with the measured traits located in chromosomal regions where genes or QTL have been previously reported.5. Marker-based allelic and haplotypic effects were shown for the high-significantly associated markers. Based on the multiple comparisons by the LSD test, the difference among means of each trait of accessions group by alleles and haplotypes were investigated, as well as the favorable alleles and haplotypes were identified. The rice accessions can be roughly categorized according to their origin places, belonging to indica or japonica subspecies, landraces or modern cultivars and so on. The accessions bearing different marker alleles had uneven distribution among such kind of categories. This information together with association results can be applied in breeding programs.
|
PDF Full Text Request