Association Mapping Of Yield And Fiber Traits In Upland Cotton

Cotton is one of the foremost cash crops and the primary source of natural textile fiber in the world. The change of modern spinning technology needs lint fiber of high quality. Improvement of lint yield and fiber quality was the main objective in cotton breeding. Upland cotton is widely cultivated and account for about95percent of cotton production around the world. Thus, it is of great significance to study the genetic basis of yield and fiber traits. Previous studies showed that yield and fiber traits are quantitative traits controlled by multiple genes and environment factors and have complex genetic pattern. With the widespread applications of molecular markers in the genetic breeding, association mapping based on the linkage disequilibrium (LD) provides a new approach for exploring the genetic mechanisms of lint yield and fiber traits. In the present studies, association mapping was conducted to dissect the genetic variance loci related to yield and fiber traits, facilitating molecular improvement of cotton yield and fiber. The primary results are summarized as follows:1.323accessions of Gossypium hirsutum were evaluated for three fiber quality traits (upper half mean length, UHML; strength, STR; micronaire value, MIC) in2007-2009at three locations, which represent the three major cotton-growing areas in China. Firstly, GMDR-GPU was used to scan5600SSRs for1D and2D significant candidate markers and obtained a reduced number of317SSRs. Then, a GPU parallel computing software QTXNetwork based on mixed linear model was used to identify loci associated with three traits. The genetic model included additive, epistasis and their environment interaction. It was revealed that additive and epistasis effects were both significant to the three traits. Epistasis effects were the leading genetic effects for UHML and additive effects were the primary genetic component of STR and MIC. In addition, the total heritability of environment interaction was relatively small for the three traits, indicating that these fiber traits are relatively stable across environments. The three fiber traits were controlled by several additive loci with moderate effects and many epistasis loci with small effects. The additive loci tended to be located in the A sub-genome, and the epistasis loci were more likely within D sub-genome as well as between A sub-genome and D sub-genome. These may provide evidences that tetraploid cottons have higher lint fiber quality than ancestor-like diploid cottons.11individual loci with additive effects and14pairwise loci with epistasis effects were significantly associated with UHML (h2>0.5%and PEW-value<10-5);14individual loci with additive effects and2pairwise loci with environment-specific epistasis effects were significantly associated with STR; eight individual loci with additive effects, one individual locus with environment-specific assistive effect and14pairwise loci with epistasis effects were significantly associated with MIC. The result showed UHML was positively associated with STR (r=0.71, p<0.01) while MIC was negatively correlated with UHML and STR (r=-0.25, p<0.01and-0.24, p <0.01, respectively). And some SSR loci (or gene) were found to have genetic effect on multiple traits, which demonstrated pleiotropy is presence in fiber traits.2.39lines and178F1hybrids of Upland cotton were planted in Alar of Xinjiang province and Anyang of Henan province in2012and2013. Five fiber traits including upper half mean length (UHML), uniformity (UI), micronaire value (MIC), strength (STR) and elongation (ELG) were measured.351SSR markers used to carry out association mapping to discover the association between genotypes and phenotypes by the software QTXNentwork. A genetic model including additive (a), dominance (d), epistasis (aa, ad, da, dd) and their environment interaction (ae, de, aae, ade, dae, dde) was used to explore the genetic basis of five traits. The results showed that loci×environment interaction effects could explain47.51～80.39%of the phenotypic variance. It suggested that five traits were sensitive to environment factor. The additive-by-environment interaction effects were the primary genetic component for UHML, UI and STR, while dominance-by-environment interaction effects were dominated in MIC. For ELG, the total heritability of dominance-by-environment interaction effects and epistasis-by-environment interaction effects were both large. In total, there were10loci detected for UHML,5for UI,6for MIC,8for STR and8for ELG. The predicted genetic effect of each QTS was provided, and improving the fiber traits could be achieved by selecting the different genotype of these QTSs. The software also provided the predicted genetic effect of homozygotes (QQ, qq), heterozygote(Qq), best line, superior line, best hybrid and superior hybrid. The breeding value superior line or superior hybrid could be attained by selecting several QTSs based on the best line or best hybrid.3. We also studied the genetic mechanism of lint yield, boll weight, boll number and lint percentage for217cotton varieties. The sum heritability of genotype-by-environment interaction effects was large for those four traits and suggested that environment factor had great impact on lint yield and yield components. The total heritability of both lint yield and boll weight was mainly contributed by additive-by-environment interaction effects and dominance-by-environment interaction effects. Epistasis-by-environment interaction effects were the primary genetic effects for boll number. Additive-by-environment interaction effects were the first genetic component for lint percentage while additive and epistasis-by-environment interaction effects were not ignorable. Six loci for lint yield, eight for boll weight, three for boll number and nine for lint percentage were detected, respectively. Some QTSs with highly significant were screened and lint yield could be improved by selecting specific QTSs in particular environments. Selecting some QTSs based on the best line or best hybrid of each trait in each environment, the superior line or superior hybrid could be gained in MAS breeding.