Doctor of Philosophy (PhD)


Plant, Environmental Management and Soil Sciences

Document Type



Cotton (Gossypium spp) is an important world crop. Although great improvements have been achieved through traditional breeding methods, cotton breeders are facing many problems, i.e., narrow genetic base, inability to use alien genes and difficulty in breaking gene linkages. Genetic transformations and quantitative trait loci (QTL) analyses are main tools used by breeders to overcome these problems. In this dissertation, an optimized cotton regeneration system from shoot apices was developed. The regeneration rate was increased to 85% by combining rooting induction, indole acetic acid (IAA) shock and graft techniques. The regeneration system is genotype-independent and the whole process takes 12 to 16 weeks. Transgenic cotton plants were obtained via Agrobacterium-mediated transformation using shoot apices as explants. Transformation rates were 0.67% and 1.01% for LBA 4404 with β-glucuronidase (GUS) gene and EHA 105 with Bar gene, respectively. Putative transgenic plants were confirmed by leaf GUS assay, kanamycin or herbicide (Liberty) leaf test, polymerase chain reaction (PCR) and southern blot analysis. Out of 151 polymorphic markers, 53 amplified fragment-length polymorphism (AFLP) markers were assigned to individual chromosomes or chromosome arms by using a set of aneuploid genetic stock. In the QTL analysis of cotton yield and yield components was conducted on an F2:3 population derived from the intraspecific cross. A previously developed linkage map was used based on same population covering 1733.2 cM (37.7%) cotton genome (4700 cM). A total of 47 markers associated with yield and yield component traits were detected. Nine and seven QTL detected by interval mapping (IM) and composite interval mapping (CIM) methods, respectively, four of which were detected by both methods. For lint yield, two main QTL, explaining 27% of variation, were detected via CIM method. No QTL was detected for bolls per plant by IM method and one QTL explaining 8.56% variation was detected by CIM method. For number of fibers per seed, 23.7% of variation was explained by two main QTL detected by both IM and CIM methods. For mean weight per fiber, two QTL were detected via CIM. No QTL was detected for seed number per boll via either method.



Document Availability at the Time of Submission

Release the entire work immediately for access worldwide.

Committee Chair

Gerald O. Myers