Using Modern Molecular and Genomic Tools to Evaluate Potentially Overlooked Genetic Variation for Fiber Quality in Obsolete US Improved Cotton Cultivars

Improvements in fiber quality phenotyping methods such as High Volume Instrumentation (HVI) and Advanced Fiber Information System (AFIS), have increased breeders ability to detect superior fiber quality; however, this also suggests that potential sources of superior fiber quality prior to the use of HVI and AFIS may have been overlooked. The purpose of this dissertation is to explore gains made in fiber quality according to HVI and AFIS, the stability of these fiber traits, and the feasibility of using genomic prediction to tap into potentially unexploited sources of variation for fiber quality traits. Genetic gains for HVI and AFIS fiber traits were analyzed using a population of 63 cultivars from the obsolete US improved cotton cultivar collection that represents the past 100 years of breeding efforts in the. All HVI and AFIS traits evaluated made statistically significant gains except for Length by number coefficient of variation, micronaire, and fineness. The statistically significant percent gains per year ranged from 0.267% to 0.025%. Many traits AFIS traits showed gains even though direct selection pressure was likely not applied for these traits, so it is inferred that these gains arose through correlations indicating genetic variation for traits unexploited. HVI traits and AFIS traits evaluated were highly stable across irrigated and dry environments tested in Corpus Christi and Weslaco, TX in a population consisting of germplasm from obsolete US improved cotton cultivar collection and germplasm with superior fiber quality developed by the Texas A&M Cotton Improvement Laboratory. It was found that selection for these traits in any environment would result in a similar list of genotypes. Genomic prediction was performed using a population of consisting of germplasm from the obsolete US cotton cultivars collection and germplasm developed for superior fiber quality from the Texas A&M Cotton Improvement Laboratory. Prediction accuracies within the obsolete US cotton cultivars ranged from 0.24-0.56 for HVI and AFIS traits, and variation explained was less than previously reported heritabilities. Prediction accuracy for yarn quality was determined using a selection index created from HVI and AFIS parameters and correlated to yarn work-to-break. Accuracy was determined to be 0.36