| Popis: |
As interest in new sweetpotato (Ipomoea batatas (L.) Lam.) markets such as starch feedstocks, anthocyanin production, chips and French fries, and other processed food products have increased, breeders have begun developing sweetpotatoes with unique compositions. This dissertation describes a three-pronged strategy to understand the genetic control of sweetpotato composition and modify it using near infrared spectroscopy (NIRS), quantitative inheritance studies, molecular markers, and exotic germplasm. In our first experiment, a 5 × 5 NCII crossing block with 25 full-sib families was designed to study the inheritance of sweetpotato starch and anthocyanin content. Linear regression modeling was used to determine the effect of general combining ability (GCA) and specific combining ability (SCA) on dry matter, total monomeric anthocyanin (TMA) concentration, fresh yield, and total dry matter and anthocyanin yield. All five traits were moderately to highly heritable with significant general combining abilities. Yield and dry matter yield had significant specific combining abilities and significant differences among parents were discovered for all traits. Yield, dry matter, dry matter yield, and TMA yield were significantly impacted by spatial gradients within the field, but TMA concentration was not. Phenotypic and genotypic correlations among traits indicated that many traits of interest shared either genotypic and/or phenotypic correlations. In our second experiment, a nested crossing block was used to estimate the heritability of sweetpotato yield and storage root composition traits in a population incorporating exotic germplasm obtained from the US sweetpotato germplasm repository and a core set of elite US sweetpotato lines crossed in a modification of the NCI design. Yield traits were recorded in the field and biochemical composition was phenotyped using NIRS. Heritability was measured on a half-sib family basis and a full-sib family basis to allow comparison between the commonly used polycross nurseries and paired crossing blocks. Parent offspring regression, which has been commonly used by sweetpotato breeders, was also used to provide another heritability estimate. Starch and sugar contents had relatively high heritabilities on both a GCA (h2 > 0.32) and SCA basis (h2 > 0.77). Yield traits had low heritability on a GCA basis (h2 < 0.16), but moderate heritability on an SCA basis (h2 = 0.21 – 0.51). Heritability trends suggested that polycross nurseries could be effective for modifying sweetpotato composition, while paired crosses would be more effective for the modification of sweetpotato yield. Based on the performance of a wide range of crosses between exotic and heirloom varieties, we hypothesize that the global sweetpotato germplasm base contains many useful alleles for continued sweetpotato improvement. Our final study involved a previously described sweetpotato quantitative trait loci (QTL) mapping population developed from a Tanzania x Beauregard cross (Cervantes-Flores 2006). This population was phenotyped using NIRS to identify QTL for sugar and starch content. In Beauregard, six QTL were associated with decreased starch and dry matter content and eight QTL were associated with increased sugar content. One QTL in Beauregard was associated with decreased yield. In Tanzania, two QTL were associated with increased starch and two QTL were associated with decreased starch; there were also two QTL associated with decreased sugars and one associated with increased sugar content; one QTL was associated with decreased culls. In most cases, newly identified QTL co-locate with those previously described. Collectively, this research represents a significant effort in sweetpotato to merge molecular markers with NIRS phenotyping, and it has opened the doorway to further developments that merge these two new technologies for sweetpotato improvement. |
