The Development Of NIRS Model For The Determination Of Oil Content And Fatty Acid Composition And The Mapping Of QTL For These Quality Traits In Cottonseed

Cotton is a major cash crop which is linked to the livelihood of a large number of people. It is an important raw material for the textile industry and plays a significant role in the global economy commanding a large domestic and international demand for many economies in Asia, America and Africa. Cottonseed is rich in protein and other important materials and has become a major source of oil and animal feed worldwide. It is next in importance to the cotton fiber which is extensively used in the textile industry. The increasing demands from the textile industry over the years have compelled cotton breeders to focus their efforts on improving fiber quantity and quality. However there has also been a gradual acceptance of cotton by-products such as the oil for domestic and industrial consumption, and the meal for use as animal feed which has began to encourage breeders to research into possible ways of improving the quality of cottonseed products. Cotton has a complex genetic base and complicated genetic mechanisms which makes progress in its improvement based on conventional breeding methods very slow. The development of molecular biology and molecular genetics makes it possible to dissect the composite genetic factors that control the quantitative traits in cotton into a single genetic factor making it possible to manipulate quantitative traits by tracing the related molecular markers and quantitative trait loci (QTLs), which may control the desired traits through marker assisted selection (MAS). Cottonseed quality traits such as oil and fatty acid contents are quantitative traits that vary continously and can be studied using quantitative genetic methods. Before traits are identified, their contents in the seeds have to be determined. In large scale breeding programs however, laboratory methods of oil and fatty acid determinations are limited to few samples and the procedures are expensive and time-consuming. The use of cheaper, faster and less destructive methods that can be applied to a larger quantity of samples are becoming more acceptable to breeders. Near infrared reflectance spectroscopy (NIRS) is one of such methods that provide the same accuracy as in the laboratory determinations.Fatty acid content in cotton is a complex quantitative trait and its genetic mechanisms would be closely related to QTLs located on the chromosomes of the diploid embryo and diploid maternal genomes. Genetic effects on inheritance of fatty acid content and quality has been widely studied in many crops. Here is the first report on QTL mapping of cottonseed fatty acid content based on a new genetic map using recombinant inbred lines (RILs) derived from the hybrid of two parents, HS46and MARCABUCAG8US-1-88. Since the quality traits of seed grown on its maternal plant could be simultaneously controlled by different genetic systems, the expression of QTLs in the diploid embryo genome and diploid maternal plant genome might be different. In this situation, these QTLs located on chromosomes in embryo and maternal plant genomes could not be easily distinguished.The present research is made up of two major parts. The first part involves the application of NIRS technology in the determination of oil and fatty (palmitic, oleic, linoleic, myristic, palmitoleic, stearic, linolenic and1-eicosenic) acid contents in cottonseed kernel using hybrid and breeding materials as well as materials from the cotton germplasm. The second part involves the identification and analysis of QTLs that control the oil and fatty acid contents respectively in an intra-specific RIL population obtained from a cross of two upland cotton parents, MRCABUCAG8US-1-88and HS46; using a relatively high density genetic linkage map constructed based on the RIL population that has three kinds of molecular markers (SSR, SRAP and RAPD) covering29linkage groups.An IF2population, constructed based on the RIL population was established and its embryo additive, embryo dominance and maternal additive effects were simultaneously studied. The contents of oil, palmitic acid, oleic acid and linoleic acid were chosen for the QTL studies because it was possible to develop robust NIRS models that could provide reliable predictions for these traits in cottonseed kernel.The main results were as follows:1. NIRS determinations for oil, palmitic acid, oleic acid and linoleic acid contents had high correlations with their respective laboratory determination methods. Using the modified partial least squares (MPLS) method, robust calibration models of high accuracy were developed for oil and fatty acid contents that could be used for the future determination of similar traits in other cottonseed samples. The equation for total oil content showed the highest R2(0.996) and1-VR (0.995) with low SEC (0.197) and SECV (0.233) and is considered to be the best equation with the highest RPD。value (14.06). This was followed by the equation for linoleic acid with high R2(6.186) and1-VR (0.974); low SEC (0.220) and SECV (0.387). Oleic acid also had a relatively high correlation with R2(0.891) and1-VR (0.817), low SEC (0.462) and SECV (0.602) and a relatively high RPD。(2.329). Myristic acid had an R2of0.778, a1-VR of0.738, an SEC of0.049, an SECV of0.054and an acceptable RPD。of1.937. Palmitoleic, stearic, linolenic and1-eicosenoic acids had very low R2(0.481,0.485,0.233and0.526respectively) and1-VR (0.409,0.453,0.112and0.501respectively), and although they also had low SEC (0.027,0.129,0.041and0.012respectively) and SECV (0.029,0.133,0.044and0.013respectively), their RPD。(1.305,1.346,1.062and1.391respectively) were below the limit for reliable predictability. In external validation, high coefficients of determination were obtained for the equations for oil content (r2=0.993) and linoleic acid (r2=0.963) whiles those for myristic, palmitic and oleic acids (r2=0.753,0.777and0.795respectively) were relatively high. The lowest values were obtained for palmitoleic, stearic, linolenic, and1-eicosenoic acids (r2=0.516,0.472,0.101and0.224respectively). The regression plots for oil content showed a high accuracy in the estimation of this trait with an r2of0.993and an SEP of0.273. This was followed by the regression plot for linoleic acid (r2=0.963, SEP=0.470). There was however lesser accuracy in the prediction of the remaining fatty acids including myristic acid (r2=0.753) and palmitic acid (r2=0.777). According to the guidelines for the interpretation of RPDv which are similar to those for RPDc, the equations for total oil content and linoleic acid are good enough to be used for quality assurance and research applications as well sample screening for breeding programs. Those for myristic, palmitic and oleic acids can be used only for screening whiles those for stearic, palmitoleic and1-eicosenoic acids are unusable for now. Although some traits did not provide the desired results, the use of NIRS methods enabled the determination with a larger quantity of samples thus enhancing the variability in the traits of interest.2. For the identification of quantitative trait loci controlling oil content in cottonseed kernel, which simultaneously related to the diploid embryo nuclear genome and diploid maternal plant nuclear genome, twelve samples each were used for both parents (MRCABUCAG8US-1-88and HS46) while361samples were used for the immortal F2generation (IF2). There were visible variations in the oil content among the parents and the IF2population as indicated by the oil content values. Among the QTLs detected were qOC-18-1between markers JESPR204(c18) and CIR099(c18) in the region of2.91cM on chromosome18; qOC-LGll between markers NAU979and NAU1162in the region of4.56cM on linkage group11; qOC-18-2located between markers NAU1184b and BNL3479(c18) in the region of11.50cM on chromosome18; and qOC-22between markers CIR253and JESPR65in the region of6.10cM on chromosome22. With respect to oil content, significant (P<0.01) embryo additive effect (aem) for qOC-18-1and qOC-22; as well as significant (P<0.01) embryo dominance effect (dem) for qOC-LG11and qOC-18-2were observed. The maternal additive effect (am) was significant (P<0.01) for qOC-18-2and a less-significant (P<0.05) for qOC-LG11. qOC-18-2had the largest embryo additive effect (aem=0.4572**). qOC-LG11had the largest and only significant embryo dominance effect (dem=-0.3660**). qOC-18-2had the largest maternal additive effect (a"’=-0.307**) followed by qOC-LG11(am=0.1579*). Maternal additive effects for qOC-18-1(am=-0.0320) and qOC-22(am=0.0112) were not significant. With regards to the impacts of the genetic effects of QTLs from different genomes, the negative aem of qOC-18-1and qOC-LG11could decrease the oil content whiles the positive aem of qOC-18-2and qOC-22could increase it. The negative dtm of qOC-LG11could decrease the oil content in cottonseed kernel. The positive am of qOC-LG11could increase oil content while the negative am of qOC-18-2could reduce the oil content in the cottonseed kernel.3. Six QTLs associated with fatty acids in cottonseed, were subsequently mapped on chromosomes5and linkage groups:LG1, LG7, LG9, LG10and LG14. Among the QTLs detected, two QTLs (qC16-0-5-15between markers BNL3992(c5) and TMB1667in the region of15cM on chromosome5, and qC16-0-LG14-1between markers RAPD-G17-270and RAPD-G17-200in the region of1cM on LG11) were found for palmitic acid. For oleic acid, only one QT(qC18-1-LG9-1) was located between markers RAPD-G4-580and RAPD-I11-380in the region of1cM on LG9whereas, three QTLs (qC18-2-LG10-8between markers NAU500a and NAU5171a in the region of8cM on LG10; qC18-2-LG1-21between markers NAU5498b and NAU3426in the region of25cM on LG1, and qC18-2-LG7-2between markers TMB179and TMC10in the region of2cM on LG7) were found in present experiment. All the fatty acid related QTLs except for qC18-1LG9-1(C18-1) and qC18-2-LG7-2(C18-2) had significant (P<0.01) maternal additive effect (am). Two of them (qC16-0-5-15for C16-0and qC18-2-LG1-21for C18-2) had significant (P<0.01) embryo dominance effect (<dem) whiles qC,6-0-5-15for C16-0; qC18-1-LG9-1for C18-1and qC18-2-LG10-8, qC18-2-LG1-21and qC18-2-LG7-2for C18-2had significant (P<0.01) embryo additive effect (dem).With regards to the impacts of the genetic effects of QTLs from different genetic systems, the embryo dominance effect (dem) and negative maternal additive effect (am)of qC16-0-5-15could decrease the palmitic acid content whiles the positive embryo additive effect (aem) of qC16-0-5-15and the maternal additive effect (am) of qC16-0-LG14-1could increase the content of palmitic acid. For oleic acid the negative embryo additive effect (aem) of qC18-1-LG9-1could decrease its content. For linoleic acid, the positive embryo additive effect (aem) of qC18-2-LG10-8and qC18-2-LG7-2could increase its content while the negative maternal additive effect (am) of qC18-2-LG10-8as well as the embryo additive effect (aem) of qC18-2-LG1-21and the embryo dominance effect (dem) of qC18-2-LG7-2could decrease its content. These results indicated that the additive effects either from embryo or maternal plant genome were very important for fatty acid content and the embryo dominance effect was also vital.