Inheritance Of Seed Traits And QTL Mapping For Protein In Upland Cotton(G. Hirsutum L)

Cotton (Gossypium spp.) is one of important economic crops in the world. The fiber which is its main product is used in the textile industry while the cottonseed, which is a byproduct is widely used in making food, animal feed, fuel and other products. Cottonseed contains many kinds of nutrients, making it a potential food-source for humans as the world population continues to rapidly grow. Cottonseed improvement has therefore become one of the major objectives in cotton breeding requiring a more detailed understanding of the genetic mechanisms in cottonseed for effective manipulation. The aims of the present study are the following:(1) Studying the inheritance of upland cotton seed traits, which include determination of genetic effects, heritability and correlation based on the embryo genome, the maternal plant genome and the plasmon. It was conducted with genetic models that were developed purposely for quantitative traits in diploid plant seed and their corresponding statistical approaches using two years data from the188inbred lines and their hybrids, all of which were cultivated within a partial diallel mating design;(2) Identifying and analyzing the QTL for protein-related traits based on embryo genome and maternal plant genome. An "immortalized F2"(IF2) population including376crosses was adopted, for this study which was conducted using the newly developed QTL mapping method that includes embryo and maternal main effects as well as their genetype×environment (GE) interaction effects on quantitative traits of dicotyledonous seeds. The main results were summarized as follows:1Calibration equations were developed with near infrared spectroscopy (NIRS) for amino acid content in cottonseeds using a total of445samples with a high variability in amino acid contents were used for developing calibration equations of amino acid content of cottonseeds. The spectral data of cotton kernel powder was processed using the first derivative mathematical treatment combined with SNV and de-trend, as well as Modified Partial Least Squares (MPLS) regression method. Chemometric models for17amino acids present in cottonseed were developed, and12of them were found to be excellent for determination of related amino acids including asparagic acid, threonine, glutamic acid, etc, with-the ratio of SD to SECV in the calibration set (RPDc) of3.735～7.132and adetermination coefficient (r2) of0.910～0.979in external validation. For these12amino acids, their values obtained by the chemical method were accurately predicted using NIRS. The RPDc of serine, methionine, tyrosine and proline ranged between2.205and2.814, and they determination coefficient (r2) were from0.800to0.830in external validation. For these4amino acids, the NIRS values were not as accurate as those from the chemical analysis, but could be used for sample screening in cotton breeding program. The equation for cystine however was not reliable as its RPDc was only1.358, which was not suitable to estimate its content in cottonseeds.2An immortal F2population (IF2) comprising of376crosses was constructed using a set of recombinant inbred lines (RIL) derived from the hybrid of HS46XMARCABUCAG8US-1-88. In this population, most cottonseed quality traits including seed index (SI), kernel index (KI), kernel percentage (KP), kernel/hull (K/H), protein content (PC), oil content (OC), gossypol content (GC) and17amino acids contents have characters of quantitative traits with normal distribution. In addition, means of most traits in some crosses of the IF2population were beyond those of the parents (HS46and MARCABUCAG8US-1-88). This population thus is suitable resource for genetic studies and breeding.3Based on phenotypic data of cottonseed in188recombinant inbred lines and their IF2population, genetic variances, heritability and correlation analysis of7important cottonseed quality traits were conducted.(1) The results of variance analysis showed:SI, KI and K/H were mainly affected by embryo additive and maternal additive effects. Secondly SI and K/H were affected by cytoplasmatic interaction; KP was mainly affected by embryo additive effect, and on a lesser note by embryo dominance and maternal additive interaction effects. OC was predominantly affected by cytoplasmic effects, and secondly by embryo additive interaction effects; The phenotypic variance of the above mentioned traits was mainly atributed to their genetypes, although environmental factors were not neglected. PC was mainly controlled by embryo dominance and maternal additive interaction effects, secondly by maternal additive and dominance main effects; GC was mainly affected by embryo dominance×environment interaction effects, also by maternal additive effect. The phenotypic variance of these two traits could be more affected by environment. The environmental factor needed to be considered in the improvement of these traits.(2) The results of heritability analysis showed:The total narrow-sense heritabilities for GC and OC were relatively high, it was thus feasible to select these traits in early generation; those for SI, K/H, KI and PC were relatively low, then these traits were not suitable for selection in early generation. Maternal general heritability was the most important for SI, KI and GC. Cytoplasmic heritability was the main parts for K/H, KP and OC; Selection advance for SI, K/H, OC, KI and GC could therefore be expected in the early generation through the combination of the performance of maternal plant and single seed selection. While for PC, it was cytoplasmic interaction heritability, and improving PC would be more efficient in the early generation by single plant selection based on the performance of maternal plant under specific environment.(3) The analysis of genetic correlation showed:The relationship between KI and GC, KI and K/H were mainly controlled by embryo additive effect, which showed a significantly negative correlation; Significantly negative maternal additive correlation were the main part between SI and K/H, SI and GC, SI and KI, SI and/KP, SI and PC, KP and PC, K/H and PC, OC and PC, KI and PC, KI and GC. The relationship between K/H and KI, K/H and GC, K/H and KP were mainly controlled by embryo additive, cytoplasmic and maternal additive effects, which showed significantly negative correlation; Significantly positive embryo additive correlation was main part between GC and OC; Significantly positive cytoplasmic correlation was main part between GC and PC. Significantly positive embryo and maternal additive correlation were main part between KP and KI. Therefore, direct selection for above pair traits could be conduceted in the early generation by single seed or plant selected. The relationship was not significant between OC and SI, OC and K/H, OC andKP, OC and KI, and it was not reliable to get high oil content in cottonseed through physical quality trait selected4QTL mapping of cottonseed protein and amino acid contents were conducted on the immortal F2population obtained from the random mating among the188recombinant inbred lines derived from the cross of two upland cotton cultivars (HS46×MARCABUCAG8US-1-88). A molecular linkage map of recombinant inbred lines constructed in our laboratory and a newly developed program called the QTL Network-CL-2.0-Seed was used for QTL analysis for protein and amino acid contents based on multiple genetic systems in two different environments. Six QTL were detected for protein content explaining58.5%of the phenotypic variation. Among them, two QTL were expressed in the embryo, While the remaining four QTL were expressed in the embryo and maternal genomes. Three of the six QTL had environmental interaction effects. Among three QTL with more than5%embryo additive heritability, two QTL were mapped on particular chromosome. This indicated these two QTL were major for improving protein content in cottonseed.QTL for sixteen amino acid including phenylalanine (phe), lysine (Lys), leucine (Leu), isoleucine (ILe), methionine (Met), threnine (Thr), valine (Val), aspartic acid (Asp), alanine (Ala), proline(Pro), serine(Ser), tyrosine(Try), histidine(His), argnine (Arg), glutamic acid (Glu) and glycine (Gly) contents were identified. The number of QTL for these traits ranged from2to8, explaining23.37%～64.32%of the phenotypic variance. Among these QTL, the range between0and2QTL were expressed in embryo genome. No QTL for Asp and Gly contents was expressed in embryo genome.Two QTL for Phe, Val, Met, Tyr and His contents were expressed in the embryo genome, respectively. Among these QTL for the sixteen aimno acid contents, the range between0and2QTL were merely expressed in maternal genome. Two for ILe and one QTL for Val, Asp and His contents were merely expressed in maternal genome, respectively. No QTL for the remaining twelve amino acid contens was merely expressed in maternal genome. Among the QTL for the sixteen amino acid contents, the range between0and6QTL were simultaneouly expressed in embryo and maternal genomes. No QTL for His content was simultaneouly expressed in embryo and maternal genomes. Six QTL for Phe content were simultaneouly expressed in the embryo and maternal genomes. The number of QTL with significant environmental interaction effect varied from1to5QTL for sixteen amino acid contents. One for Thr, Asp, Glu and Gly and Six QTL for Phe content had environmental interaction effects, respectively. The results of mapping QTL for Phe, Val and Ile content were excellent. A total of eight QTL for Phe content, six QTL for Val and Ile contents explained54.5%,64.32%and60.95%of phenotypic variance, respectively. These QTL could play an important role for the improvement of their corresponding traits.