【作者】 杨代刚；

【导师】 喻树迅；

【作者基本信息】 中国农业科学院 ， 作物遗传育种， 2011， 博士

【摘要】 棉花是我国也是全球最重要的纤维作物,以陆地棉为主。研究陆地棉品种双列杂交的遗传效应及表达谱,为选择骨干亲本、预测杂种优势及分子设计育种提供理论依据。本文以来源不同的10个陆地棉品种为亲本,按双列杂交配制45个正交F₁及其45个F₂,2008～2010年在安徽望江、河南安阳和新疆阿克苏进行同步试验,调查了8个农艺、经济性状,检测了5项纤维品质指标,2008年取盛花期幼蕾进行基因芯片分析。以ADAA与环境互作的遗传模型和条件分析方法及QTLNetwork V3.0新型软件进行数据分析。主要结果如下:1、8个农艺、经济性状的遗传主效应及互作效应不同程度起作用。单株铃数、单铃重、衣分、子指、皮棉产量的加加上位性效应和衣分、子指的加加上位性×环境互作效应正向极显著。单铃重、衣分、子指的h_G²较高,早代选择和异地选择有效;株高、果枝数、单株铃数受环境影响较大,直接选择不大可靠。果枝数、单株铃数、子棉产量和皮棉产量H_GE²较高（31.4%～51.8%）,在特定环境可利用杂种优势。以9018、中棉所41、sGK958和2028为亲本可同步改良2～3个产量组分（特别是衣分）而显著提高后代皮棉产量。2028的衣分和子指的加性效应和加加上位性效应、sGK958的衣分和子指的加性效应及其衣分的加加上位性效应均正向极显著,育种价值较大。以3392154-55为亲本可改良后代铃重而提高皮棉产量。2、纤维品质的各遗传效应及其与环境的互作效应差异较大。比强度主要受加性效应控制,纯系育种是遗传改良的有效途径。上半部平均长度的h_G²和HG2较高,可在早代选择,特定环境也可利用杂种优势。整齐度指数、伸长率和马克隆值V_e / V_P超过40%,直接选择效果较差。新陆中9号和sGK958是优质亲本,育高产优质品种首选sGK958作亲本。3、遗传相关分析表明,提高衣分改良皮棉产量效果好,不同环境选择有效;但选高产又优质的品种,衣分不宜过高。子指可间接选择纤维品质。特定环境中,提高单铃重可提高皮棉产量和比强度;选植株高的品种,也可间接改良皮棉产量。选纤维长、比强高、马值适中的品系和优势组合是可能的。4、主效应杂种优势分析表明,子棉、皮棉产量F₁和F₂的平均优势和超亲优势均较大,纤维品质平均优势较小,超亲优势均为负值。7个F₁的皮棉产量和纤维品质较协调。5、对皮棉产量的贡献率分析结果,株高、果枝数、单株成铃数、单铃重和衣分的CR_D（C→T）正向极显著且较大（57.70%～86.80%）,选显性效应高的组合,可提高其皮棉产量。衣分的CR_{AE（C→T）}和CR_{DE（C→T）}正向极显著且较高（35.91%～36.43%）,在特定环境,选高衣分可显著提高选系的皮棉产量;选衣分显性效应高的组合,其皮棉产量会提高。6、取盛花期幼蕾,利用基因芯片技术进行表达谱统计分析。检测到73个与皮棉产量显著相关的数量性状转录子（QTTs）,其中10个遗传率大于3%。与皮棉产量对应的加性方差小,显性方差大;与产量构成因素对应的加性方差较大,显性方差较小。9018和sGK958的基因对皮棉产量的加性效应贡献值较高,可用于高产育种。相同基因对不同亲本的加性效应贡献值不尽相同。基因对组合皮棉产量显性效应总预测值的贡献多数较高,对其显性效应贡献值具有显著影响的QTT不同,对其皮棉产量及其构成因素有影响的QTT不同,说明其作用机制也不同。更多还原

【Abstract】 Cotton is the most important fiber crop in China and the world, and upland cotton accounts for more than 95% cotton production. Yangtze River Basin, Yellow River Basin and Xinjiang are the three major cotton growing regions in China. Study of the genetic effects and the expression profile of upland cotton cultivars from the three main growing areas should be helpful for selection for the backbone parents, heterosis prediction and molecular breeding in cotton.In this study, 10 representative upland cotton cultivars from China’s three major cotton-growing regions, using a diallel mating design method, were used to produce 45 F₁ hybrids and their 45 F₂ generations. Field experiments were carried out in 2008, 2009 and 2010, in three typical ecological regions for cotton production in China, including Wangjiang in Anhui Province, Anyang in Henan Province and Aksu in Xinjiang Uyghur Autonomous Region. Eight morphological and yield traits were investigated and five fiber quality characters were surveyed. Gene microarray analysis was conducted using flowering buds at full bloom stage in 2008. Statistical analysis was conducted using the genetic model including additive-dominance-epitasis and its interaction effect with environment, condition analysis method, QTLNetwork V3.0. The main results are as follows:（1） For the eight morphological and yield traits, the main genetic effects and their interaction with environment play a role at varying degrees. Additive-by-additive epistatic effects on boll number per plant, boll weight, lint percentage, seed index and lint yield, interaction between additive-additive epistatic effect and environment on lint percentage and seed index were positively significant. General heritability in the narrow sense （h_G² ） for boll weight, lint percentage and seed index were high, indicating that early generation selection was effective. Plant height , number of fruiting branches and boll number per plant were greatly influenced by environment, indicating that direct selection is less reliable. Interaction heritability in the broad sense (H_GE² ) for the number of fruiting branches , boll number , seed-cotton yield and lint yield were relatively large （31.4% 51.8%）, indicating that heterosis can be used in specific environments. Parents 9018, CCRI41, sGK958 and 2028 can be used as hybrid parents to concurrently improve two or three yield components （especially lint percentage） , leading to significantly increase lint yield. The additive effect and additive-addtive epistasis on lint percentage and seed index in parent 2028 was large and positively significant, and should be used in selection breeding. Using 3392154-55 as parent was expected to effectively improve boll weight and thus increase lint yield.（2） For the fiber quality traits, both the main genetic effects and their interaction with the environment play a role, but the difference was larger. Fiber strength was mainly affected by additive effects, which can be genetically improved by conventional breeding. h_G² and general heritability in the broad sense （H_G² ） of upper-half-mean length were large, and early generation selection can also be conducted and heterosis can be utilized in specific environment. V_e / V_P of uniformity index, elongation and micronaire value were above 40%, which means direct selection may be less effective. Xinluzhong 9 and sGK958 were two elite parents with high fiber quality, but sGK958 was good at both lint yield and fiber quality.（3） Genetic correlation analysis showed that lint yield can be improved by increasing lint percentage, and selection was effective in different environments. To breed cultivars with high yield and good fiber quality, lint percentage should not be very high. In certain circumstances, higher boll weight should result in the increasing of lint yield and fiber strength . Good quality can be achieved selection for seed index. In some specific environments, selecting taller individuals should indirectly improve lint yield. In different environments, indirect selection and heterosis utilization on the fiber strength can be conducted through selecting on upper-half-mean length.（4） The heterosis analysis based on the main effect showed that average population mid-parent heterosis and average population better parent heterosis for seed-cotton yield and lint yield in F₁ and F₂ were large. The average population mid-parent heterosis for the fiber quality traits were small and their average population better parent heterosis were negative. Seven F₁ hybrids exhibited both high lint yield and good fiber quality.（5） The contribution ratio analysis of 6 agronomic and economic characteristics to lint yield showed that CR_{P （C→T）} of 5 out of the 6 traits were significantly positive, the CR_{AE （C→T）} of the tested traits were significantly positive, among which lint percentage was large （35.91%）, which means that in a particular environment, selecting on that trait was expected to significantly improve lint yield . The CR_{D （C→T）} of plant height, number of fruiting branches per plant, boll number per plant, boll weight, and lint percentage were significantly positive and large （57.70%～86.80%）, which means that selecting for hybrids with high dominant effect was expected to improve lint yield. The CR_{DE （C→T）} of lint percentage was significantly high （36.43%）, which means that selecting for hybrids with lint percentage with high dominant effect was expected to increase lint yield.（6） Taken young flowering buds as materials, expression profile analysis indicated that 73 quantitive trait transcripts （QTTs） , among which 10 QTTs’heritability were over 3%, significantly related to lint yield were measured in flowering buds. The additive variance for lint yield was lower while the dominance variance was larger. The additive variance for the yield components （lint percentage, boll number per plant and boll weight） were larger while the dominance variance was lower. Predicted values of additive effects of 9018 and sGK958 were higher at the gene expression level.They were good parents for high yield. However, Xinluzhong 9 and CCRI45 went against high yield. The same gene contribute different additive effect on different parents.The total predicted value of dominant effect of different crosses was higher. The QTTs, which have siginificant effects on dominance contribute value and on lint yield and its components, were different. Hence, their mechanism was different.更多还原