【作者】 徐新福；

【导师】 唐章林； 柴友荣；

【作者基本信息】 西南农业大学 ， 作物遗传育种， 2005， 硕士

【摘要】 杂种优势是生物界普遍存在的一种现象,利用杂种优势是提高作物产量和改良品质最有效的途径之一。长期以来,人们一直致力于探讨杂种优势的遗传基础及其预测方法,以便能更好地利用杂种优势。分子标记的发展与应用,为深入研究作物的杂种优势现象提供了新的、有效的技术平台。 油菜是世界范围内广泛种植的主要油料作物之一。我国油菜种植面积和总产均约占世界的1/3,均居世界首位,其中杂交种种植面积超过栽培总面积的30%。我国的油莱杂种优势利用研究居国际领先地位,但在杂优育种中,还没有一个简便而有效的预测指标来指导亲本选配。在多种作物上分子标记技术被应用于杂种优势的预测及遗传基础研究,而在油菜上这方面的研究报道还比较少。 本研究通过对甘蓝型油菜杂交亲本DNA分子标记数据和F₁生物学性状数据分析,估计了标记座位的效应和性质（加性、显性）,通过分析杂种及亲本标记型值与杂种表现、杂种优势的关系,研究甘蓝型油菜杂种优势的遗传基础,利用逐步回归分析建立了分子标记预测杂种优势的模型。主要结果如下: 1.杂交亲本的遗传关系分析 研究材料中的6个不育系分别来自不同的选育单位,遗传差异较大,11个恢复系中除17号材料955由华中农业大学选育外,另外10个均由西南农业大学选育而成,而且多个材料含有GH01的血缘。 用45对SSR引物和35对AFLP引物对17个杂交亲本基因组DNA模板进行扩增,共得到319个多态性位点。利用这319个标记位点计算17个亲本材料的分子标记遗传距离,采用UPGMA法进行系统聚类分析。聚类结果与血缘组成并不一致,含有GH01血缘的材料分布在不同聚类组中,说明经过多年选育,材料的遗传组成在很大程度上得到改变。总体上,不育系与恢复系间遗传差异大于它们各自内部的遗传差异。更多还原

【Abstract】 Heterosis is a common phenomenon in biosphere. Application of heterosis is one of the most effective ways to enhance yield potential and to improve crop quality. For long time many researchers have endeavored themselves to elucidate the genetic basis and prediction methods of heterosis intended to utilize plant hertorosis more efficiently. The development and application of molecular marker technologies have provided a novel and effective platform for the deep study of heterosis in crops.Rapeseed is one of the most widely planted oil crops worldwide. The planting area and output of rapeseed in China is about one third of the world’s total amounts and ranked first among all the rapeseed growing nations. Hybrid cultivars account for over 30% of China’s total acreage of rapeseed. The research on the utilization of rapeseed heterosis has being kept ahead of the world, but there has not being effective and simple indexes in the breeding of heterosis to direct parent mating. Molecular marker technoliogies have been widely used in the researches on the prediction of heterosis and basic hereditary principles of many crops, but little has been reported on rapeseed heterosis prediction though molecular markers.In this study about the heterosis prediction of rapeseed, the effects and features of the marker seats （additive or dominant） were estimated based on the analysis of the relationships between DNA fingerprint data of parents and the F₁’s biological traits; the hereditary basis of the heterosis was elucidated according to the analysis of relationships between the marker values of hybrids and parents and the heterosis values; and a mathematic model for heterosis prediction by molecular markers was set up according to the stepwise regression analysis. The results are as follows:1. Analysis of hereditary relationships among hybrid parentsIn this research, the 6 sterile lines were from different beedmg institutes and have distinct hereditary distances, while the 11 restorer lines were all bred by Southwest Agriculture University and many of them have the consanguinity of line GH01, except for line 995 which was bred by Huazhong Agricultural University.319 polymorphic loci were detected through the amplification of genomic DNA templates of 17hybrid parents using’ 45 pairs of SSR primers and 35 pairs of AFLP primers. The hereditary distances of the 17 parents were calculated according to the 319 polymorphic loci, and cluster analysis was carried out using UPGMA method. The results of cluster analysis were not consistent to the consanguinity compositions of the parents. Materials derived from GH01 were clustered into different groups, indicating that the hereditary compositions of materials had greatly changed through years of selection. In general, hereditary differences between sterile lines and restorer lines were greater than those within each type.2. Hereditary analysis of yield traitsVariance analysis was carried out for 9 traits related to yield potential of 84 materials including hybrids, parents and CK, and 66 hybrid crosses. Extremely significant differences were detected among all the 84 materials and 66 crosses.Within the 9 traits tested the yield per plant of crosses showed the topmost positive colony average heterosis ratio, with a highest value of 34.99%. The heterosis ratios of yield and its components were as follows: yield per plant > plant silique numbers > seed number per silique > 1000-seed weight. The colony average heterosis ratio of 1000-seed weight was the lowest with a value of only 3.97%.The dominant and additive hereditary variances of the 9 traits tested all reached the significance level of Pooooi, which indicated that the dominant and additive hereditary effects widely existed within these 9 traits. In plant height, dominant hereditary effect was higher than additive hereditary effect. But in yield per plant, first effective branching node, siliques of main branches and seed number per silique, dominant and additive hereditary variances were near in values, demonstrating that these two effects were equal in importance. In the rest 4 traits, the additive variances were overwhelming over the dominance variances.3. Screening of trait-related molecular markers and effect estimationFor the 9 traits tested, different numbers of markers with dominant or additive effects were screened. Each marker locus had different effects. Some loci had only additive or dominant effects, while others had both of them. These marker loci had different effects on the size and direction of the traits. In each trait, loci with additive effects were more than those with dominant effects, indicating that dominant heredity was more influential than additive heredity. Interaction loci were not detected in yield per plant, but occurred within 24 marker loci on other 8 traits. Different screening methods resulted in different numbers of specific marker loci.4. Relationships between genetic distances （GD） and hybrid performance and heterosisThe general genetic distance （GGD） of parents had no obvious relationship to the performance of 9 traits tested except for the effective silique number of main branches. Though the correlations between two kinds of special genetic distances （SGD₂, SGD₃） and several F₁ traits performance reached significant level of P_0.01, the correlation coefficient and determination coefficient were both更多还原