【作者】 雷东阳；

【导师】 陈立云； 谢放鸣；

【作者基本信息】 湖南农业大学 ， 种子科学与工程， 2008， 博士

【摘要】 稻米品质主要包括外观品质、加工品质、营养品质、蒸煮和食用品质等。稻米外观品质性状主要包括粒长、粒宽、粒长宽比、垩白和透明度等性状,是稻米商品品质的第一要素,直接影响稻米的商品价值。由于稻米品质一般由多基因控制,且易受环境条件的影响,因此利用传统的育种方法改良这些性状具有较大的难度。本研究利用多个群体系统地研究了稻米外观品质性状的遗传基础,定位了控制稻米外观品质性状的QTL,同时对其进行了相关遗传分析,为利用分子育种手段改良稻米外观品质建立了基础。主要结果如下:（1）利用所构建的“Lemont×特青”重组自交系（RIL）群体,采用混合线性模型和复合区间作图法,对旱、雨季条件下获得的稻米外观品质性状进行QTL定位及上位性和环境互作（Q×E）效应分析。结果表明,亲本Lemont和特青间米粒外观品质性状差异显著;外观品质性状尤其是垩白度在旱季和雨季不同条什下变化较大。共检测到20个控制粒长、粒宽、粒长宽比和垩白度的QTL,其中在第7染色体上检测到同时控制粒长、粒宽和长宽比的基因座位,位于标记区间OSR4-RM336之间。共检测到29对上位性互作位点,表明上位性效应对稻米外观品质性状具有重要的作用。共有5个主效QTL和8对上位性效应位点与环境之间互作显著,表明稻米外观品质性状的一些QTL对环境敏感。基因型与环境互作在不同性状和不同基因（QTL）间差别很大,就稻米外观品质性状而言,QTL与环境互作对垩白度及粒长宽比影响较大,分别占14.07%和17.78%的表型变异,而对粒长和粒宽的影响相对较小,分别只占2.77%和2.15%的表型变异。一些QTL（如ck4b、gl7a、gw5和lw3b）仅表现为加性效应,未检测到显著的环境互作,但另一些QTL（如ck3、gl3和lw9）与环境发生显著的互作。在实际育种上应利用在不同环境稳定检测到的QTL进行分子标记辅助选择。（2）分析了特青背景高代回交导入系（ILs）的遗传结构。在160个SSR标记位点上均不同程度地检测到供体等位基因,不同染色体及同一染色体的不同位点的供体等位基因导入频率各不相同,平均导入片段以第4染色体最多,第7染色体最少,表明染色体本身会影响等位基因导入。每个IL的供体基因平均占10.30%,变幅为0%～35%;平均导入片段数量为8.7个,单片段平均长度为17.8cM。选择56个导入系构建特青背景的全染色体组导入系群体,每个染色体由3～7个导入片段覆盖,非目标区域的导入成分有待标记辅助回交纯化。（3）利用特青背景的高代回交导入系对粒长、粒宽、粒长宽比和垩白度等性状进行了QTL分析。部分在重组自交系中定位的QTL尤其是主效QTL（qCD-7、qGL-7、qGW-7）在ILs中能够得到证实。但也有一部分QTL在两个群体的定位结果中有差异,尤其是利用ILs检测到控制粒宽的qGW-2,可解释表型变异的28.05%,但这个QTL在RIL群体里没有检测到,这可能是因为遗传背景的差异造成的。与RILs群体定位的结果一样,我们发现第7号染色体的OSR4-RM505区间是垩白度、粒长和粒宽的共同基因座位;另外还发现第6号染色体的RM439-RM340区间同时控制垩白度、粒宽和粒长宽比等性状。结果表明,利用ILs和RILs的定位QTL结果存在一定的差异,影响稻米外观品质相关性状的一些QTL位于同一染色体的相同或者相邻区域上。（4）选用IR75589-31S、IR60、IR70和重组自交系或高代回交导入系中选取的90个株系进行不完全双列杂交,以获得的270个杂交组合为材料,对杂交稻米外观品质性状表现及其与亲本的关系进行了研究。就杂交组合外观品质性状相对优势而言,除粒长宽比以外,其它外观品质性状的中亲优势都表现为正向组合个数大于负向。杂种稻米外观品质性状表现的差异因性状而异,以垩白度的变异系数最大,杂种米粒垩白度、粒长、粒宽和粒长宽比普遍介于双亲之间。稻米外观品质性状间的相关、回归和通径分析表明,垩白度、粒长、粒宽及粒长宽比间的相关系数均达极显著水平,粒长、粒长宽比与垩白度呈显著负相关,粒宽与垩白度显著正相关;粒长、粒宽和粒长宽比对垩白度的直接通径系数分别为-1.9409、2.1738和2.7805。直线回归分析显示外观品质性状间回归方程为y=-39.892x₁+141.389x₂+77.341x₃-269.906,从方程可以看出,当x₁、x₂、x₃变化量相同时,粒宽（x₂）对垩白度（y）的影响最大,其次为粒长宽比（x₃）。SSR分子标记遗传距离与杂种稻米外观品质相关性状杂种优势关系分析表明,亲本间的遗传距离在0.2914到0.4205之间,平均遗传距离为0.3520;杂交亲本遗传距离与外观品质性状间的相关系数偏小（-0.0307,0.1358,-0.1408）,均未达到显著水平,外观品质性状杂种优势与亲本间的遗传差异大小无关。（5）采用禾谷类作物种子品质性状的遗传模型及其统计分析方法,对杂交稻米外观品质性状进行了遗传研究,结果表明,杂交稻外观品质性状主要受遗传控制且遗传效应以加性效应为主。粒长、粒宽和粒长宽比以种子直接加性效应为主,母体效应也起到一定的作用,垩白度主要受到种子直接效应和母体效应控制,同时还受到细胞质的影响。粒长、粒宽和粒长宽比三性状的狭义遗传率均较高,都在80%以上,其中以粒宽为最高,达89.90%,垩白度的遗传率相对来说低一点,为72.94%。稻米外观品质各性状的均以种子直接遗传率最高,其次为母体遗传率,细胞质遗传率最低。更多还原

【Abstract】 environmentally sensitive.Another important finding of this study was that genetic by environment interactions（GEIs） were trait and gene（QTL） specific.The four traits differed considerably with respect to the effects of QEs.QE had larger effect on CD and LWR,accounting for 14.07%and 17.78%of the phenotypic variation,respectively.Very little of the phenotypic variation observed for GL and GW was due to QE only accounting for 2.77%and 2.15%of the variation,respectively.Some QTLs（e.g.ck4b, gl7a,gw5 and lw3b） had only additive main effects,while others（e.g.ck3,gl3 and lw9） also had an additive×environment interaction effects.The information obtained in this study should be very useful for manipulating the QTLs for these traits by molecular MAS.（2） A total of 217 Near-Isogenic Introgression Lines（ILs） in the Teqing background with introgressed Lemont segments were analyzed using 160 rice microsatellite markers. The donor segments from different chromosomes and chromosomal locations were introgressed into the recipient genome in different frequencies,suggesting that some known internal mechanisms could affect the transmission of genomic segments.The total introgressed segments were highest in chromosome 4 and lowest in chromosome 7.The average donor genomic portion in the ILs was 10.30%,ranging from 0 to 35%.The mean number of donor segments in a IL was 8.7 with 17.8 cM average length of introgressed segments.A set of 56 lines was selected to build ILs in Teqing background with approximately 3 to 7 lines for each chromosome.There are a few introgressive segments on none target regions requiring further elimination by marker-aided backcrossing.（3） The ILs with Teqing genetic background were used to dissect the genetic basis of CD,GL,GW and LWR by using QTL Network 2.0.Some of the main QTLs which were detected in the RILs,such as qCD-7,qGL-7 and qGW-7,were also identified in the ILs. But there are still some differences about the two populations.A major QTL（qGW-2） was detected for GW contributing 28.05%of the phenotypic variance.Surprisingly,the major QTL was not identified in the RILs.This may be due to the different genetic background.Among these QTLs,one common QTL for CD,GL and GW in interval OSR4-RM505 on chromosome7 was identified.This QTL also was identified in the RILs. A common QTL for CD,GW and LWR in interval RM439-RM340 on chromosome 6 was also identified.These results suggested that there were some differences in QTLs between the RILs and ILs.Moreover,there were some common QTLs for the closely relative traits at the same or near chromosome regions.（4） Ninety-three parents including IR75589-31S,IR60,IR70 and ninety recombinant inbred lines or advanced backcrossing introgression lines with different indica-japonica degrees were chosen to compare four characters of grain quality between F₁ hybrids and their parents.Based on the relative heterosis for all characters in all hybrids,the frequency of positive mid-parent heterosis was higher than that of negative mid-parent heterosis except for LWR.Grain appearance quality among F₁ hybrids varied significantly with greatest variation on CD.The performance of most F₁ hybrids for all the four characters was between the low-value parent and high-value parent.F₁s were found to be highly correlated with the male parent,female parent and average of the parents for all the four characters.Correlation analyses also revealed different relationships among the four characters tested.CD was significantly negatively correlated with GL(R=-0.5628^**) and LWR(R=-0.7498^**) but positively correlated with GW(R=0.7539^**).The results of path analysis showed that magnitudes of direction from GL,GW and LWR to CD were-1.9409,2.1738 and 2.7805,respectively. Regression analysis of rice grain appearance traits resulted in linear regression equation between CD（y） and GL（x₁）,GW（x₂）,LWR（x₃） of y=-39.892x₁+141.389x₂+77.341x₃ -269.906.This shows that GW is the most important factor affecting rice CD.The genetic distance of the rice parents and the relativity analysis were determined using SSR markers.The genetic distance between three female and ninety male based on 161 SSR markers varied in the range from 0.2914 to 0.4205 with an average of 0.3520.The relative coefficients between SSR genetic distance and heterosis of four grain appearance traits were partial to small（CD=-0.0307,GL=-0.1358,GW=0.1147,LWR=-0.1408）. The genetic distance of rice parents selected and the grain appearance heterosis was not significantly correlated.Because of this,SSR markers selected in this study couldn’t predict rice appearance quality heterosis,thus,the genetic analysis based on molecular markers predicting the grain appearance quality needs further studies.（5） The genetic effect of appearance quality traits on hybrid rice were studied by using the genetic model and analysis method for quality traits.The results indicated that all appearance quality traits were mainly influenced by directive additive variance（V_A） and maternal additive variance(V_Am).GL,GW and LWR were controlled mainly by directive effects and less by maternal additive effects.CD on the other hand,was mainly controlled by direct additive and maternal additive effects.GL,GW and LWR had very high narrow sense heritability values（86.03%,89.70,and 83.78%,respectively） while CD had relatively lower narrow sense heritability（72.94%）.Heritability for all the four grain appearance traits was more of direct heritability than maternal heritability.更多还原