【作者】 孙建昌；

【导师】 陈耀锋；

【作者基本信息】 西北农林科技大学 ， 作物遗传育种， 2012， 博士

【摘要】 水稻是中国的第一大粮食作物，我国大约2/3的人口以大米为主食。近年来随着农业科学技术的发展和应用，水稻单产和亩产呈现连续递增的势头。然而育种家们在追求高产品种选育的同时，品种间的遗传基础越来越狭窄，突如其来的灾害一旦发生可能会造成不可估量的损失和严重后果。拓宽现有品种的遗传基础，培育遗传背景复杂的高产优质品种对我国水稻生产的可持续发展有重要意义。地方品种是在水稻生产和驯化过程中经自然和人为长期选择的产物，与育成品种相比，遗传背景复杂，具有丰富的遗传多样性和异质性，对环境具有很强的适应性，是丰产、优质、抗病虫、抗逆等重要性状的优异基因源。是育种家们培育各类品种所需材料取之不竭的源泉。因此，有效保护水稻地方品种对我国水稻生产的可持续发展意义重大。育成品种是经一代代育种家们定向培育，是许多高产、优质、抗逆、抗病虫等优异基因的聚合体，更符合人类对粮食的需求和生产，在生产上发挥着重要作用。同时也是品种培育和改良的最核心的遗传资源。利用不同地理来源育成品种的群体进行基因发掘对实际的生产利用有重要意义。本研究以云南水稻地方品种和来自世界各地的粳稻品种为试验材料，对云南水稻地方品种保护机制及不同地理来源粳稻主要农艺性状进行管理作图，获得如下结果：1.利用20对SSR引物对原产于云南的16份水稻地方品种和2份选育品种进行单个品种群体内的遗传多样性分析。结果表明，87.5%的地方品种群体内SSR标记多态性高于选育品种，而12.5%的地方品种群体内SSR标记多态性与选育品种相近。81.2%的地方品种群内的等位基因数（Na）和Nei基因遗传多样性指数（He）高于选育品种，而18.8%的地方品种群体内Na和He与选育品种相同或略小。水稻地方品种间群体内He的差异较大，其变异范围为0.0146～0.5117，黄板所-2、麻线谷-1、麻线谷-2的群体内遗传多样性较高，分别为0.4214、0.5117和0.4489。87.5%地方品种的杂合度（Ho）明显高于选育品种；地方品种群体间遗传多样性差异很大，其中1/4的遗传差异性来源于地方品种的群体内，差异呈极显著水平(P<0.001)。RM333、RM257和RM180在供试云南地方品种群体内的多态性、等位基因数、多样性指数和变异百分率均较高，适合应用于云南水稻地方品种群体内遗传多样性检测。2.选择原地保护（2007年收集）和异地保护（1980年左右收集）的8对同名相同云南水稻地方品种为试验材料，比较原异地保护地方品种的表型差异性。调查的7个农艺性状中，除剑叶宽变幅1980群体大于2007群体外，其余6个性状均表现为2007群体变幅大于1980群体。变异系数表现为九月糯、香谷、大白糯、麻线谷均有6个性状2007群体的变异系数高于1980群体；接骨糯和冷水谷分别有和4个农艺性状2007群体高于1980群体。比较8对地方品种1980群体和2007群体各农艺性状的平均变异系数，除穗抽出度外，其余6项农艺性状均表现为1980群体小于2007群体。以上结果表明均表明云南水稻地方品种2007群体内表型变异比1980群体表型变异更大，说明原地保护方式更利于产生表型变异。3.原、异地保护的8对同名相同水稻地方品种SSR标记差异性分析表明，在原地保护地方品种的每个群体检测到等位基因43～88个，平均每条SSR引物等位基因为2.15～4.40个；而在异地保护地方品种的每个群体检测到等位基因33～65个，平均每条SSR引物等位基因为1.65～3.25个。除齐头谷外，其余7对地方品种的原地保护群体的等位基因数显著高于异地保护群体。在8对地方品种群体中，齐头谷和黄版所异地保护群体的Nei基因多样性指数无显著或显著高于原地保护群体外，其余6对品种的原地保护群体的Nei基因多样性指数均显著高于异地保护群体。除齐头谷的异地保护群体特异等位基因数高于原地保护群体外，其余7对地方品种原地保护群体的特异等位基因数均高于异地保护群体，且原地保护群体的特异等位基因数为异地保护群体的2.1～5.0倍。AMOVA分析表明，原、异地保护同名地方品种间群体遗传结构差异极显著（P<0.001），变异百分率均在20％以上。原、异地保护的同名云南水稻地方品种的遗传结构和遗传多样性具有显著的差异，原地保护群体具有更丰富的等位基因和基因多样性。4.选择He较小的接骨糯、冷水谷、齐头谷和He较大的麻线谷，分析20、40、60、80、100株群体与原群体间的遗传差异性。以原群体2%以上的等位基因为对照，40～60株群体可以保持对照98%的等位基因数，80株群体可以检测到对照全部等位基因。以原群体5%以上等位基因为对照，40株群体可以检测到98%的等位基因，60株能完全检测到对照的等位基因。有效等位基因数(Ne)和He分析表明，遗传多样性较低的接骨糯和冷水谷在繁殖株数为40株时，可以保持原群体的Ne和He，而遗传多样性较高的麻线谷需80株群体来保持原群体的Ne和He。结果表明地方品种更新繁殖群体受该品种群体内异质性的影响，利用SSR标记分析认为有效保持水稻地方品种群体遗传完整性的更新繁殖株数为40～80株。5.选取两个遗传多样性差异大的云南水稻地方品种，利用混合取样法对每次选取120株连续繁殖4次，分析各群体间的遗传结构差异。等位基因数差异显著，但这种变化主要由频率很低的稀有等位基因增加或消失引起。有效等位基因和Nei遗传多样性指数在很小的变异范围内浮动，群体间差异不显著。其中异质性高的麻线谷的浮动幅度更大一些，说明异质性高的水稻地方品种更容易发生遗传侵蚀。参试品种4次繁殖的相似系数分别在0.98和0.99以上。2～4次繁殖随机抽取40株和80株群体分析其与原群体的遗传相似性，冷水谷40株群体的相似系数在0.99以上，而麻线谷需80株群体保持0.99以上的相似性，进一步证明了40～80株群体在繁殖次数中可以保持原群体的遗传相似性。以上结果表明云南水稻地方品种更新群体为120个单株时，连续繁殖4次后，保持了原群体的遗传完整性。根据地方品种群体内遗传多样性大小不同，40～80株群体可以保持原繁殖群体的遗传多样性。由此推测，利用40～80或更大的繁殖群体多次繁殖不会显著影响原材料的遗传结构和完整性。6.利用主要农艺性状和SSR标记对347份粳稻种质进行群体结构分析和关联作图。表型分析表明，来自不同国家或地区品种间农艺性状差异较大，各性状变异系数在5.46～27.36之间，极差值（除生育期外）占平均值比率范围为67.7～149.6%。Structure群体结构分析表明参试群体分3个亚群，生态类型相似地区的品种基本归在1个亚群。利用TASSEL软件的GLM模型对7个农艺性状2年数据关联作图表明，在公共图谱上共线性或非共线性的位点组合广泛存在连锁不平衡（LD），但不平衡程度的r2>0.5组合数少，只占总位点数的5.83%。148个位点中有76个位点共计216个（次）和7个农艺性状显著关联，共50个（次）位点与两年表型数据均显著关联。76个位点中与生育期关联位点31个，株高19个，穗长22个有效穗数和千粒重各20个，穗粒数28个，结实率24个；有46个位点与2个或2个以上性状关联。在与各性状的关联位点中一些位点与前人利用QTL定位位点的位置相同或接近。结果表明关联作图法定位是传统QTL定位的有效补充，两者结合起来可能是水稻复杂性状新基因定位和发掘的有效方法。更多还原

【Abstract】 Rice is the first food crop in China, the staple food of people about2/3is rice in China.With the development and applications of agricultural science and technology, rice yield is themomentum of continuous increment in recent years. However, the genetic basis of varieties ismore and more narrow while breeders are pursuing high-yielding varieties at the same time.Unexpected disaster, such as diseases, pests, etc. may result in incalculable losses and seriousconsequences. It is important to sustainable development of rice production for broadeningthe genetic basis of the varieties and nurturing the complex genetic background, high-yield,high-quality rice varieties. Landraces is the product of long-term choice by natural andman-made in rice production and domestication. Compared with breeding rice, landraces havemore complex genetic background, richer genetic diversity and heterogeneity, and strongeradaptability to the environment. It is an inexhaustible source for breeders to nurture all kindsof varieties which contain many excellent genes such as high yield, high-quality, disease andinsect resistance, stress tolerance etc. Therefore, the effective protection of rice landraces issignificant for sustainable development of rice production in China.By breeders oriented cultivation of generation to generation, bred varieties is a polymerof many excellent yields, good qualities, tolerance, resistance to insect pests and other genes.Bred varieties plays important roles in cereals production and meeting human demand. Bredvarieties are also the core of genetic resources for breeding and improvement. It is importantto discover excellent genes using bred varieties come from different geographical orecological places.In this study, Yunnan rice landraces and japonica varieties from around the world as testmaterial, we studied to protection mechanism of Yunnan rice landraces and associationmapping of main agronomic traits of japonica rice. The main results were summarized asfollows:1. Genetic diversity within populations of16rice landraces and2advanced cultivarsfrom Yunnan were analyzed using20SSR markers. The results showed that the SSR markerpolymorphism within populations for87.5%of rice landraces was higher than that ofadvanced cultivars, while the SSR marker polymorphism within populations for12.5%of landraces was similar with that of advanced cultivars. The number of alleles (Na) and Nei’sgenetic diversity index (He) were showed that81.2%of rice landraces were higher than thoseof advanced cultivars, and18.8%of landraces were same or slightly smaller than advancedcultivars. The He within populations of rice landraces were significantly different, rangedfrom0.0146to0.5117, while HBS-2, MXG-1and MXG-2were the highest with0.4214,0.5117and0.4489, respectively. Compared heterozygosity within rice varieties,87.5%of ricelandraces were significantly higher than improved cultivars. AMOVA showed that1/4geneticdifference among rice landraces were from within populations of rice landraces (P<0.001).Compared the polymorphism, Ne, He and percentage of variation within populations of ricelandraces among20SSR markers, RM333, RM257and RM180were higher than others. Itspeculated that RM333, RM257and RM180were suitable to testing genetic diversity withinrice landraces of Yunnan.2. The phenotypic differences eight pairs of rice landraces from in situ (collected in2007)and ex situ (collected in1980) conservation programs with single-origin were studied. Surveyof7agronomic traits, in addition to the FLW variation range of1980population was greaterthan that of2007population, the remaining six traits showed that the2007population wasmore than1980population. the coefficient of variation showed that JYN, XG, DBN and MXGwere6agronomic traits which2007populations was higher than1980populations, andJGN and LSG was4. Compared the average coefficient of variation between8pairsagronomic characters of rice landraces populations collected in1980and2007, six agronomictraits showed that2007population was higher than1890population except for PE. Theseresults indicate that the phenotypic variation within Yunnan rice landraces of in situconservation programs was higher than that of ex situ conservation programs.3. The genetic structure and diversity of eight singer-origin pairs of rice landraces fromin situ and ex situ conservation programs were studied using20pairs of microsatellitemarkers with high polymorphism. The number of alleles detected in the populations from insitu conservation ranged from43to88with the mean number of alleles per locus rangingfrom2.15to4.40, while the number of alleles detected in the populations from ex situconservation ranged from33to65, and the mean of alleles per locus ranged from1.65to3.25.Compared to the ex situ populations, the number of alleles, the number of specific alleles andthe genetic diversity index showed a significant increase in the in situ populations. Further,the numbers of specific alleles from in situ populations were2.1to5.0times greater than inex situ populations except for rice landrace QTG. An AMOVA showed that thewithin-landrace genetic structure differed significantly between in situ and ex situ conservation treatments with differences exceeding20%. The analysis of genetic similarityreached similar conclusions to those of the AMOVA. Compared with ex situ conservationprograms, the rice landraces under in situ conservation programs had more alleles and highergenetic diversity in Yunnan of China.4. Selected JGN, LSG and QTG with smaller He and MXG with higher He, geneticdifferences were analyzed among20,40,60,80,100populations (randomly selected fromoriginal populations respectively) and original populations.40to60populations can contain98%alleles, and80populations can be detected in all alleles compared the originalpopulations of alleles frequency more than2%. Compared alleles more than5%of originalpopulations,40populations can be detected98%alleles while60can be fully detected.Changes of Ne and He showed that JGN, LSG and QTG with low genetic diversity in thereproduction need40populations to keep Ne and He of original populations while MXG withhigh genetic diversity need80populations. The results showed that the number of populationsto contain genetic integrity in updating was affected by the heterogeneity of the rice landraces.In our study,40～80populations can effectively maintain the genetic integrity of ricelandraces to update using SSR markers.5. Selected2rice landraces of Yunnan with significant differences in genetic diversity,120single plants of populations were planted four times consecutively using a bulk systemgetting seeds. Genetic differences among populations were analyzed. The number of alleleswere significant difference, but this change mainly caused by increasing or disappearing rarealleles with low frequency. The Ne and He changed within the small fluctuations, and nosignificant difference among populations. MXG with high heterogeneity floating range wasbigger than that of LSG, this indicated that rice landraces with high heterogeneity are moreprone to genetic erosion. Similarity coefficient of rice landraces planting4times were0.98and0.99, respectively. Analyzed similarity coefficients among40and80populations(randomly selected from the original populations) and the original populations planted2to4times,40populations of LSG and80populations of MXG were above0.99. It was provedthat40to80populations can be kept genetic similarity of the original groups to planting4times continuously. The result showed that using120individuals to update rice landraces canmaintain the genetic integrity of original populations.40to80individual populations containsimilarity among4times planting. It speculated that used of40to80individual populationscan maintain genetic diversity for repeatedly breeding.6. Association mapping of347japonica varieties using the main agronomic traits andSSR markers. Analysis of Phenotype showed that the agronomic traits and coefficient of variation of varieties from different national or regional were large variations. Coefficient ofvariation of traits from5.46to27.36except for MD, accounted for the average ratios from67.7to149.6%. Analysis of Structure showed that the tested groups divided into three subsets,the variety of similar areas or ecological types classified in same subgroup. Agronomic traitswith2years data and148SSR markers were associated mapping using GLM model ofTASSEL software. LD was detected extensively not only among syntenic markers but alsoamong nonsyntenic ones, while the loci pairs with r2>0.5accounted for only5.83%of thetotal ones.76sites totaling216(times) were significantly agronomic traits while50(times)associated2years data at the same time within148loci. Among76loci, MD associated with31loci, PH is19, PL is22, PPP is28,1000-GW is20, SPP and Ssr is24respectively.46lociwere associated with two or more traits. Some loci located in this paper were the same orclose to the loci of QTL mapping by previous. The results showed that association mappingmethod is an effective complement to traditional QTL mapping, and combine the twomethods is the effective way to map and excavate rice complex gene.更多还原