【作者】 高明刚；

【导师】 李斯深；

【作者基本信息】 山东农业大学 ， 作物遗传育种， 2014， 博士

【摘要】 小麦是世界上重要的粮食作物，选育高产优质小麦品种是当前小麦育种的主要任务之一。构建高密度遗传图谱，将优异的种质资源转化为基因资源，从中发掘重要基因和QTL，对小麦分子标记辅助选择育种、分子聚合育种以及重要基因的图位克隆具有重要意义。本研究以“山农0431×鲁麦21”RIL（2010年F7）群体176个系为材料进行遗传连锁图谱的构建和产量相关性状的QTL分析。主要结果如下：利用SSR、DArT和SNP分子标记，构建了一张高密度遗传图谱。该图谱覆盖21条染色体，由42个连锁群组成，包含5916个位点，其中独立位点4530个（3605个DArT、850个SNP、75个SSR位点）。图谱全长2929.96cM，染色体平均长度为139.52cM，其中，3B染色体最长，为209.95cM；3D最短，为31.15cM。A、B、D基因组的长度分别是1065.36cM（36.4%）、1072.95cM（36.6%）和791.65cM（27.0%），标记数分别是1591（35.1%）、2278（50.3%）和661（14.6%）。该图谱的分子标记间的距离范围从6A的0.31cM到1D的2.18cM，图谱的总平均密度是0.65cM。将RIL群体在8个环境中种植：2011年菏泽（E1）、2010年泰安（E2）、2011年泰安（E3）、2011年泰安旱地（E4）、2011年烟台（E5）、2011年淄博（E6）、2012淄博旱地（E7）和2012年淄博（E8）。方差分析表明，8个环境下的14个产量相关性状在基因型和环境间均表现为极显著差异（p≤0.001）。粒长的遗传力最高，为76.9%；籽粒产量最低，为28.10%。相关分析表明：株高（PH）分别与每平方米穗数、千粒重、籽粒产量、穗长、基部不育小穗数、顶部不育小穗数和粒宽7个性状表现显著的正相关关系（p≤0.05）。对于产量性状，籽粒产量与每平方米穗数和穗粒数均表现为显著负相关，而产量三要素彼此之间均表现为显著负相关。对于穗部性状，穗长、总小穗数和可育小穗数彼此之间均表现为极显著正相关，基部不育小穗数与总小穗数表现为极显著正相关，而与可育小穗数表现为显著正相关。顶部不育小穗数和可育小穗数之间表现为显著负相关；对于籽粒大小性状，籽粒长宽比与粒长呈显著正相关，而与籽粒密度呈显著负相关。本研究共检测到204个QTL，分布于除3D外的20条染色体上，单个QTL解释平均表型变异的3.84%~33.31%。其中，123个QTL表现为正加性效应，其增效效应来源于山农0431，81个QTL表现为负加性效应，其增效效应来源于鲁麦21。E3环境下的株高QTL的LOD值最高，为17.03，解释表型变异的33.31%。控制产量及构成因素的的QTL有56（27.5%）个，分布于除3D、4B和6D外18条染色体上，单一QTL解释表型变异的5.26%~20.15%；控制株高及穗部的QTL有82（40.2%）个，分布于除3D外的20条染色体上，单一QTL解释表性变异的4.65%~33.31%；控制籽粒大小性状的QTL有66（32.4%）个，分布于除3D和6D外19条染色体上，单一QTL解释表型变异的3.84%~24.36%；204个QTL中，19个QTL在4个或以上环境中同时被检测到，是相对稳定的QTL（RHF-QTL）。这些QTL分布于9条染色体上：1A、2B、2D、3B、4B、4D、5A、5B和6A，涉及株高、千粒重、穗长、基部不育小穗数、总小穗数、粒长、籽粒长宽比和籽粒密度在内的8个性状，解释平均表型变异的6.59%~24.05%。其中3个RHF-QTL（QPh-2B，QTss-6A和QBsss.2-6A）在各环境中加性效应表现为负值，其增效效应来自于鲁麦21，其余16个RHF-QTL加性效应均表现为正值，其增效效应来自于山农0431。7个RHF-QTL（QPh-4D、QSl.2-1A、QTss-6A、QBsss.2-6A、QFfd-5B、QGl.2-5B和QGlw-5B）解释平均表型变异的百分比大于10%，分别为15.59%、11.38%、18.07%、24.05%、10.27%和13.74%，为稳定的主效QTL。检测到包含3个或3个以上QTL置信区间重叠的QTL簇21个（C1~C21），分布于10条染色体上：1A、1B、1D、2B、2D、3A、3B、4D、5B和6A，包含84个QTL（41.18%）。除4个RHF-QTL（QPh-2B、QTss-4B、QBsss.2-6A和QBsss-5A）外，其余15个RHF-QTL（15/19×100%=78.95%）集中分布于10个QTL簇（C2、C8、C10、C11、C15、C16、C17、C18、C20和C21）中，分布于7条染色体（1A、2B、2D、3B、4D、5B和6A）上，包含46个QTL，9个QTL加性效应为负值，表明其增效效应来自于鲁麦21，另有37个QTL加性效应为正值，其增效效应来自于山农0431。这10个QTL簇是重要的QTL簇。更多还原

【Abstract】 A high genetic map was constructed of wheat (Triticum aestivumL.) and QTL analysis foryield traits was conducted using a population of176recombinant inbred lines (RILs) derivedfrom “Shannong0431×Lumai21”(F7in2010). The main results are as follows:A total of74500molecular markers were used to screen the polymorphism of RILs andtheir parents. A number of7955polymorphic loci (6237DArTs,1519SNPs and199SSRs)were identified and used to construct the genetic map. A genetic map with5916loci wasfinally constructed covering all the21chromosomes. In which,4530loci were unique loci,including3605DArTs,850SNP and75SSR loci; and the other1386loci showedco-segregation with other markers. The final map spanned a total length of2929.96cM across42linkage groups, with an average chromosome length of139.5cM. The largest chromosomewas3B (201.95cM), and the shortest was3D (31.15cM). The total number of mapped lociper chromosome ranged from30(3D) to447(3B) with an average of215.71loci perchromosome. The density ranged from0.31(6A) to2.18(1D) cM/marker with an averagedensity of0.65. The map length and locus number was unequally divided among the threegenomes:1065.36cM (36.4%),1072.95cM (36.6%) and791.65cM (27.0%) in lengths; and1591(35.1%),2278(50.3%) and661(14.6%) loci for the A, B and D genomes, respectively.The seven homologous groups also varied in locus number and map length: group7containedthe most loci and length (899loci covering471.16cM); whereas group4had the smallest lociand length (402loci covering245.25cM).The parents of the RILs displayed remarkable differences for the investigated traits in theeight environments and their average value (AV). For the RILs, the variance for genotypesand environments of all of the14investigated traits were significant at the P≤0.001level,indicating that the environments and genetic background were both important in explainingthe overall phenotypic variations. The for the investigated traits ranged from28.10(GY) to76.90%(GL), which were over50%for PH, TGW, SL, GL, GW, GLW, BSSS, TSSS andTSS; whereas lower than50%for GY, FSS, SN, FFD and GNS. The coefficients of variation(CVs) ranged from3.38%for GW in E4to197.53%for TSSS in E2. Transgressivesegregation was observed for all of the traits in the eight environments and AV.The Pearson correlation coefficients for PH appeared significant positive correlations with SN, TGW, GY, SL, BSSS, TSSS and GW; and negative correlation with GNS. The yieldcomponent traits, SN, GNS and TGW, showed extremely significant negative correlation. Thesignificant positive correlations were obtained between GY and SN/GNS. For spike traits, thestrong positive correlations were simultaneously obtained between SL, TSS and FSS. BSSSappeared significant positive correlations with TSS, and negative correlation with FSS.Significant negative correlations were found between TSSS and FSS. Some correlationcoefficients were significant between GNS/TGW/GY and spike traits. For the grain size traits,the significant positive correlations were obtained between the GL and GLW, and significantnegative correlations between the GLW and FFD. Some correlation coefficients weresignificant between yield traits and grain size traits.A total of204additive QTLs (370QTLs for trait-environment combinations) weredetected on20chromosomes except for3D for all of the investigated traits in the eightenvironments and their AV. An individual QTL in different environments explained3.84-33.31%of the phenotypic variations. Of which,123QTLs showed positive additiveeffects with Shannong0431increasing the effects of QTLs, whereas81QTLs were negativeeffects with Lumai21increasing the QTL effects. The highest LOD value for a single QTL inthe different environments was17.03for PH in E4. Fifty-six QTLs were identified for yieldcomponent on20chromosomes except for3D,4B and6D. An individual QTL in differentenvironments explained5.26-20.15%of the phenotypic variations. Eight-two QTLs wereidentified for PH and spike triats on20chromosomes except for3D. An individual QTL indifferent environments explained4.65-33.31%of the phenotypic variations. Sixty-six QTLswere identified for PH and spike triats on19chromosomes except for3D and6D. Anindividual QTL in different environments explained3.84-24.36%of the phenotypicvariations.Nineteen relatively high-frequency (RHF) or relatively stable QTLs (9.27%) expressing inmore than four environments and/or in AV were located, including eight traits (PH, TGW, SL,BSSS, TSS, GL, GLW and FFD) with the average contributions ranging from6.59%to24.05%. Of which, all the RHF-QTLs except for QPh-2B, QTss-6A and QBsss.2-6A showedpositive additive effects with Shannong0431increasing the effects of QTLs. SevenRHF-QTLs (QPh-4D, QSl.2-1A, QTss-6A, QBsss.2-6A, QFfd-5B, QGl.2-5B and QGlw-5B)were stably main effect QTLs with explaining more than10%of phenotypic variations.A number of21QTL clusters (C1-C21) with more than three traits were mapped on10chromosomes:1A,1B,1D,2B,2D,3A,3B,4D,5B and6A, which were related to all of theinvestigated traits and involved84QTLs (84/204×100%=41.18%) and15RHF-QTLs (16/19×100%=84.21%). Cluster C2, C8, C10, C11, C15, C16, C17, C18, C20and C21detected in at least one RHF-QTL and over nine trait-environments combinations, which wereconsidered the most important QTL clusters, and these QTL cluster interval should be usefulfor marker-assisted selection (MAS).更多还原