【目的】灌浆是玉米籽粒形成的重要生理过程,直接决定了籽粒的最终产量。了解玉米籽粒灌浆特性相关性状对粒重形成的作用,解析灌浆特性的遗传基础,为玉米高产育种实践提供指导。【方法】以中国玉米骨干自交系黄早四(HZS)、旅28(Lv28)为亲本构建的包含172个家系的重组自交系(recombination inbred line,RIL)群体为试验材料。首先,利用Logistic模型与Richards模型,进行玉米籽粒灌浆过程拟合度的比较分析。其次,利用方差分析、相关性分析及回归分析分别比较亲本籽粒灌浆特性的差异,研究群体中不同灌浆特性相关性状的关系及其对百粒重的贡献。然后,利用GBS方法,对群体进行基因型分型,选择亲本间多态性标记,构建遗传图谱。最后,利用完备区间作图法(inclusive composite interval mapping,ICIM)进行灌浆特性与生育期相关性状的QTL分析。【结果】籽粒灌浆一般呈现慢-快-慢的变化趋势,可分为缓增期、快增期以及减缓期3个阶段。通过比较不同灌浆模型的拟合度发现,基于Richards模型的预测值与表型值间的决定系数显著高于Logistic模型。比较亲本间灌浆特性的差异发现,黄早四的平均灌浆速率为旅28的1.28倍,但旅28的灌浆持续时间为黄早四的1.07倍,亲本之间在灌浆特性方面差异明显。群体表型相关性分析发现,除缓增期灌浆持续时间(T1)外,其他灌浆特性相关性状均与百粒重(HKW)存在显著的正相关关系。回归分析发现,快增期灌浆持续时间(T2)与灌浆速率(G2)可分别解释百粒重表型变异的57.50%和30.00%。利用多态性SNP标记构建了全长为1 471 c M,标记间平均遗传图距为1 c M的遗传图谱。多个环境下共检测到26个灌浆特性相关QTL、3个百粒重相关QTL及14个生育期相关的QTL,分布在玉米除第7染色体外的其他染色体上,LOD值介于3.27—9.05,单个QTL贡献率为5.97%—21.16%。同时,利用联合环境分析发现,控制不同性状的QTL定位在染色体相同或相近的位置,形成了多个分布于玉米基因组bin 1.05、bin 2.03、bin 4.05、bin 4.06、bin 7.04、bin 9.04的QTL富集区域。其中,在位于bin 4.05(48.24 Mb—135.73 Mb)和bin 9.04(110.40 Mb—114.73 Mb)的区间之内,共定位到多个仅与灌浆速率相关的主效QTL。【结论】Richards模型能够更好地模拟玉米籽粒的灌浆过程。在灌浆特性相关性状中,快增期灌浆速率与灌浆持续时间对于玉米粒重的增加具有重要作用。单环境检测发现,灌浆持续时间相关位点仅能在单环境中得以检测,表现为环境敏感类型。联合环境分析发现,在bin 4.05和bin 9.04区间内分别检测到仅与灌浆速率相关的主效QTL,可作为玉米籽粒灌浆研究的重点区域。 【Objective】 Grouting is an important physiological process for the formation of maize grain and directly determines the final grain yield. To understand the role of grain traits related to the formation of grain weight, to understand the genetic basis of grain-filling characteristics and to provide guidance for high-yield breeding of maize. 【Method】 A recombination inbred line (RIL) population containing 172 lines constructed from HZS and Lv28 maize inbred lines of maize in China was used as experimental materials. First, the logistic model and the Richards model were used to make a comparative analysis of the fitting degree of corn grain filling process. Secondly, variance analysis, correlation analysis and regression analysis were used to compare the differences of grain filling characteristics of parents to study the relationship between different grain filling characteristics and their contribution to the 100-grain weight. Then, using GBS method, the genotypes of the population were genotyped, and the polymorphic markers between parents were selected to construct the genetic map. Finally, QTL analysis of grain filling traits and growth-related traits was performed using inclusive composite interval mapping (ICIM). 【Result】 Grain filling generally showed a slow-fast-slow change trend. It could be divided into three stages of slow growth period, rapid growth period and slowing period. By comparing the fitting degrees of different grouting models, it is found that the coefficient of determination between the predicted value and the phenotype value based on the Richards model is significantly higher than that of the Logistic model. Comparing the differences of filling characteristics between parents, we found that the average filling rate of Huangzao 4 was 1.28 times of that of Lu 28, but the duration of filling of Lu 28 was 1.07 times of that of Huangzao 4, and there was significant difference in the grouting characteristics among parents. The phenotypic correlation analysis showed that there was a significant positive correlation between the grain filling traits and the weight per 100 kernel weight (KWWW) except for the duration of filling (T1). Regression analysis showed that the duration of filling (T2) and filling rate (G2) during fast accretion period could explain 57.50% and 30.00% of the phenotypic variation of the 100-grain weight, respectively. A genetic map with a total length of 1 471 cM and an average genetic distance of 1 c M between markers was constructed by using the SNP marker. A total of 26 QTLs for grain filling traits, 3 QTLs for 100-kernel weight and 14 QTLs related to their growth stages were detected in a variety of environments. The distribution of QTL on other chromosomes except maize chromosome 7 was between 3.27-9.05 , A single QTL contribution rate of 5.97% -21.16%. At the same time, using the joint environmental analysis, it was found that QTLs controlling different traits were located at the same or similar chromosomes, resulting in multiple QTLs rich in maize genomes such as bin 1.05, bin 2.03, bin 4.05, bin 4.06, bin 7.04 and bin 9.04 Set area. Within the interval of bin 4.05 (48.24 Mb-135.73 Mb) and bin 9.04 (110.40 Mb-114.73 Mb), a number of major-effect QTLs related only to the filling rate were co-localized. 【Conclusion】 The Richards model can better simulate the grain filling process of corn. Among the traits related to grain-filling traits, the grain-filling rate and the duration of grain-filling were important for the increase of grain weight. Single environmental test found that the grout duration related sites can only be detected in a single environment, the performance of environmentally sensitive types. According to the joint environmental analysis, main QTLs related only to grain filling rate were detected in bin 4.05 and bin 9.04, respectively, and could be used as the key area for corn grain filling research.