中文摘要：

本研究对国家大豆改良中心种质创新计划中296份稳定品系进行多环境田间试验获得蛋白质含量数据,并利用227对PAV标记和93对SSR标记基因型数据进行蛋白质QTL关联分析。结果表明,供试材料蛋白质含量变幅为34.50%-53.84%,平均遗传率达92.57%,品系与环境间存在极显著互作。6个环境下与蛋白质含量显著（P≤0.05）和极显著（P≤0.01）关联的位点分别有139、44个,分布于大豆基因组的20个染色体,其中在2个以上环境被重复检测到的有8个。以平均蛋白质含量〉45%为标准筛选出64个新品系,分别来自33个杂交组合。基于AMMI模型的品种稳定性分析发现高蛋白含量材料Di变幅达0.08-1.27,来自菜豆5号/泰兴黑豆、南农73-932/早熟18等组合的新品系在6个环境表现稳定;基于分子标记可将高蛋白品系聚为6类,具有共同亲本的材料多聚在一起,且高值材料（〉47%）和低值材料的杂交组合也被区分开。对5个在多环境都检测到的关联位点进行分析,发现64份高蛋白材料分别含0-4个优异等位变异,材料间优异等位变异的构成存在差异。聚合不同优异等位变异可能创造更高蛋白质含量的新品系。

英文摘要：

Improvement of protein content is a major breeding goal for high quality soybean. In the present study,296 new breeding lines developed from a germplasm enhancement plan at National Center for Soybean Improvement were planted under six different environments,and their seed protein contents were investigated. The genotype data of 227 PAV and 93 SSR markers were used to detect QTL for protein content and to reveal the genetic relationship of the elite lines. The results indicated that the protein content of the sample ranged from 34. 50% to53. 84%,and its average heritability value reached 92. 57%. ANOVA showed that the interactions between genotypes（ lines） and environments（ the combinations of year and location） were significant. A total of 139（ P ≤0. 05） and 44（ P≤0. 01） markers（ QTL） distributed over all the 20 chromosomes were identified to be significantly associated with protein content under the six environments. Among them,8 loci were repeatedly detected inat least two environments. According to the criterion that the average protein content should be above 45. 00%,64 new lines come from 33 crosses were selected. Stability analysis based on AMMI model pointed out that Diof the high protein content lines ranged from 0. 08 to 1. 27 and the lines from the crosses of Caidou 5 / TXHD,NN73-932 / Zaoshu 18 behaved stable under the six environments. Based on the molecule maker data,the high protein content lines could be classified into 6 groups and most lines from the same crosses were classified as the same group. The crosses with high protein content（ 47%） lines could be distinguished as specific groups. Analysis of the five marker loci detected in more than one condition indicated that there were 0- 4 elite alleles in 64 high protein content lines and differences of the elite allele distribution existed in some of these lines. It is feasible to create new lines with higher seed protein content by pyramiding different elite alleles.