大豆种质群体蛋白质含量全基因组关联分析及优化组合设计育种效果的研究

Soybean (Glycine max (L.) Merr.) is a major crop with high protein and oil content. Previous soybean breeding in China was mainly focused on oil content improvement and developing soybean cultivars with both high-yield and high-protein content was rarely concerned. Presently, domestic soybean is mainly used for protein processing and consumption of soy products, and high-protein soybean has become an important direction of soybean breeding in China. It was generally understood that soybean was domesticated in ancient China and several thousand years’ selection by farmers formed the unique Chinese soybean germplasm population which accumulated abundant genetic variation across all traits and it is a gene reservoir to the modern soybean breeding. There is abundant genetic variation in protein content (PC) ranging from 29.3% to 52.9% with an average of about 40%. To exploit and utilize the useful genes, a large representative sample composed of 400 accessions from the soybean germplasm population is tested in this study, and the PC is examined by near infrared spectrometer with multiple environments and replicates experiments. Genome-wide association study (GWAS) of PC is conducted with genome-wide SNPLDB marker (Linkage disequilibrium blocks constructed by genome-wide SNP marker) and QTL locations and allele effects are estimated. According to the QTL-allele matrix, breeding potential of all possible parental crosses is predicted. Then the optimal crosses of excellent lines are prepared, the target QTL genotype is examined in progeny to evaluate effectiveness of optimal cross design, marker-assisted selection (MAS) is used to create disruptive high PC (53%) materials (lines) in parallel. In this study, samples are selected from a large germplasm population; the SNPLDB marker can be used to detect multiple alleles (different from SNP); the developed QTL-allele matrix contains important genetic information of population; the cross prediction approach can provided an example for marker-assisted breeding by design; the progeny derived from all crosses can be further used for creating novel materials (lines).

以往全国强调高油育种，高产高蛋白大豆品种稀缺。目前国产大豆主要用于蛋白加工和豆制品消费，高蛋白已成为我国大豆育种重点方向。我国数千年农民选种和豆腐加工形成了独特的大豆种质群体，蛋白质含量变幅大（29.3%-52.9%），蕴藏着丰富的基因资源。本项目采用大豆种质群体代表性大样本，通过多环境重复试验分析群体蛋白质含量的表型变异；利用全基因组SNP构建的连锁不平衡区块标记（SNPLDB）进行全基因组关联分析，查明蛋白质含量的QTL-等位变异（QTL-allele）体系；基于QTL-allele矩阵进行标记辅助优化组合设计，预测育种潜力；配制最优组合并通过标记辅助后裔选择评估设计育种效果，创制突破53%的高蛋白新种质，选育高产高蛋白新品系。本研究取材为种质大群体，所用标记可以检测复等位变异，所获QTL-allele矩阵涵盖群体全基因组遗传信息，所建亲本优配方法可为分子设计育种提供范例。

大豆是人类植物蛋白质最重要的来源，蛋白质含量40%左右，所提供的蛋白质约占全世界蛋白质消费总量的70%。以往全国强调高油育种，高产高蛋白大豆品种稀缺。目前国产大豆主要用于蛋白加工和豆制品消费，提高蛋白质含量已成为我国大豆育种重点方向。根据研究内容和研究目标，本项目获得以下主要结果：（1）以279份江淮流域代表性大豆种质资源构成的群体为材料，完成了多环境下大豆蛋白质含量的精准表型鉴定；利用RAD-seq测序获得全基因组59845个SNP标记，并在此基础上构建了8148个具有复等位变异的SNPLDB标记。利用限制性两阶段全基因组关联分析方法，检测到26个与大豆蛋白质含量显著关联的位点，其中第20号染色体30.99至31.17 Mb (Glyma 1.01) 区域的QTL是控制蛋白质与脂肪含量最显著关联的位点，推测该位点是大豆蛋白质和脂肪含量的关键主效位点。（2）根据江淮流域大豆种质群体蛋白质含量的QTL-等位变异矩阵进行设计育种，配制最优组合，创制了蛋白质含量53%以上的新种质2份。（3）以Linhefenqingdou和Meng 8206杂交衍生的大豆重组自交系群体为试验材料，通过RAD-seq测序技术，构建了SNP标记高密度遗传图谱，采用复合区间作图法和基于混合线性模型的复合区间作图法，挖掘到4个控制蛋白质与脂肪含量的主效和稳定表达的重要QTL，基于RNA测序数据，预测了15个大豆蛋白质与脂肪含量的候选基因。本项目发表论文4篇，培养研究生2名，获得新品种权1项，完成了预期目标。研究结果可为大豆品质性状育种提供理论依据，所用亲本优配方法可为分子设计育种提供技术支撑。