用全基因组关联作图和共表达网络分析鉴定油菜种子硫苷含量的候选基因

Identification of Candidate Genes for Seed Glucosinolate Content of Rapeseed by Using Genome-wide Association Mapping and Co-expression Networks Analysis

油菜籽饼粕是畜禽养殖中重要的蛋白原料,但饼粕中的硫苷是一种抗营养物质,食用过多会对禽畜产生毒害,因此挖掘油菜籽粒硫苷含量的候选基因对油菜种子低硫苷育种具有重要现实意义。本研究连续4年种植1个含157份材料的油菜自然群体,结合重测序数据对种子硫苷含量进行全基因组关联分析(GWAS),并对15份低硫苷和15份高硫苷材料进行种子发育早期的转录组测序,通过权重基因共表达网络分析(WGCNA)鉴定种子硫苷含量的候选基因。用GWAS共检测到45个与种子硫苷含量显著相关的SNP,单个位点解释的表型变异为13.5%~23.3%,主要分布在A09、C02和C09染色体的3个区间中,覆盖5个已知的硫苷代谢基因。用WGCNA分析发现高、低硫苷材料之间的2275个差异表达基因,可分为12个基因模块,其中1个模块的基因显著富集在已知的硫苷生物合成途径,对该模块内163个基因的权重分析得到13个候选基因。经检测,GWAS和WGCNA共得到的18个候选基因中,有14个候选基因的表达量与种子硫苷含量显著相关(r=0.376~0.638,P<0.05)。用两种方法鉴定到1个共同的候选基因Bna C02g41790D(基因名MAM1),与该基因连锁的5个SNP构成5种单体型,等位基因效应分析发现,自然群体中63%的材料(99/157)为Hap 5,平均硫苷含量为50.79μmol g^(-1),与另外4种单体型(95.04~110.28μmol g^(-1))存在极显著差异(P<0.01)。本研究结合GWAS和WGCNA两种方法鉴定了油菜种子硫苷含量的候选基因,可为复杂性状候选基因的筛选提供参考。

Seed meal of rapeseed （Brassica napus L.） is a valuable protein source for livestock raising. However, high seed glucosinolates （GSL） content is harmful and toxic to livestock. Therefore, identifying candidate genes of seed GSL content is important in rapeseed breeding for low seed GSL. In this study, a genome-wide association study （GWAS） for seed GSL content was conducted using 157 rapeseed lines grown in four consecutive years. Meanwhile, a weighted gene co-expression network analysis （WGCNA） was carried out in early seed development stage of 15 low and 15 high seed GSL content lines for detecting candidate genes. In total, 45 SNPs found by GWAS significantly associated with seed GSL contents, explaining 13.5%–23.3% of the phenotypic variance per SNP. These SNPs were mainly detected from three intervals on chromosomes A09, C02 and C09, covering five known GSL metabolism genes. A total of 2275 differentially expressed genes （DEGs） were identified by RNA-Seq between rapeseed lines with low and high seed GSL contents. These DEGs were clustered into 12 modules by WGCNA, of which one module （contains 163 DEGs） was mainly enriched in the GSL biosynthetic process. By using a weighted analysis for this module, 13 hub-genes were detected including nine known GSL metabolic genes. Among the 18 candidate genes identified by GWAS and WGCNA, 14 genes showed significant correlation between their expressions and the seed GSL contents （r = 0.376–0.638,P 〈 0.05）. Furthermore, one gene, BnaC02g41790D （MAM1）, was detected by both GWAS and WGCNA. Five haplotypes were formed by five SNPs that significantly linked with BnaC02g41790D, and 63% of the rapeseed population （99/157） were found to carry Hap 5 with significant lower seed GSL contents （an average of 50.79 μmol g-1） compared with those carrying the other four haplotypes （95.04–110.28 μmol g-1）. By GWAS and WGCNA, our study not only identified the candidate genes for seed GSL content of rapeseed, but also provided a guidance for digging candidate genes for other complex traits.