Abstract: 【Objective】To assess the dynamic and underlying genetic control of glucosinolate (GS) content in seeds and silique walls at different developing stages. We conducted Associative Transcriptomics (AT) in a panel of European winter oilseed rape. 【Method】The glucosinolate content in developing and mature seeds and silique wall samples from the rapeseed panel (n=113) was determined by HPLC. Using mRNA-seq data derived from the young leaves of B. napus, functional genotypes were inferred for the panel, which included both sequence variation (i.e. Single Nucleotide Polymorphism, SNP) and transcript abundance (Gene Expression Markers, GEM). A total number of 355 536 SNPs and transcript abundance were scored across the rapeseed panel using a reference sequence of 116 098 ordered coding DNA sequence (CDS) gene models for resolving the genetic factors controlling GS variation. All significantly associated markers were further traced back to a CDS gene model, of which the function can be inferred by sequence blast and prediction using public Brassica and Arabidopsis databases. 【Result】The GS content in seeds and silique walls 15 days after pollination (DAP) was very low, with an average of 7.58 μmol?g^-1 (1.69-20.45 μmol?g^-1) and 4.81 μmol?g^-1 (1.47-25.23 μmol?g^-1), respectively. The GS content for mature seeds, seeds at 25 DAP and silique walls at 25 DAP varied from 8.87 to 111.83 μmol?g^-1, 2.17 to 147.21 μmol?g^-1 and 0.73 to 130.77 μmol?g^-1, respectively. The observed extensive phenotypic variations for GS content at these two stages made them suitable for AT analysis. In total, 256,397 informative SNPs with minor allele frequency (MAF) ＞ 0.01 and 53,889 GEMs with mean expression level＞0.4 were plotted after AT analysis. With these SNPs markers, a total number of 167, 158 and 3 SNPs were shown to be significantly associated (-log₁₀P＞6.71) with GS content of mature seeds, seeds at 25 DAP, and silique walls at 25 DAP, respectively. Among these, 5 association peaks distributing across chromosome A2, A9, C2, C7 and C9 were found for mature seeds. Meanwhile, 127, 16 and 24 GEMs were significantly (-log₁₀P＞6.03) associated with GS content of mature seeds, seeds at 25 DAP, and silique walls at 25 DAP, respectively, which also unambiguously defined 5 association peaks in A2, A8, A9, C2, and C9. The association peaks in A9, C2 and C9 were common for both mature seeds and silique walls at 25 DAP. Based on the Arabidopsis homologous gene annotations, 25 genes derived from the significant SNPs or GEMs were inferred and predicted to be involved in GS metabolic pathways. The other genes derived from AT analysis, though not directly involved in GS metabolism, were classified as transcription factors, factors responding to stimulus or involved in cellular process, catalytic activity, and thus were also predicted to play an important role in GS accumulation in the different tissues of Brassica napus. 【Conclusion】The GS content in silique walls was found to be high and positively correlated with that of seeds, indicating that silique wall was an important organ for GS synthesis or translocation. A total number of 328 SNPs and 144 GEMs were found to be significantly associated with GS content for different tissues. From these, 25 genes were predicted to be directly involved in GS accumulation or translocation, and 73 genes with uncharacterized function were also inferred.

Key words: Brassica napus, glucosinolates, dynamics, Associative Transcriptomics, SNP, GEM