转录组+代谢组专题 | 转录组+代谢组关联分析内容及分析思路

那么转录组和代谢组的数据如何通过生物信息学的方法有机联系起来，而不是两个数据的简单堆叠呢？

转录组与代谢组的关联分析方法主要包括以下内容：

1.基于KEGG通路的注释和富集分析

2.基于皮尔逊的相关性分析

3.基于降维的模型构建确定关联关系

We examined the expression of structural genes of the anthocyanin biosynthesis pathway, as well as transcription factor gene families (MYB, bHLH, and WD40) known to regulate anthocyanin biosynthesis in other species. The expression of ZjANS (Zj.jz022481171) was 219.96-fold higher at DAP30 than at DAP110, a result consistent with the high content of cyanidin in young jujube fruit skin. Meanwhile, metabolomic analysis showed that cyanidin-3-O-rutinoside and peonidin-3,5-O-diglucoside were the major anthocyanin components of jujube fruit skin. Cyanidin-3-O-rutinoside is the product of glycation modification, and peonidin-3,5-O-diglucoside is the product of methylation. Therefore, genes encoding glycosyltransferases and methyltransferases are likely to participate in the biosynthesis of these anthocyanins. In the transcriptomic data, expression of the UDP-glucuronosyltrans ferase gene ZjUGT79B1 (Zj.jz027401002) was higher at DAP30 than at DAP110. By contrast, expression of the methyltransferase gene ZjCCoAOMT (Zj.jz041523079) increased with maturity, suggesting that it may be involved in the methylation of other flavonoids. In addition, two MYB homologues ZjMYB113 (Zj.jz044581035) and ZjMYB5 (Zj.jz005919060)  three bHLH homologues ZjGL3a (Zj.jz029235018), ZjTT8 (Zj.jz001627021), and ZjGL3b (Zj.jz022481028), and three WD40 homologues ZjWDR1 (Zj.jz028347005), ZjWDR2 (Zj.jz010873005), and ZjWDR3 (Zj.jz043343182) were also identified .（Shi Q et al. 2020. Journal of Agricultural and Food Chemistry, 68(51).）

The DEGs and DEMs in 4FADvs4WD, such as caffeoyl-CoA Omethyltransferase, Quercetin and 3,7-Di-O-methylquercetin, were found to be related to flavonoids biosynthesis (Fig. 4a). While, except for flavonoids biosynthesis (e.g. Chalcone synthase 3, Chalcone synthase1, flavanone 3-hydroxylase, flavonoid 3′,5′-hydroxylase, Quercetin, Kaempferol, Myricetin, Phloretin and Myricetin) (Fig. 4b), the co-expression network of DEGs and DEMs in 8FADvs8WD were mainly enriched in ascorbate metabolism (e.g. L-ascorbate oxidase, aldehyde dehydrogenase family 3, GDP-mannose-3′,5′-epimerase perakine reductas, L-ascorbate) (Fig. 4c) and glutathione metabolism (e.g. glucose-6-phosphate 1-dehydrogenase 1, glutathione S-transferase, CYS-GLY and Glutathione reduced form) (Fig. 4d). The results showed that the FA could modulate the co-expression of DEGs and DEMs related to ascorbate metabolism, glutathione metabolism and flavonoids biosynthesis during drought stress.（Sun Jianhao et al.BMC Genomics, 2020.）

基于降维的模型构建确定关联关系

Integration analysis was performed using from one to three joint components and one orthogonal component per dataset (transcriptomic and metabolomics). The model used with three joint components explained more than 90% of the total variation of the transcriptome (R2t) and metabolome (R2m) (Figure3A). Almost 100% of the modeled transcriptome and metabolome variations were considered joint variation because they could be explained by variation in the complementary model [metabolomics (RmCORR/ R2m) and transcriptomic (RtCORR/R2t) models, respectively]. Based on the loading coefficients thresholds calculated after 1000 permutations, we identified 10 transcripts and 16 metabolites as having the most influence on the model (Figure 3B).