Abstract
Lily is a popular flower around the world not only because of its elegant appearance, but also due to its appealing scent. Little is known about the regulation of the volatile compound biosynthesis in lily flower scent. Here, we conducted an approach combining two-dimensional analysis and weighted gene co-expression network analysis (WGCNA) to explore candidate genes regulating flower scent production. In the approach, changes of flower volatile emissions and corresponding gene expression profiles at four flower developmental stages and four circadian times were both captured by GC-MS and RNA-seq methods. By overlapping differentially-expressed genes (DEGs) that responded to flower scent changes in flower development and circadian rhythm, 3,426 DEGs were initially identified to be candidates for flower scent production, of which 1,270 were predicted as transcriptional factors (TFs). The DEGs were further correlated to individual flower volatiles by WGCNA. Finally, 37, 41 and 90 genes were identified as candidate TFs likely regulating terpenoids, phenylpropanoids and fatty acid derivatives productions, respectively. Moreover, by WGCNA several genes related to auxin, gibberellins and ABC transporter were revealed to be responsible for flower scent production. Thus, this strategy provides an important foundation for future studies on the molecular mechanisms involved in floral scent production.