Dataset

Nectar spurs (tubular outgrowths of floral organs) have long fascinated biologists. However, given that no model species possess nectar spurs, there is still much to learn about their development. In this study we combined morphological analysis with comparative transcriptomics to gain a global insight into the morphological and molecular basis of spur outgrowth in Linaria. Whole transcriptome sequencing was performed on two closely related species at three key developmental stages, one with a spur (Linaria vulgaris), and one without a spur (Antirrhinum majus). A list of spur-specific genes was selected, on which we performed a gene enrichment analysis. Results from our RNA-seq analysis agreed with our morphological observations. At an early stage, genes associated with cell division are highly expressed in the developing spur. At a later stage in spur development, genes associated with cell wall remodelling are expressed. Our list of spur-specific genes was enriched for genes connected to the plant hormones cytokinin, auxin and gibberellin. We present a global view of the genes involved in spur development in L. vulgaris, and define a suite of genes which are specific to spur development. This work provides a catalogue of candidate genes for spur outgrowth and development in L. vulgaris which can be investigated in future studies. The raw reads and the assembled transcriptomes are available from https://www.ebi.ac.uk/ena with the project accession PRJEB52140. “Lv_trinity-clusters-0.5_withheader.txt” shows which assembled transcripts (trinityID) belong to each gene (corsetID) when clustered by corset (Corset v.1.07; https://github.com/Oshlack/Corset) with distance parameter 0.5. “Lv_trinity-counts-0.5_replicates_renamed.txt” includes the read counts for all the samples per each gene (corsetID). In the sample names, “rA”, “rB”, “rC” and “rD” are the replicate IDs; “V” refers to ventral tissues and “D” refers to dorsal tissues; “Inter” refers to the intermediate stage; “Early” and “Late” refer to the early and late stages in the development.

The counts data should be normalized with software such as DESeq2 or edgeR before being used to evaluate differential expression.