Fig. 7

Preparation of a consensus transcriptome using previously published RNA-seq data. Single transcriptomes were de novo assembled with Trinity using samples with paired-end design, then combined and similar contigs were subsequently clustered with CD-HIT-EST. Poor quality clusters were removed on the basis of cluster length, composition and expression (FPKM). For trying purposes, we also constructed with Trinity a transcriptome on a dataset including reads from all samples with paired-end design