Fig. 7

Transcriptional divergence between different ploidy wheat and epigenetic governed stage-specific regulation. a Comparisons of gene expressions between hexaploid wheat (AABBDD) and ancestor (AABB). Pearson correlation index was calculated between gene expression in hexaploid wheat and counterpart in ancestor (left). Genes were clustered into three categories based on Pearson index (right). Transcription data was generated from public data [20]. b Schematic representation of genomic features differences between dysfunction and conserved cluster genes generated in a. Statistics were generated from Additional file 1: Fig. S10b-f. c, Coding sequence similarity and Ka/Ks comparisons among different gene sets in a. Dysfunction genes could be clustered into genes with relaxing and strong negative selection based on the Ka/Ks value. d The proportion of relaxing and strong negative selection genes in dysfunction gene set and TEs insertion frequency difference in the promoters. e Epigenetic modification enrichment at TEs insertion loci. The background generation method is the same as Fig. 2.e. f The expression pattern of representative gene TaUGT91C1 in our data and public data (left) and epigenetic regulations (right). g Comparisons of homoeologs expressions among subgenomes of hexaploid wheat. Each row represents an individual gene A-subgenome of hexaploid wheat, and the genes were ranked by activated order (left). Pearson correlation index was calculated between homoeologs expression in different subgenomes of hexaploid wheat as those in a. h H3K27me3, chromatin accessibility and H3K27ac contribution to gene expression across wheat embryogenesis and key TFs binding strength at different developmental stages. The top 100 variable motifs calculated by chrVAR were used as key TFs. i Stage-specific transcriptional divergence regulation model in wheat embryogenesis. Histone modification H3K27me3 and chromatin accessibility contributed to the early- and late-stage, while H3K27ac and TFs-binding contributed to the mid-stage. Fisher exact test was used for significance calling in d and e (*p <  = 0.05; **p <  = 0.01; ***p <  = 0.001). Fisher’s least significant difference (LSD) was used in h for significance calling