Fig. 7: The gene expression networks of different breeding programs.
a Venn diagram of candidate genes, eGenes regulated by active-eQTGs, and DEGs. b Proportion of candidate genes in the orange blocks of the panel (Fig. 7a) where the Spearman’s rank correlation coefficient between active-eQTGs and eGene has a |SCC| > 0.3 and p value < 0.05. Differences in proportions were assessed using a one-sided two-proportion z-test, with p < 2.2 × 10–16. c Gene network from the orange blocks in Fig. 7a, showing only genes that satisfy p-GWAS < 1 × 10–5, p-TWAS < 1 × 10–4, p-SMR < 1 × 10–5, and |SCC| > 0.6. Nodes represent eGenes or active-eQTGs. Orange edges indicate positive expression correlations (SCC > 0.3 and p value < 0.05) between eGenes and active-eQTGs, whereas the blue edges represent the negative correlation (SCC < -0.3 and p value < 0.05). No edges were drawn between eGenes and active-eQTGs that did not meet these thresholds. To reduce complexity, only eGenes associated with agronomic traits that met all of the following criteria were visualized: GWAS p value < 1 × 10–5, TWAS p value < 1 × 10–4, SMR p value < 1 × 10–5, |SCC| > 0.6, and classification as DEGs. The small regulatory network on the right shows co-expression between the cloned genes (TaISA2 and VRN1-5A) and active-eQTGs. MCCs stands for Modern Chinese Cultivars. USMCs stands for United States Modern Cultivars. LRs stands for landraces. Source data are provided as a Source Data file.
