Extended Data Fig. 3: Drivers of variation in transcript abundance.

a, Expression of HvCBF2 and HvCBF4 by tissue and copy number. Each genotype had 3 biological replicates per tissue, except for HOR10350 (caryopsis n = 2), HOR7552 (root n = 2) and HOR84148 (inflorescence n = 2). Significant correlation can be found for HvCBF2 in the coleoptile (r² = 0.85), inflorescence (r² = 0.62), root (r² = 0.5) and shoot (r² = 0.71) and for HvCBF4 in the inflorescence (r² = 0.44) and shoot (r² = 0.49). The boxplot whiskers show minimum and maximum values, the upper bound of the box represents 75th percentile, the lower bound 25 percentile and the centerline the median. b, Principal component cluster analysis of the genes expressed in the inversion. The region contained 2508 expressed PanBaRT genes in at least one of the four tissues. The PCA splits the 201 genotypes into two clusters in every tissue corresponding to those genotypes carrying the inversion versus those carrying the wild type. c, Alignment of chromosome 7H between cultivar RGT Planet and the linear pan-genome generated through PSVCP highlighting the genomic 140 Mb inversion. d, Hi-C interaction matrix of chromosome 7H from the cultivar Morex. The division into A and B compartments emerges as principal component 1 in a PCA of Hi-C interaction frequencies. The A/B compartment boundary on the long arm of 7H is quite distinct and part of the inversion in RGT Planet.