Association of genes in different regulatory categories with the proxies of pleiotropic constraint in nelumbo, Arabidopsis and Gossypium hirsutum. A, Differences in expression breadth for genes with different regulatory categories in Nelumbo with 19 different tissues and Arabidopsis with 21 different tissues (Kruskal–Wallis test, *** P value < 0.001 after Benjamini–Hochberg correction for multiple comparisons). The violin plot shows the summarizing ranges of genes with different regulatory categories in different tissues (violin contours). B, Average expression level (log-transformed FPKMs) for genes with different regulatory categories in Nelumbo with 19 different tissues and Arabidopsis with 21 different tissues (Kruskal–Wallis test, *** adjusted P value < 0.001 for multiple comparisons). Horizontal solid lines represent median values, and dashed lines represent lower/upper quartiles. C, The ratio of the observed number of WGD genes (duplicate genes retained from ancient WGD) to the expected number of WGD genes. The proportion of WGD genes that have both putative paralogs retained for genes in different regulatory categories (colored bars). D, The ratio of the observed number of orphan genes (genes that lack detectable similarity to genes in other species) to the expected number of orphan genes. The proportion of orphan genes under cis- and/or trans-regulation (colored bars). In c and d, the black dashed line indicates the ratio of 1, i.e. observed numbers = expected numbers. Ratios > 1 indicate there are more genes than expected, while ratios < 1 indicate there are fewer genes than expected (χ2 test of independence, *: P value < 0.05, **: P value < 0.01, ***: P value < 0.001). Corresponding violin plots and bar graphs for other stages and other duplications are shown in Supplemental Figures 17–19.