Fig. 5.

HFR1 is an important regulator for flower development and extends ABC model. a The known regulatory interactions within the ABC model (AP1, AP2, AP3, PI, and AG). b, c The predicted edges among the five genes by FDCRNet under the condition that HFR1 is lowly expressed and highly expressed respectively. d The correlations among the five genes in ABC model are dynamically changed along with the expression of HFR1. We binned the samples into ten groups by HFR1’s expression level. We then computed the Pearson correlation coefficient (PCC) among the five genes in each group. Given two genes, we binned the PCCs in ten groups into two categories, low and high, by k-means clustering. The high PCC categories are colored in orange and low PCC categories are colored in blue. The p-values of conditional mutual information are computed in each category by the permutation test. e, f Validation of the modulation relationship (HFR1-HAC1-AG) in an independent floral organ data. Consistent with the FDCRNet results, the correlation between HAC1 and AG is strengthened by HFR1’s high expression. g Comparison of H3K27ac signal of AG on flower and leaf tissues