Fig. 3: FOXA1 interacting with co-factors, driving breast cancer susceptibility.

a The numbers of genetic variants (multiplied by 1000) that are occupied by specific TFs or co-occupied by two TFs, stratified by FOXA1. b Barplots show the association strengths for genetic variants occupied by specific TFs occupancy. The “+” and “−” symbols underneath the x axis represents genetic variants with and without occupancy of specific TFs. P-value on the top shows the significance of the interaction between FOXA1 and TF pairs. c Heatmap shows the TF binding signals for the flanking regions of genetic variants co-occupied by FOXA1 and TF pairs. The heavy color in each panel denotes the binding signal of the TF of interest. d Barplots show the proportion of the detected FOXA1 motif based on motif enrichment analysis for flanking regions of genetic variants (i.e., ±25, 50, and 100 bp). Disruptive motif models are illustrated in the right panel. e Heatmap shows a correlation pattern based on gene-expression data of 22 TFs in the GTEx data. The color change from blue to red denotes the correlation coefficients changes from −0.5 to 1. f Barplots show association (regression coefficients) of TF score with breast cancer risk in BCAC data. g Barplots show an association of TF score with breast cancer risk in BCAC data, stratified by FOXA1. In these figures, the two-sided nominal P-values and \(\triangle {\chi }^{2}\) were derived from the generalized mixed models and the error bars denote 95% confidence intervals.