Fig. 4.

Transcriptional regulation by AR enhancers. a Cumulative distribution function correlating the distance (bp) of each enhancer class to the promoters of androgen-upregulated genes. b Chromatin loops at ARBS (VCaP ChIA-PET) were overlapped with the enhancer classifications to identify those ARBS that looped to a promoter (± 5 kb from the TSS) of an androgen-upregulated gene. c The violin plots shows the number of AR chromatin interactions in each enhancer class. d Schematic representation of AR ChIA-PET data transformed into graph network. e Calculation of the relative interaction frequency in the graph network between androgen-upregulated gene promoters (Up), androgen-downregulated gene promoters (Down), and each ARBS enhancer class. f With the interaction graph network, the betweenness centrality in the largest connected graph was calculated for each enhancer class (ns p > 0.05, ***p < 10⁻⁹). g At AR-regulated genes, individual ARBS CRE (induced, inactive, constitutive) were inhibited with CRISPRi (blue) in LNCaP cells to determine their impact on AR transcription. Gene expression was quantified by qPCR and normalized to non-targeting gRNA controls (white bar). The TSS of each gene was also targeted with CRISPRi as a positive control (black bar) (3 biological replicates ± SD;***p < 10⁻⁹). h Androgen-induced expression of genes regulated by only inducible enhancers (n = 102) or both inducible enhancers and other ARBS (n = 58) (**p < 10⁻⁴). i Evolutionary conservation from 100 vertebrate species of different ARBS enhancer classes compared to genomic regions with ARE motif but no AR binding (n = 2783)