Skip to main content


Fig. 3 | Epigenetics & Chromatin

Fig. 3

From: CGGBP1 regulates CTCF occupancy at repeats

Fig. 3

CTCF occupancy at CTCF motifs and repeats depends on CGGBP1. a Genome browser views showing repeat-masked (RM) CTCF reads distribution in CT, KD and OE samples in a region on chromosome 21. For comparison, UCSC CTCF-binding sites, CTCF reads from prostate epithelial cells (ENCFF098DGZ) and A549 (ENCFF9810JS) are shown. Widespread differences between cell types and similarities between CT, KD and OE can be observed. The motifs discovered in CT, KD and OE are shown on the right. The scales of the Y-axes are different between the samples. b CTCF occupancy at the indicated motifs (a) is low under normal levels of CGGBP1 with KD causing a strong increase. ce Genome browser views showing CTCF reads distribution at repeat-unmasked CTCF peaks exclusive for CT (c) (Y-axis data range 0–50, 27 kb region on chr21 approximately start from 36908 kb), KD (d) (Y-axis data range 0–125, 27 kb region on chr21 approximately start from 6028 kb) and OE (e) (Y-axis data range 0–250, 27 kb region on chr21 approximately from 12718 kb). CTCF peaks and subtle differences in CTCF motifs enriched in each dataset are shaded in gray. f CTCF motif counts and repeat content in repeat-unmasked CTCF peaks were plotted for CT, KD and OE samples, respectively. Between CT and KD, a striking shift in CTCF occupancy from repeats to motifs is seen. g PCA analysis between the ChIP samples to find the patterns of differences in CT, KD and OE datasets due to CGGBP1 levels showed that CT was nearly equidistant and different from KD (majorly PC1) and OE (majorly PC2) with no similarity between KD and OE. h Clustering and correlation heatmap for CT, KD and OE samples show that CT is weakly inversely correlated with KD and OE with no correlation between KD and OE. ik CTCF reads signal was plotted for CT (blue), KD (red) and OE (green) in 1 kb flanks of peaks center. Differential peaks as described below were split into without motifs (left flanks) or with motifs (right flanks) and plotted separately. The peak identities are mentioned on the left Y-axis of each block. The sample from which signal is derived is mentioned in the inset for area under the curve difference calculation. Peak identities are as follows: CT-positive KD-negative peaks (CT–KD peaks), KD-positive CT-negative peaks (KD–CT peaks), CT-positive OE-negative peaks (CT–OE peaks), OE-positive CT-negative peaks (OE–CT peaks), OE-positive KD-negative peaks (OE–KD peaks), KD-positive OE-negative peaks (KD–OE peaks). The labels on the upper left and upper right of each block indicates the area under the curve (AUC) difference calculated by comparing the read signals plotted for KD and CT (KD–CT) for Fig. 3i, OE and CT (OE–CT) for Fig. 3j and KD and OE (KD–OE) for Fig. 3k. AUC differences show that repeat-containing motif-free peaks are specifically enriched in CT the most, and the differences between CT, KD and OE are lowest at motif-containing peaks. The horizontal axes showing the mean signals have the following range of magnitudes: 0–100 units for “with motifs” and 0–60 units for “without motifs”

Back to article page