Fig. 6

NL interactions are independent from H3K27me3 in mESCs. A Average H3K27me3 signal around mESC LAD borders, classified by CTCF presence. Human reads from spiked-in HEK293T cells were used to calibrate the mouse ChIP-seq data (Additional file 1: Fig. S10A, see the “Methods” section). Solid lines and shaded areas represent the mean signals and 95% confidence intervals of the mean, respectively. B Profile of LaminB1 z-scores along a representative genomic locus for CTCF-AID mESCs treated with DMSO and the H3K27me3 inhibitors GSK126 and EED226 (3 days) and with DMSO or IAA to induce CTCF depletion (last 24 h). Data are processed in 10-kb bins and are averages of n biological replicates. The pink arrows highlight example regions that lose LaminB1 signal upon CTCF depletion. Calibrated H3K27me3 ChIP-seq signals before and after CTCF depletion are added for comparison. C Average LaminB1 z-scores around LAD borders are shown for the samples in panel B, as described in Fig. 1E. D Quantification of the LaminB1 change outside LAD borders, as described in Fig. 1F. E Heatmap showing Spearman correlation coefficients between the IAA-induced LAD differences for untreated CTCF-AID cells (control) and for cells treated with DMSO and the H3K27me3 inhibitors