Fig. 1

NL interactions are largely maintained upon CTCF depletion. A Positioning of CTCF binding sites around LAD borders in various mouse and human cell lines. n shows the number of LAD borders for every cell type (see the “Methods” section for more details). Data are from refs [25, 26, 33,34,35]. B LAD borders are classified as CTCF borders if a CTCF site is positioned within 20 kb outside of the LAD (overlapping with the enriched CTCF density). The percentage of LAD borders that is not within 10 kb of active genes (FPKM > 1) is highlighted. Data are from refs [25, 26, 35,36,37,38,39]. C Average scores for LaminB1 DamID (log₂ ratio over Dam), CTCF and cohesin binding (ChIP-seq coverage), and various publicly available epigenetic data sets around mESC LAD borders, classified by CTCF presence. Solid lines and shaded areas represent the mean signals and 95% confidence intervals of the mean, respectively. Data are from [26, 40,41,42,43,44]. D LaminB1 pA-DamID z-scores along a representative chromosome for mESC parental (PT) and CTCF-AID cells, for a time course of IAA addition in (0, 6, 24, and 96 h). Data tracks are processed in 10-kb bins and averages of n biological replicates. Orange arrows highlight example regions with variable signals following CTCF depletion. E Average LaminB1 z-scores (except 96 h) around LAD borders (from mESC PT pA-DamID data, Additional file 1: Fig. S2E). The solid line and the shaded area represent the mean signal and 95% confidence interval of the mean, respectively. The red arrow highlights the position of the CTCF enrichment. F Quantification of the LaminB1 changes with 0 h at the red arrow in E for all individual LAD borders. A Wilcoxon test was used to test for statistical significance between borders with and without CTCF, followed by Benjamini–Hochberg multiple testing correction. G Cumulative density profile of LAD border positioning as determined by hidden Markov modeling of the CTCF-AID samples, relative to the nearest LAD border from the PT clone. A distance cutoff of 100 kb was used to prevent comparisons between different LADs. H Similar plot to F for CTCF treated with 6 h of IAA, but further segmented by the number of CTCF sites at the LAD border