Fig. 2

Long-range absolute contact probability is specifically modulated for each cell type. a Left, a schematic representation of pairwise distance measurements between consecutive and nonconsecutive borders, with color code and positions as in Fig. 1b. Right, normalized Hi-C counts vs. microscopy absolute contact probability for consecutive and nonconsecutive domain borders for embryo and S2 cells. Solid black and red lines represent exponential and power-law fits, respectively. Matrix resolution = 10 kb. N for microscopy pairwise measurements is provided in Supplementary Fig. 1f–h. N = 2 for Hi-C data, from at least three and two biological replicates, respectively. b Absolute contact probability vs. mean physical distance between probes for consecutive and nonconsecutive TAD borders (filled circles). Solid lines represent power-law fittings with the scaling exponent described in Supplementary Fig. 2b. Triangles represent measurements within TADs. c Matrix of relative frequency of normalized Hi-C counts for late embryo vs. S2 cells for chromosome 2L. Contact frequency ratio is color coded according to scale bar. Matrix resolution = 50 kb. N = 4, biological replicates. d Log–log plot of normalized Hi-C counts between TAD borders vs. genomic distance for embryo and S2 cells. Solid lines represent the average contact frequency for randomly chosen positions in the genome. Matrix resolution = 10 kb. N = 2, biological replicates. e,f Log–log plot of the mean physical distance vs. genomic length for (e) active and (f) inactive/repressed chromatin domains for different cell types. Mean distance values were normalized by the pre-exponential factor from the power-law fit of each data set (Supplementary Fig. 2d, e). Solid lines show the power-law fits, with the scaling exponent β shown in the panel. Circles and triangles are depicted as described in panel 2b. Error bars represent SEM. N > 140 for each data point, from more than three biological replicates (Supplementary Fig. 1)