Fig. 9: Structural organization calculated from Hi-C and imaging data.
a Comparison between the mean distance matrix inferred from Hi-C contact map (lower triangle) and the experimental measured average distance matrix (upper triangle) for Chr2 from the cell line IMR90. The distance scale is given on the right. b Direct comparison of pairwise distances, \(\langle {r}_{ij}^{{{{{{{{\rm{Imaging}}}}}}}}}\rangle\) versus \(\langle {r}_{ij}^{{{{{{{{\rm{Hi-C}}}}}}}}}\rangle\). Each dot represents a pair (i, j). Dashed line, with a slope of unity, is a guide to the eye. c Principal component dimension 1 (PC1) for imaging and Hi-C data. The correlation matrix is computed from the connectivity matrix K, and then PCA is performed on the resulting correlation matrix. d Scatter plot of Qk(k = 8) and Fk(k = 8) for 1000 conformations. The conformations are randomly chosen from the total of ~3000 conformations measured in the imaging experiment (green). For Hi-C, 1000 conformations are randomly generated using HIPPS/DIMES (red). e Histogram of kij for A-A, B-B, and A-B. kijs are obtained using Hi-C contact map. f Genomic-distance normalized \(\langle {k}_{ij}(s)\rangle=(1/(N-s))\mathop{\sum }\nolimits_{i < j}^{N}\delta (s-(j-i)){k}_{ij}\) for A-A, B-B, and A-B. 〈kij(s)〉 are shown for s between 0 and 233 Mbp (left) and for between 0 and 10 Mbp (right).
