Fig. 3: The strength of crosslinking may play a pivotal role in determining the balance between reliability and sensitivity in compartment identification.

a The principal component analysis (PCA) results from the “PC1” values of all libraries in K562 (left) and GM12878 (right) cells. Only the first (PC1) and second (PC2) principal components are displayed, along with their respective ratios of explained variances. b The DNA methylation rates, gene expression levels, H3K27me3 ChIP-seq signals, and ATAC-seq signals in regions designated as compartment B (left in each pair) and compartment A (right in each pair) under each crosslinking condition in genome regions with inconsistent compartment assignments in K562 cells. The p-values of Mann-Whitney U tests comparing values in each pair are indicated below (after \(-{\log }_{10}\) transformation). Group-wise sample sizes (from left to right): DNA methylation: 1501, 903, 1142, 1262, 1243, 1161, 1247, 1157, 959, 1445; gene expression: 779, 675, 797, 657, 572, 882, 650, 804, 564, 890; H3K27me3 signal: 903, 1501, 1262, 1142, 1161, 1243, 1157, 1247, 1445, 959; ATAC-seq signal: 903, 1501, 1262, 1142, 1161, 1243, 1157, 1247, 1445, 959. c The chromosome-wise Spearman correlation coefficient (SCC) values of the “PC1” vectors determining compartment assignments between replicates of all crosslinking conditions in K562 cells. Chromosomes 1 through 22 and X are marked from left to right in each condition. d The percentage of the genome with consistent compartments in biological replicates for each crosslinking condition. e The genome-wide saddle plots demonstrating the separation of compartments A and B in K562 cells.