HI-C Terminology

Cited from Paper "A 3D Map of the Human Genome at Kilobase Resolution Reveals Principles of Chromatin Looping"

'We define the ‘‘matrix resolution’’ of a Hi-C map as the locus size used to construct a particular contact matrix and the ‘‘map resolution’’ as the smallest locus size such that 80% of loci have at least 1,000 contacts. The map resolution is meant to reflect the finest scale at which one can reliably discern local features.'

'We began by probing the 3D partitioning of the genome. In our earlier experiments at 1 Mb map resolution (Lieberman-Aiden et al., 2009), we saw large squares of enhanced contact frequency tiling the diagonal of the contact matrices. These squares partitioned the genome into 5–20 Mb intervals, which we call ‘‘megadomains.’’We also found that individual 1 Mb loci could be assigned to one of two long-range contact patterns, which we called compartments A and B, with loci in the same compartment showing more frequent interaction. Megadomains—and the associated squares along the diagonal—arise when all of the 1 Mb loci in an interval exhibit the same genome-wide contact pattern. Compartment A is highly enriched for open chromatin; compartment B is enriched for closed chromatin.'

'Two of the five interaction patterns are correlated with loci in compartment A (Figure S4E). We label the loci exhibiting these patterns as belonging to subcompartments A1 and A2. Both A1 and A2 are gene dense, have highly expressed genes, harbor activating chromatin marks such as H3K36me3, H3K79me2, H3K27ac, and H3K4me1 and are depleted at the nuclear lamina and at nucleolus-associated domains (NADs) (Figures 2D, 2E, and S4I; Table S3). While both A1 and A2 exhibit early replication times, A1 finishes replicating at the beginning of S phase, whereas A2 continues replicating into the middle of S phase. A2 is more strongly associated with the presence of H3K9me3 than A1, has lower GC content, and contains longer genes (2.4-fold).'

'The other three interaction patterns (labeled B1, B2, and B3) are correlated with loci in compartment B (Figure S4E) and show very different properties. Subcompartment B1 correlates positively with H3K27me3 and negatively with H3K36me3, suggestive of facultative heterochromatin (Figures 2D and 2E). Replication of this subcompartment peaks during the middle of S phase. Subcompartments B2 and B3 tend to lack all of the above-noted marks and do not replicate until the end of S phase (see Figure 2D). Subcompartment B2 includes 62% of pericentromeric heterochromatin (3.8-fold enrichment) and is enriched at the nuclear lamina (1.8-fold) and at NADs (4.6-fold). Subcompartment B3 is enriched at the nuclear lamina (1.6-fold), but strongly depleted at NADs (76-fold).'

'Upon closer visual examination, we noticed the presence of a sixth pattern on chromosome 19 (Figure 2F). Our genome-wide clustering algorithm missed this pattern because it spans only 11 Mb, or 0.3% of the genome. When we repeated the algorithm on chromosome 19 alone, the additional pattern was detected. Because this sixth pattern correlates with the Compartment B pattern, we labeled it B4. Subcompartment B4 comprises a handful of regions, each of which contains many KRAB-ZNF superfamily genes. (B4 contains 130 of the 278 KRAB-ZNF genes in the genome, a 65-fold enrichment). As noted in previous studies (Vogel et al., 2006; Hahn et al., 2011), these regions exhibit a highly distinctive chromatin pattern, with strong enrichment for both activating chromatin marks, such as H3K36me3, and heterochromatin-associated marks, such as H3K9me3 and H4K20me3.'