Hi-C and Chromatin Conformation Capture

Hi-C (High-throughput Chromosome Conformation Capture) is a genome-wide molecular assay that quantifies the three-dimensional spatial proximity of all chromatin loci within intact cell nuclei by coupling proximity-based DNA ligation with paired-end high-throughput sequencing. Within a eukaryotic nucleus of 5 to 10 μm diameter, approximately 2 meters of linear genomic DNA must be folded across four orders of magnitude while preserving dynamic access for transcription factors, dna-replication machinery, and DNA repair complexes. Evolved from targeted Chromosome Conformation Capture (3C) pioneered by Job Dekker in 2002, Hi-C introduced by Erez Lieberman-Aiden and Dekker in 2009 replaced locus-specific PCR primers with biotinylated nucleotide fill-in and streptavidin enrichment, enabling unbiased all-versus-all contact matrices across entire karyotypes.

Quantitative analysis of Hi-C contact matrices has established that interphase chromosomes are organized into a multi-scale hierarchy: whole chromosome territories (>100 Mb), transcriptionally active and inactive A/B compartments (1–10 Mb), ctcf-and-tad-insulation Topologically Associating Domains (TADs) (100 kb–1 Mb), and fine-scale enhancer–promoter chromatin loops (10–200 kb). Beyond functional epigenetics, because intra-chromosomal contact probability decays predictably as an inverse power-law function of linear genomic distance, Hi-C and sequence-agnostic variants such as Omni-C have become essential scaffolding technologies for assembling fragmented contigs into telomere-to-telomere chromosome-scale genome references.

Cross-Linking, Digestion, and Biotinylated Proximity Ligation

The canonical in situ Hi-C protocol proceeds through five sequential biochemical steps inside intact, permeabilized nuclei:

1. Formaldehyde Fixation: Cells are treated with 1% to 3% formaldehyde (CH₂O), which forms covalent methylene bridges between primary amino groups of DNA-bound proteins (histones, CTCF, cohesin) and adjacent DNA bases lying within 2 to 10 Å, freezing native 3D chromatin contacts in place.
2. Chromatin Fragmentation: Cross-linked DNA is digested with a 6-cutter restriction endonuclease (such as HindIII, recognizing A^AGCTT, ~4 kb resolution) or a 4-cutter enzyme (such as MboI or DpnII, recognizing ^GATC, ~1 kb resolution) to create 5′ overhangs.
3. Biotinylated End Fill-In: Klenow DNA polymerase I fills in the 5′ restriction overhangs using dNTPs supplemented with a biotinylated nucleotide (commonly biotin-14-dATP or biotin-14-dCTP), generating blunt, biotin-tagged DNA termini.
4. In Situ Proximity Ligation: T4 DNA ligase is added under dilute intra-nuclear conditions, favoring intramolecular ligation between cross-linked DNA fragments that occupy the same protein complex over random intermolecular collisions.
5. Cross-Link Reversal and Streptavidin Capture: Proteinase K digestion at 65 °C reverses formaldehyde cross-links, DNA is sheared to 300–500 bp, and chimeric ligation junctions containing the internal biotin tag are captured on streptavidin-coated magnetic beads for paired-end Illumina sequencing.

Evolution from 3C to Micro-C and DNase Omni-C

To overcome sequence bias and resolution ceilings inherent to restriction enzymes, the 3C technological family has diversified across scope and fragmentation chemistry. While 3C (one-versus-one), 4C (one-versus-all via inverse PCR), and 5C (many-versus-many via multiplexed ligation-mediated amplification) interrogate targeted loci, Micro-C replaces restriction enzymes with micrococcal nuclease (MNase), digesting linker DNA between individual nucleosomes to achieve 147–200 bp single-nucleosome resolution. Conversely, Omni-C digests fixed chromatin with sequence-independent endonuclease DNase I, eliminating restriction-site dropout in AT-rich or repetitive satellite arrays and yielding uniform coverage ideal for de novo chromosome scaffolding and structural variant detection.

Power-Law Contact Decay and the Fractal Globule

When paired-end Hi-C reads are mapped to a reference genome, binned into genomic intervals of width Δ*x*, and normalized via iterative correction and eigenvector decomposition (ICE) or Knight–Ruiz (KR) matrix balancing to remove GC-content and mappability biases, the resulting symmetric matrix M_i,j records interaction frequencies between loci i and j. Across distances s = |*i* − *j*| spanning 500 kb to 7 Mb, average intra-chromosomal contact probability P(s) follows an inverse power-law scaling relationship:

P(s) ∝ s^−γ

In interphase mammalian nuclei, the scaling exponent γ equals approximately 1.0 to 1.08, matching theoretical predictions for a non-equilibrium fractal (crumpled) globule polymer state (γ = 1.0) rather than an entangled equilibrium globule (γ = 1.5, which occurs only up to saturation) or an ideal random walk polymer (γ = 1.5). A fractal globule packs contiguous linear segments into unknotted local domains, allowing rapid local unfolding for gene activation without topological entanglement. During prometaphase chromosome condensation, condenser I and II complexes reorganize the polymer, shifting the contact curve to feature a prominent peak around 10–12 Mb reflecting helical arrays of nested chromatin loops.

A/B Compartments and Loop-Extrusion TADs

Subtracting the expected distance decay P(s) from the observed contact matrix and computing the Pearson correlation matrix reveals a striking plaid ("checkerboard") pattern across each chromosome. Decomposition of this correlation matrix via Principal Component Analysis (PCA) extracts the first principal eigenvector (PC1), which partitions the genome into multi-megabase A compartments (positive PC1, gene-dense, early-replicating euchromatin enriched in H3K27ac and H3K36me3 histone-modification marks) and B compartments (negative PC1, gene-poor, late-replicating heterochromatin enriched in H3K9me3 and nuclear lamina-associated domains).

At sub-megabase scales (100 kb to 1 Mb), Hi-C matrices display self-interacting diagonal squares termed Topologically Associating Domains (TADs), frequently punctuated at their apices by intense corner peaks ("dots"). These domains are generated independently of compartmental phase separation by ATP-dependent cohesin loop extrusion: the ring-shaped cohesin complex loads onto chromatin and extrudes a progressively larger DNA loop at 0.5–1.5 kb s⁻¹ until it is stalled on both strands by convergently oriented CTCF zinc-finger proteins bound to asymmetric 19-bp recognition motifs.

3D-DNA and YaHS Assembly Algorithms

Because intra-chromosomal contact frequency exceeds inter-chromosomal contact frequency by more than an order of magnitude—and decays monotonically with physical separation along the chromosome arm—Hi-C and Omni-C data provide long-range linkage information across 10 kb to >200 Mb, far surpassing the 15–100 kb span of PacBio HiFi or Oxford Nanopore long reads. Scaffolding algorithms such as 3D-DNA, SALSA2, and YaHS (Yet another Hi-C Scaffolder) ingest draft long-read contigs alongside aligned Hi-C read pairs to solve a global optimization problem in three stages:

1. Misassembly Breaking: Contigs showing abrupt depletions in local diagonal contact coverage are cleaved at chimeric join points.
2. Chromosome Clustering and Ordering: Contigs are partitioned into linkage groups corresponding to the species' haploid karyotype n, and ordered along each chromosome axis to maximize the monotonic decay of P(s) along the main matrix diagonal.
3. Orientation Assignment: Because the two ends of each contig exhibit asymmetric contact intensities with adjacent neighbors, graph-based orientation solvers orient contigs in 5′→3′ sense, producing chromosome-scale scaffolds across complex genomes such as the 52-chromosome, 4.6 Gb karyotype of the trapdoor spider Mygalomorphae (Aptostichus stephencolberti).

How does Omni-C improve upon restriction-enzyme Hi-C for genome assembly?

Standard Hi-C relies on restriction enzymes that cut at specific 4-bp or 6-bp motifs, leaving blind spots in AT-rich pericentromeric satellites, telomeres, or sequences lacking recognition sites. Omni-C replaces restriction enzymes with sequence-agnostic DNase I endonuclease digestion, generating uniform genome-wide read coverage (~80% of the genome covered within 2-fold of the mean) and increasing the density of informative long-range contacts at contig boundaries for chromosome scaffolding.

What is the difference between A/B compartments and Topologically Associating Domains (TADs)?

A/B compartments (1–10 Mb) arise from affinity-driven phase separation of euchromatin (A) and heterochromatin (B) and persist—or even strengthen—when the cohesin complex is acutely depleted. Conversely, TADs (100 kb–1 Mb) are formed by active ATP-dependent cohesin loop extrusion halted by convergent CTCF boundaries; depleting cohesin (RAD21) erases TADs worldwide while leaving A/B compartments intact, whereas depleting CTCF blurs TAD boundaries and causes aberrant enhancer–promoter cross-talk.

How can Hi-C contact maps detect balanced chromosomal translocations and inversions?

In a Hi-C contact heatmap mapped against a reference genome, structural variants produce characteristic off-diagonal geometric signatures because sequences that are distant in the linear reference are brought into direct 3D contact in the rearranged chromosome. A balanced chromosomal inversion generates a symmetric "bowtie" or "butterfly" pattern across the breakpoint coordinates, while a reciprocal inter-chromosomal translocation produces sharp high-intensity L-shaped wings in the inter-chromosomal contact block.

1. Lieberman-Aiden, E., van Berkum, N. L., Williams, L., Imakaev, M., Ragoczy, T., Telling, A., et al. (2009). "Comprehensive mapping of long-range interactions reveals folding principles of the human genome." Science, 326(5950), 289–293. https://doi.org/10.1126/science.1181369
2. Rao, S. S. P., Huntley, M. H., Durand, N. C., Stamenova, E. K., Bochkov, I. D., Robinson, J. T., et al. (2014). "A 3D map of the human genome at kilobase resolution reveals principles of chromatin looping." Cell, 159(7), 1665–1680. https://doi.org/10.1016/j.cell.2014.11.021
3. Dudchenko, O., Batra, S. S., Omer, A. D., Nyquist, S. K., Hoeger, M., Durand, N. C., et al. (2017). "De novo assembly of the Aedes aegypti genome using Hi-C yields chromosome-length scaffolds." Science, 356(6333), 92–95. https://doi.org/10.1126/science.aal3327
4. Zhou, C., McCarthy, S. A., & Durbin, R. (2023). "YaHS: yet another Hi-C scaffolder." Bioinformatics, 39(1), btac808. https://doi.org/10.1093/bioinformatics/btac808