ChIP-Seq

ChIP-Seq (Chromatin Immunoprecipitation followed by Massively Parallel DNA Sequencing) is a molecular and bioinformatics methodology used to map the genomic binding coordinates of DNA-associated proteins—such as transcription factors, RNA Polymerase II, CTCF insulator proteins, and chromatin remodeling enzymes—as well as covalent histone post-translational modifications in living cells. Introduced independently in 2007 by four research groups led by David S. Johnson, Ali Mortazavi, Richard M. Myers, and Barbara J. Wold; Artem Barski and Keji Zhao; Tarjei S. Mikkelsen, Eric S. Lander, and Bradley E. Bernstein; and Gordon Robertson and Steven J. M. Jones, ChIP-Seq replaced hybridization-based ChIP-on-chip microarrays by coupling antibody immunoprecipitation directly with massively parallel sequencing.

By sequencing 20 to 50 million immunoprecipitated chromatin fragments per sample across the 3.055-gigabase human genome, ChIP-Seq resolves transcription factor binding footprints to within ±10 to 30 base pairs and charts the combinatorial histone code across promoters, enhancers, and heterochromatin domains—forming the experimental foundation of the ENCODE (Encyclopedia of DNA Elements) and NIH Roadmap Epigenomics reference consortia.

Formaldehyde Cross-Linking, Chromatin Shearing, and Antibody Immunoprecipitation

A standard cross-linked ChIP-Seq (X-ChIP) experiment proceeds through five sequential biochemical stages:

1. Reversible Formaldehyde Cross-Linking: Intact living cells (10⁶ to 10⁷ cells) or homogenized tissues are incubated with 1% formaldehyde (H₂CO) for 8 to 15 minutes at room temperature. Because formaldehyde is a small, dipolar molecule with a ~2 Å reaction radius, it diffuses across cell and nuclear membranes to form reversible covalent methylene bridges (–CH₂–) between primary exocyclic amino groups of DNA nucleobases (adenine, guanine, cytosine) and the ε-amino groups of lysine or guanidino groups of arginine on bound transcription factors and core nucleosome histones (H2A, H2B, H3, H4). Addition of 125 mM glycine quenches unreacted H₂CO.
2. Cell Lysis and Chromatin Fragmentation: Isolated nuclei are lysed and the cross-linked chromatin is sheared into 150 to 500 bp DNA fragments using focused acoustic sonication (Covaris or Bioruptor). Alternatively, in Native ChIP (N-ChIP)—used for stably bound histone modifications without formaldehyde fixation—chromatin is digested enzymatically with Staphylococcus aureus micrococcal nuclease (MNase), a Ca²⁺-dependent endo-exonuclease that cleaves linker DNA between nucleosomes to yield ~147-bp mononucleosome core particles.
3. Antibody Immunoprecipitation: Cleared chromatin is incubated overnight at 4 °C with a ChIP-grade monoclonal or polyclonal antibody recognizing the target transcription factor epitope or modified histone tail residue (e.g., anti-H3K27ac or anti-CTCF). Antibody-chromatin complexes are captured onto Protein A or Protein G paramagnetic beads while unbound genomic DNA is washed away under high-salt and detergent stringency.
4. Cross-Link Reversal and Input Control Processing: Eluted complexes (alongside an un-immunoprecipitated Input DNA control or IgG mock control taken prior to step 3) are incubated at 65 °C for 4–16 hours in 200 mM NaCl with RNase A and Proteinase K to hydrolyze the formaldehyde methylene bridges and digest proteins.
5. Library Preparation and Sequencing: Purified ChIP and Input DNA fragments (1–50 ng) are end-repaired, ligated to indexed adapters, amplified by 8–12 cycles of PCR, and sequenced to 20–30 million single-end or paired-end reads for focal factors (or 40–60 million reads for diffuse histone marks).

Epigenomic Chromatin Signatures and Tethered-Nuclease Alternatives

Bimodal Strand Shift Modeling and Dynamic Poisson Peak Calling (MACS2)

After mapping ChIP-Seq and Input control reads to the reference genome (via Bowtie2 or BWA-MEM2) and filtering out PCR duplicates and ENCODE Blacklist regions (repetitive pericentromeric and satellite regions that produce artificial signal pileups), enriched binding loci are identified using Model-based Analysis of ChIP-Seq (MACS2 / MACS3), developed by Yong Zhang and X. Shirley Liu in 2008.

Because short-read sequencers sequence only the 5′ ends (first 50–150 bp) of each 200–300 bp immunoprecipitated double-stranded DNA fragment flanking a transcription factor bound at coordinate x₀, reads mapping to the forward (+) strand accumulate upstream of x₀ while reads mapping to the reverse (−) strand accumulate downstream of x₀—producing a bimodal peak distribution separated by the mean fragment length d. MACS2 empirically estimates d from paired peaks, shifts all (+) and (−) reads toward the center by d / 2 (or uses paired-end fragment coordinates), and models background read counts using a dynamic Poisson distribution:

λ_local = max(λ_BG, λ_1k, λ_5k, λ_10k)

evaluated across the Input DNA control in sliding 1-kb, 5-kb, and 10-kb windows to neutralize local open-chromatin sonication bias and copy-number variations. Downstream de novo motif discovery suites (MEME-ChIP, HOMER) scan the 100–200 bp sequences beneath peak summits to reconstruct Position Weight Matrices (PWMs) of the bound transcription factor.

Tethered-Nuclease Alternatives: CUT&RUN and CUT&Tag

Because conventional ChIP-Seq shears the 3-Gb genome and relies on immunoprecipitation pull-down, background noise from non-specifically bound genomic fragments requires 10⁶–10⁷ cells and 30 million reads per sample. To eliminate formaldehyde cross-linking, sonication, and bulk chromatin solubilization, Steven Henikoff and colleagues introduced CUT&RUN (Cleavage Under Targets and Release Using Nuclease, 2017, by Peter Skene and Steven Henikoff) and CUT&Tag (Cleavage Under Targets and Tagmentation, 2019, by Hatice Kaya-Okur and Steven Henikoff).

In CUT&RUN and CUT&Tag, unfixed permeabilized cells or nuclei are immobilized on concanavalin A magnetic beads and incubated with a primary antibody against the target protein, followed by a chimeric fusion protein combining Protein A/G with Micrococcal Nuclease (pA-MNase) in CUT&RUN or Protein A/G with hyperactive Tn5 transposase (pA-Tn5) loaded with sequencing adapters in CUT&Tag. Upon activation with Ca²⁺ (CUT&RUN) or Mg²⁺ (CUT&Tag), the antibody-tethered enzyme cleaves or tagments only the DNA directly flanking the bound target protein in situ, leaving the remaining 99.9% of the genome intact inside the nucleus. This signal-to-noise improvement reduces required sequencing depth to 3 to 8 million reads and enables profiling of 100 to 1,000 primary cells or single cells.

Why is an Input DNA control (rather than an IgG mock control alone) mandatory for accurate ChIP-Seq peak calling?

Acoustic sonication does not fracture genomic DNA uniformly: accessible euchromatin at active gene promoters and enhancers shears more readily into 150–300 bp sequenceable fragments than compact heterochromatin, creating false-positive read pileups at open promoters even without antibody enrichment ("sonication bias"). In addition, cancer genomes harbor focal gene amplifications (copy-number gains) and repetitive regions that generate proportional spikes in baseline read depth. Sequencing 1% to 5% of sheared chromatin prior to antibody addition (Input DNA) captures these local chromatin accessibility and copy-number biases at sufficient depth, whereas a non-specific IgG mock immunoprecipitation recovers so little DNA (<50 pg) that PCR amplification yields low-complexity library duplicates unsuitable for local background modeling.

How does Spike-In Normalization (ChIP-Rx) detect global genomic losses or gains of a histone modification?

Standard ChIP-Seq normalizes read counts to mapped reads per million (RPM / CPM), which assumes that the amount of bound protein or histone mark per cell remains invariant across conditions. However, if a pharmacological inhibitor (such as an EZH2 methyltransferase inhibitor) erases 90% of H3K27me3 across the genome, standard RPM normalization scales up the remaining 10% residual peaks to fill the 30-million-read sequencing budget, masking the global depletion. ChIP with Reference Exogenous Genome (ChIP-Rx) spikes a fixed proportion of Drosophila melanogaster S2 chromatin into every mammalian sample prior to immunoprecipitation, using the invariant Drosophila ChIP read count as an internal quantitative ruler.

What is the difference between Narrow Peak calling and Extended Domain calling in ChIP-Seq?

Sequence-specific transcription factors (such as CTCF, MYC, or ERα) and promoter-proximal histone marks (H3K4me3, H3K27ac) bind discrete focal loci spanning 150 to 1,000 bp, producing high-amplitude narrow peaks with a well-defined central summit suitable for motif discovery. Conversely, repressive Polycomb domains (H3K27me3), constitutive heterochromatin (H3K9me3), and co-transcriptional elongation marks (H3K36me3) coat contiguous arrays of hundreds of nucleosomes across 10 kilobases to several megabases at lower per-base read density. Calling diffuse marks with narrow-peak parameters fragments contiguous domains into thousands of discontinuous sub-peaks; MACS2 --broad and SICER link adjacent enriched windows into contiguous domains.

1. Johnson, D. S., Mortazavi, A., Myers, R. M., & Wold, B. (2007). "Genome-Wide Mapping of in Vivo Protein-DNA Interactions." Science, 316(5830), 1497–1502. https://doi.org/10.1126/science.1141319
2. Barski, A., Cuddapah, S., Cui, K., Roh, T. Y., Schones, D. E., Wang, Z., Wei, G., Chepelev, I., & Zhao, K. (2007). "High-Resolution Profiling of Histone Methylations in the Human Genome." Cell, 129(4), 823–837. https://doi.org/10.1016/j.cell.2007.05.009
3. Zhang, Y., Liu, T., Meyer, C. A., Eeckhoute, J., Johnson, D. S., Bernstein, B. E., Nusbaum, C., Myers, R. M., Brown, M., Li, W., & Liu, X. S. (2008). "Model-Based Analysis of ChIP-Seq (MACS)." Genome Biology, 9(9), R137. https://doi.org/10.1186/gb-2008-9-9-r137
4. Skene, P. J., & Henikoff, S. (2017). "An Efficient Targeted Nuclease Strategy for High-Resolution Mapping of DNA Binding Sites." eLife, 6, e21856. https://doi.org/10.7554/eLife.21856