Protein–DNA binding maps
Chromatin immunoprecipitation followed by sequencing. Formaldehyde crosslinks the protein to DNA in the cell, the chromatin is sonicated or enzymatically fragmented, and an antibody pulls down the target from millions of cells. Reads are compared with an input made from the same chromatin.
Our ruleSimulations prepare you for the bench but don't replace hands-on practice.
ChIP-seq
Chromatin immunoprecipitation followed by sequencing. Formaldehyde crosslinks the protein to DNA in the cell, the chromatin is sonicated or enzymatically fragmented, and an antibody pulls down the target from millions of cells. Reads are compared with an input made from the same chromatin.
How it differs from the standard method
- Needs many cells (typically 1–10 million per immunoprecipitation) and a ChIP-validated antibody.
- Fragment size, and therefore resolution, is set by the shearing.
- High background compared with CUT&RUN and CUT&Tag, so deeper sequencing (about 20 million usable fragments per replicate for a factor) and an input of similar depth.
- Native ChIP without crosslinking works for histones, which stay on DNA through digestion.
1Capture the DNA a protein sits on, then count it
All three methods answer the same question: where in the genome is this protein, or this histone modification? Each one uses an antibody to single out the target, recovers the stretch of DNA underneath it, and sequences that DNA.
Fix, shear, pull down
Formaldehyde crosslinks proteins to DNA in the cell. The chromatin is sheared to a few hundred base pairs by sonication, the antibody pulls the target out of the whole soup, crosslinks are reversed and the DNA is sequenced beside an input made from the same chromatin.
Cut where the antibody is
Permeabilised nuclei on beads are given the antibody, then protein A/G fused to micrococcal nuclease. Calcium on ice switches the nuclease on; it cuts on both sides of the bound protein and the small fragment diffuses out of the nucleus to be collected. Most of the genome is never touched.
Insert adapters where the antibody is
Protein A/G fused to a Tn5 transposase loaded with sequencing adapters. After high-salt washes remove unbound enzyme, magnesium activates Tn5 and it inserts adapters next to the target. The fragments are ready to amplify: a library from the nuclei in a day.
2What the reads physically are
In ChIP-seq each read pair is one sonicated fragment, so the reads from a site spread over about a fragment length on either side of it: shear to 200–500 bp and the peak is a few hundred base pairs wide; leave fragments near a kilobase and neighbouring sites blur together. In CUT&RUN the fragment is what the nuclease cut out, often under 150 bp for a transcription factor, which gives sharp peaks.
3How it goes wrong
- No control: amplified regions, open chromatin and sticky repeats give more reads in any library. Called against a flat genome average they come out as peaks.
- A poor antibody: weak enrichment and peaks at the sites of whatever else it recognises. A western blot does not validate a ChIP antibody; loss of signal after knockout or knockdown does.
- Too little or too much fixation and shearing: weak capture, broad peaks or damaged epitopes.
- Low complexity: too few cells or too little material gives a library of few distinct molecules. Extra sequencing then only re-reads them, and duplicates left in stack into single-position towers that a Poisson test calls as peaks.
- Untethered enzyme: warm MNase digestion in CUT&RUN and low-salt washes in CUT&Tag let the enzyme cut or tagment accessible chromatin away from the antibody.
- Blacklist regions: satellites and other high-signal regions collect reads in every library and should be removed before calling.
4Reading the result
The fraction of reads in peaks (FRiP) says how much of the library is signal: about 1% or more is acceptable for a transcription-factor ChIP, and CUT&RUN and CUT&Tag routinely put a quarter to over half of their reads in peaks because their background is so low. Look at the duplicate rate, the number of usable fragments, the width of the peaks, whether the strongest peaks carry the expected motif, and whether two biological replicates agree.
Key terms
Practise these as flashcards →Chromatin from the same cells, sheared the same way but not immunoprecipitated; the ChIP-seq control that carries copy number, shearing and mappability biases.
The same procedure with a non-specific antibody of the same species; the usual control in CUT&RUN and CUT&Tag.
A region with significantly more reads than the control predicts, reported with a fold enrichment, p-value and q-value.
Fraction of reads in peaks: how much of the library is signal rather than background.
The number of distinct molecules in a library; when it is low, extra reads are duplicates and add nothing.
A list of genomic regions that give anomalously high signal in nearly every experiment, removed before peak calling.
Protein A/G fused to micrococcal nuclease or to Tn5 transposase: the enzymes that CUT&RUN and CUT&Tag tether to the antibody.
Irreproducible discovery rate: a method that ranks peaks by how consistently two replicates find them.
When to use it
- To find every binding site of a transcription factor or chromatin protein across the genome, or the domains a histone modification covers.
- ChIP-seq when crosslinking is needed to trap a transient interaction or when results must compare with large public ChIP-seq collections.
- CUT&RUN when material is limited (around 10⁵ cells) and sharp, low-background peaks are wanted for a factor or a mark.
- CUT&Tag for histone modifications from very few cells, or when a same-day library is wanted.
Compared with alternatives
Maps open chromatin rather than one protein; no antibody, and often the first experiment before choosing a factor to profile.
Tests whether a protein binds a specific DNA sequence in vitro; use it to confirm a motif, not to map sites in cells.
The pull-down chemistry behind ChIP; worth mastering on its own before chromatin.
ChIP-qPCR checks enrichment at a few known sites cheaply, before or instead of sequencing.
Tests whether a bound region actually regulates transcription, which binding alone does not show.
Sub-techniques
Choosing one here switches the whole page, including the protocol and the lab, to it.
Chromatin immunoprecipitation followed by sequencing. Formaldehyde crosslinks the protein to DNA in the cell, the chromatin is sonicated or enzymatically fragmented, and an antibody pulls down the target from millions of cells. Reads are compared with an input made from the same chromatin.
- Needs many cells (typically 1–10 million per immunoprecipitation) and a ChIP-validated antibody.
- Fragment size, and therefore resolution, is set by the shearing.
- High background compared with CUT&RUN and CUT&Tag, so deeper sequencing (about 20 million usable fragments per replicate for a factor) and an input of similar depth.
- Native ChIP without crosslinking works for histones, which stay on DNA through digestion.