Summarising ChIP-seq and controls needed

Question

As part of my study I have to present on a method of discerning binding sites of proteins to DNA, for which I have looked at ChIP-seq. This is a brief overview of what I have so far, and was wondering if there are any particularly glaring omissions that I have made so far.

Cells treated with formaldehyde to cross-link proteins such as transcription >factors (i.e. HIF) covalently to DNA and incubated for 2-30 minutes. After >incubation glycine is added to terminate the cross linking and dissipate the >formaldehyde.

Cells are lysed, and sonication (application of sound waves to agitate a >sample) is used to fragment the DNA into sizes between 200-1000 base pairs.

HIF–DNA complexes are selectively immunoprecipitated by using antibodies >specific to the HIF protein of interest, complexed with a matrix of either >agrose or superparamagnetic beads.

HIF-DNA-antibody-matrix complexes are isolated through the use of either a >centrifuge or magnetic field (depending on the matrix employed). Several >washings take place to further purify the sample.

Cross-linking of the protein to the DNA is reversed by way of heating, >separating the DNA and protein-antibody-matrix complex. Proteinase K is used to >digest proteins, leaving only the DNA.

Oligonucleotide adaptors (primers) are then added to the small stretches of DNA >that were bound to the HIF to enable rapid simultaneous sequencing.

DNA sequences are then mapped to the reference genome to determine the binding loci.

I would also like to discuss controls that are used in these procedures. So far I have found limited mention of negative controls, positive and negative control loci, non-template control and positive control antibody, but am having trouble discerning what any of this actually is referring to.

I would seriously appreciate any help available on this topic!

Answer 1

Good experiment! ChIP-seq is good for a whole genome look at transcription factor binding. KEY POINT: The data generated are only as good as the antibody. Considerable effort needs to taken to prove antibody is specific for only the protein of interest. It is for this reason, that some TFs still haven't been characterized by ChIP-seq.

As far as controls for your experiment:

The breaking of DNA from sonication isn't uniform from experiment to experiment, and sonication will preferentially break at some regions of open chromatin more often. Controls for this:

1. Isolate DNA from cells that have been x-linked and fragmented under same conditions but don't use antibodies to IP. This is "Input" DNA.
2. You can conduct a "mock" ChIP using a control AB that reacts with some non important non-nuclear protein, such as IgG.

Then sequence these samples to same depth as the Immunoprecipitated DNA.

Here is a good paper that is a little dated, but still relevent: Go to Control Sample Section

Here is a link to the guidelines for ChIP-seq for ENCODE PDF

Last but not least, if you are working on a much smaller scale, don't forget about things like DNA footprinting and EMSA to identify the ACTUAL binding site.