How to analyze ChIP PCR data - (Feb/13/2011 )

Hi guys, i'm new to doing chip and i've been reading up several posts on performing chip and stuff. However, i'm still very confused over the controls needed and analysis required. Will appreciate if someone can simplify the explanation for me.

For e.g., i need to do 1 ip with my antibody of interest. what other controls or PCR reaction will i need to do in addition to the PCR with the primers amplifying the regions i think the protein binds?

I know i'll need an input control, which is an aliquot of the sample right after sonication. So i'll use that to run pcr using the primers amplifying the regions i think the protein binds also. This is a control to know my primers work?

I was told i need to do another ip control using another IgG that will not pull down my protein of interest. For e.g., an antibody against a tag protein that are not inherently present in the cells. Then the same PCR will be done. This is a control for background, and i should expect no pcr product?


Please help me clarify! TIA

Yes, you are correct about input and the mock IP (IP done with pre-immune IgG or with an antibody to something not expressed in your cells). The mock IP may give you a PCR signal depending on what type of beads or plates you use. For instance, agarose beads tend to give a lot more background than magnetic. One important thing is that the IP for your factor of interest at a positive control site (where you expect it to bind at high level) should be several fold above that of the mock.

Another control that is extremely important is to run PCR using one or more negative control primers. These are primers for a region where you don't expect your factor to bind. If you're doing something like pan H3 then this is next to impossible but for most factors it's not hard to find a site like this. True enrichment is the difference you get between the signal at your site of interest and the negative control site(s).

Hi, thank you so much for explaining it to me is such simplified terms. Thanks!! So for the negative control primers, i can just design it anywhere on the genome that i think do not bind to the TF? Like maybe in the exons or intergenic regions?

I would opt for intergenic regions, far from any known enhancers.

Hi, 1 more question. I told my colleage about the negative primers. And he told me that we only need that if we are doing real time pcr. So if i'm just doing normal pcr i don't need it is it?

I'm not sure what your colleague is referring to but you need to run PCR for a negative control site regardless of what kind of PCR you are running. It's the most basic control for ChIP. Much more important than the mock IP.

KPDE on Wed Feb 16 16:19:54 2011 said:

I'm not sure how much to believe my colleague too. He said he just followed what the papers do. According to him, the articles that published the chip data didn't mention about the negative control primers and so he didn't do. Anyways, i guess i'll trust you guys more since you guys are experienced in chip. I'll try to design some negative control primers then. Thanks a lot!

Yeah, I've seen quite a bit of ChIP data in papers that needs proper controls. If it helps, here's the reason for the negative control primers:

In practice there can be wide differences in the amount of "non-specific" pulldown of chromatin by different sera so that there is no ideal mock IP. The mock is, at best, a gross determination of background but isn't a perfect standard against which you can compare all antibodies equally. Instead, the true background determination is a region of the chromatin where your factor doesn't bind. You can then express your results as either a ratio over this background or just compare them as separate data points.

KPDE on Fri Feb 18 17:10:51 2011 said:

I'm taking over a project of my collegue, doing CHIP for histone di-trimethylation of K4 and K9 (markers of heterochromatin and euchromatin) and PCR for a promoter of our gene of interest. This would, regarding to your comment, require negative control primers for genes that are either transcriptionally active (for heterochromatin) or repressed (for euchromatin). But the samples were then treated by azacidine, that makes in general everything transcriptionaly active (well not all, but it's hard to say which gene would be affected and which not). How can this be solved?

Trof,

Wouldn't gene free regions be a good negative control? for H3K9me3 you could check alpha-satellite repeat sequences or transposable elements. In theory by picking any reference gene shouldn't enrichment at your promoters in question be relative? So the changes you see at your reference gene would account for the general effects of chemical treatment?

My impression was that 5-azacitidine was an inhibitor of DNA methyltransferases, so since not all promoters are regulated by DNA methylation I think that a search in the literature would give you some good candidates for control genes not regulated by DNA methylation........good luck!