What do people use as sequencing control in ChIP-Seq - (Jul/28/2009 )

Hi, I am attempting to do ChIP-Seq. I have 0 experience in high-throughput data analysis. I am wondering what I need to include as a contol for the sequencing step. I did have IgG control in the ChIP step.

The sequencing itself does not require a control but I concerned that a control is needed when finding the peaks. I have done some research but still not clear about what I should use for sequencing cotrol. Could anybody help me? Thanks a bunch!

I include an input as control for sequencing run vs ChIP samples. This way you sequence the input for the particular cell or tissue type you're using. This paper in PLOS is really interesting. The link is below


PLOS ChIP-SEQ Input Controls

Thank you epi123 for your reply.
Is it that you take some whole chromatin DNA and run it through the library DNA preparation procedure (fragment end-blunting, adaptor modification and size selection and so on)? Do I need a "no antibody" or "no chromatin" control? Thank you a lot!

One more thing: is the PCR-enrichment step necessary? Thanks.

I check my IgG nonspecific background binding by qPCR on a control locus. For example, for H3K9Me3 look at enrichment at SAT2 locus over background. If IgG looks good, then I know that my ChIP is specific. So, to date, I have not examined IgG by sequencing. But I always include the chromatin input as a control and run it through the library construction in parallel with my ChIP sample. Since my samples are low amount of DNA, i do proceed to the PCR amplification step.

Hey,
in my opinion sequencing of the input DNA is the best control you can do as you have enough material to sequence (this will also tell you more about biases you're having in your sample such as CG-bias or "chromatin biases" (don't know the proper term, but open chromatin is more easily fragmented so you naturally see higher signal there even in your input). With IgG controls you hardly pull down any DNA and I would imagine it is hard to make ok sequencing libaries for that. We never do this control.
With the Illumina platform I don't think it's easily possible to omit the PCR-enrichment step, you simply don't have enough ChIP'd DNA.

Hey,
in my opinion sequencing of the input DNA is the best control you can do as you have enough material to sequence (this will also tell you more about biases you're having in your sample such as CG-bias or "chromatin biases" (don't know the proper term, but open chromatin is more easily fragmented so you naturally see higher signal there even in your input). With IgG controls you hardly pull down any DNA and I would imagine it is hard to make ok sequencing libaries for that. We never do this control.
With the Illumina platform I don't think it's easily possible to omit the PCR-enrichment step, you simply don't have enough ChIP'd DNA.

Hello ,

I have to prepare a ChIP Seq library for the first time of my life and I have the same question as TracyDuke. You talk about input DNA as a control, but could you explain what does it means ? I'm sorry I don't speak very well english so I don't understand what is an "input control" and a "bias" ...

And the Illumina's protocol recommends to do a QPCR to quantify the sample's DNA, but if I need do sequence my sample, it's because I don't know it, so how can I do a QPCR ... ? I don't understand this step So can I quantify my samply with a Bioanalyzer 2100 Agilent ?
Is there someone who can help me and give me some answers please ??

Sorry for my bad english

frigo

"Input control" means just sonicated chromatin, without going through the immunoprecipitation part.

you are correct here. If you absolutely have zero knowledge about where the protein might bind in the genome, or what gene(s) is/are regulated by your protein of interest, or you don't have any info about its binding motif, so that you could do some bioinformatic search about where are the potential binding site in the genome, then you can't qPCR at all.

you can be brave about it and just go ahead to sequence it, and hopefully the sequencing result will tell you something useful. However, since you don't know much about this protein in the beginning, there is indicator to tell you whether your ChIP-seq was successful. You have no reference for you to say what you get is signal or noise. So...at the end, you may very well be in a situation that your lab has spent thousands of dollars and several months time and still wondering whether your experiment worked. Trust me on this, this is exactly what our lab is going through right now.

yes you can. This is what we routinely use to check the library quality, i.e. how much DNA (ng) we have for the library and its fragment size.

no worry, your English is fine.

Will

Before even getting to sequencing, you must validate that your ChIP assay alone is working. This is essential and non-negotiable, at least in my lab. To do this, you need a region that you know (or suspect) is bound by your factor of interest and you can test this by qRT-PCR. If you don't know any of these, you better start looking for a positive control region. The major reason for performing this step is to confirm that your antibody is specific and is not just pulling down random junk. Combined with a correct negative control, this is a very powerful way to show that your ChIP technique is good and you are ready to move to sequencing. The same qRT-PCR will be performed before AND after library prep to ensure that your sequencing library is good to go. It's a significant bit of work. (caveat: if you really can't find a positive control region, one solution might be to perform silencing of your factor of interest and proceed straight to sequencing)

The reason why an input control is so important in the sequencing step is because ChIP-seq data is not distributed evenly across the genome and, even in the input sample, you will see clustering of reads around gene regions and especially around promoter areas. If you did not have an input control, you might assign a ChIP peak in your transcription factor sample when, in fact, it is just noise (i.e. false positive). If you have ever compared input and TF ChIP-Seq data side by side, this will be immediately apparent. By comparing the number and distribution of reads between input and your sample, you essentially correct for this background read distribution.

Hope that helps.

Thanks a lot for your answers William Chiu and Dukey, yes it helps !
Thank you