High duplicates in ChIP samples (lib prep issue?)

[vadoue]

Hi,

I am working with ChIP-seq data with enrichment close to background. Only sample for the histone mark H3K4me3 look fine with well defined peaks. The others samples (H3K9me3/ac) are really bad (similar to input) and have a high rate of duplicates reads (up to 75% for the input).

To what I understand I should have a lower duplicate rates for the input samples as it should be of high diversity, but it is not the case.

I started with 10ng of DNA using the Illumina Truseq ChIP protocol as indicated with the size selection on gel.

The enrichment from ChIP-qPCR was ok, then I am wondering if something could have been wrong during the library prep (particularly to explain to high duplicates rate in the input).

Thanks

[Wallysb01]

Its hard to say what could be going on from what information you gave. Maybe you’re over amplifying, maybe you have far more or far less starting material than you think? Do you have bioanalyzer results from your input, ChIP, input library and ChIP library? How are you quantifying the amount you have? After your size selection maybe you should run another bioanalyzer to be sure you’re still getting the sizes you want and at acceptable concentrations/total yields? I don’t know, this is the first time I’m reading through this kit and it sounds very involved....

I’ve had good luck with the ThruPlex kit from Rubicon, others in our lab have used the Nextera kit for ChIP and they work very well too.

[vadoue]

Thanks for your answer!

You can find attached the sonication profiles for some of the samples (some too concentrated...) as well as the libraries profiles.

I used a Qubit for quantification of the DNA concentrations.

I suspect maybe the step of size selection on the gel, however I just followed the protocol.

Thanks again for any feedback!

[Wallysb01]

Your input has a lot of really big chromatin fragments in it still (7-17, 9-14) or its very short with a odd curve shape overall (8-14). It would seem you still need to optimize the sheering and fixing conditions (it may be over-fixed or under-sheered or both). You may want to get a better input before getting too caught up in the library prep, since this input will negatively effect the ChIP. What you might be finding is that the 7-17 and 9-14 inputs are heavily biased by what regions of chromatin were sheerable and could actually make it into the desired size range. Which could be what results in the high duplicate rates, since there is only so many parts of the genome that you’re selecting for. And for the 8-14 input is hard to tell what is going on. That is either oversheered or the DNA concentration was just too high and the bioanalyzer run couldn’t handle it (see how the peak is up over 500). You may need to dilute your samples to put on the HS DNA chips or just load less.

[vadoue]

The problem came from the library prep. I was doing the PCR after size selection on agarose gel. The problem was that as I had to cut without seeing any band I was selectionning few DNA which was after amplying by PCR. Now I am doing 12 PCR cycle first, and then the size selection on E-Gel where I can actually see the DNA now. I have now a proper % of duplicates.

[huguesparri]

I totally agree with Vadoue: perform the PCR befopre the size selection to lower the amount of duplicates.
And logically, your results will be even better if your sonication gives you a narrow variety of sizes.