ChIP seq DNA starting material

[oliPap]

Hello,
I'm new in the ChIP-seq field and had issues with my first experiment. I started with 5 ng of ChIP-DNA (as measured with Qubit, bioanalyser gave slightly different results)then followed the Illumina protocol with 18 cycles of PCR. When the sequence went back, I found >95% of duplication. The only peaks identified by macs seem to be PCR artifact. The antibody I used was validated by western blot and mass spec analysis after immunoprecipitation. So I know that the IP step is working.

I was considering re-doing the experiment increasing the amount of DNA to start with and maybe decreasing the number of PCR cycles.

Illumina recommends 5<<10 ng of DNA as starting material. Does anybody tried to start with more ie 30-50 ng? Looks to me as the only way to get better results

any help will be greatly appreciated.
many thanks
Olivier

[pmiguel]

Did the bioanalyzer show the ChIP-DNA in a good size range (~100-700bp)?

Personally I don't understand why Illumina's ChipSeq kit uses an agarose gel band extraction. Even if you are pretty good at extracting DNA from agarose, it is good way to lose >90% of the DNA.

--
Phillip

[oliPap]

Hi Phillip,
thank you for your answer
yes the size of the fragment was fine according to the bioanalyser.
Did you succesfully run Chip seq without the gel extraction step?

thanks
Olivier

[pmiguel]

Any chance you could post the bioanalyzer chromatogram of your input DNA? I ask because we always run this as QC, but I never see example chromatograms in ChipSeq kit manuals of this step. (They always have a picture of the library, after PCR amplification, instead.)

We switched to the Bioo ChipSeq kit -- it doesn't use a gel extraction kit. I think that as a result of this it is able to specify 1 ng of input DNA.

That said, the libraries we made from the Illumina ChIPseq kit starting with 5ng of DNA were in the 40-60% duplicate read range. We did once make a set of libraries that had a similar level of duplication to yours -- but in most cases those started with an amount of DNA below the level of detection of the Qubit. (For one ul -- if we had used the entire sample, for fluorimetry we may have been able to detect its concentration. But there was no point doing that...)

--
Phillip

[oliPap]

bioanalyser data here:

http://i63.tinypic.com/fcnxaa.jpg

[oliPap]

sorry, the previous link is not working.
here:http://imgur.com/a/CRUzA

[pmiguel]

sorry, the previous link is not working.
here:http://imgur.com/a/CRUzA

That is input? Wow, that looks great.

--
Phillip