Moderately Low Yield ChIP

[jazz710]

Hello all,

I have been doing qChIP successfully for a while now, and I am trying to make the venture into ChIP-seq. My problem is that I have very small samples (vertebrate embryos) and after I pool and concentrate and do my magic ChIP dance, I get yields of about 2-8ng per ChIP (10% Input: 25-77ng).

The issue I'm having is that those concentrations are sort of in the middle of low-but-usable (~10ng) and the super low concentrations people like to do amplification (LinDA, WGA) or modified library prep (eg. Bernstein's Nano-ChIP-seq).

I feel like there's a gray area in this low (but not micro-low) levels where I'm not sure the best way to proceed.

Also, I remember reading somewhere that some people would spike their IP with input DNA to get the concentrations up. This seems like it would have massive affects on quantification, right?

Thanks for your help. This forum is seriously the greatest.

Bob

[ETHANol]

If you are doing acoustic shearing you can get away with skipping the gel extraction and use AMPure to get rid of adapter-dimer. If you are using mononucleosomal DNA, in my experience you should do, the gel size selection. For gel size selection I recommend a pre-amplification of a few PCR cycles and you can get away with 2 ng but I'd probably try to up it to 5 ng.

[jazz710]

Library Prep in house

So my original plan was do send my ChIP-DNA to Duke for Library Prep and sequencing. In your opinion, with these low values, should I do library prep in house? Is there a kit you recommend for low ChIP levels? I've seen kits thrown around here specifically for that but opinions are also so mixed.

Also, with my ChIP values so low and my Input values ~10x higher, from what I'm reading I should do my library prep with an equal concentration to my ChIP DNA right? My values are:

Treatment 1 IP/Input: 2.4ng/24ng
Treatment 2 IP/Input: 2.2ng/25ng
Treatment 3 IP/Input: 1.99ng/65ng
Treatment 4 IP/Input: 8.0ng/78ng
Treatment 5 IP/Input: 4.79ng/77ng
Treatment 6 IP/Input: 5.98ng/62ng

So it looks like I only need to do a 1% Input in the future to make those numbers more comparable but if I wanted to sequence these, should I be diluting my input to match my IP?

Thanks.

[jazz710]

Bioanalyzer Today

Additionally, my DNA was quantified using the Qubit. Today I am going to run a HS Bioanalyzer ChIP to check for fragmentation level.

I have heard a lot of people here talk about making libraries with these levels of DNA, but only if they ensure that most of their DNA is in the appropriate sequencing range for Illumina (100-500bp pre-adapter). I will report back today about my Bioanalyzer run, but has anyone here had any experience with double fragmentation ChIP-seq? (Efficient Double Fragmentation ChIP-seq Provides Nucleotide Resolution Protein-DNA Binding Profiles, PLOS ONE). They claim to make a workable library from as little as subnanogram amounts of ChIP DNA.

[sequencer2012]

Hello jazz710,

We have similar problems, and get low yields for some of our IPs.
I am curious about your statement regarding spiking the IP with some input DNA (I actually found your post by googling 'IP spiking with input'...).

Do you have the reference or a link to the website where the spiking is mentioned?

Thanks
Alex

[jazz710]

Ugh

Honest to goodness Alex, I wish I did. I know I saw it somewhere (I think on this forum) but I just cannot get the search string right to find it again.

The way I remember it, I basically think they suggested that if you add equal amounts of Input DNA to each ChIP sample, that when you normalized against input it would just 'come out in the wash' so to speak. While I understand the idea, something just didn't sit quite right with me so I wanted to find a citation to support it and can't, to date.

Now I'm just existing in the world of low-DNA ChIP-seq.

[sequencer2012]

Hello jazz710
I have started a new thread here:

http://seqanswers.com/forums/showthread.php?p=89600#post89600

I have had two replies so far, suggesting it would be ok to make a library with 1 ng of DNA, possibly with adjusting the protocol. Someone mentioned spiking with DNA from a different organism, hoping few reads will map to my genome of interest.

Alex

[jazz710]

1ng

I've also committed to just making a library with <10ng. Spiking just made me nervous, because I'm my own bioinformaticist and that sounds like extra filtering to me. I'm actually sending my samples out tomorrow so in a few weeks I'll know if it worked.

Good luck with your stuff!

Bob

[Orange and snow]

It seems as though there may be some hope to making ChIP libraries from 1ng of ChIP DNA. Unfortunately, I need to push that limit a bit further. We work with sperm and where, in some cases, only 1% of the genome remains histone bound. - Then if you are looking for modifications, it may only be 1% of that. I am lucky if I can even get 1ng of ChIP DNA. The company Zymogen says they can routinely make libraries from as little as 50pg, and Diagenode also sells a library prep kit that states being able to use down to 50pg. http://www.diagenode.com/media/catal...kit-manual.pdf -- Has anyone ever tried this kit, or had luck with pg quantities of ChIP DNA for library prep?

[jazz710]

I'm sure not many people are keeping up with this old thread, but here is a publication talking about 'spike-in' DNA for ChIP-seq.

http://genome.cshlp.org/content/early/2014/06/04/gr.168260.113.abstract