SEQanswers

Go Back   SEQanswers > Applications Forums > Epigenetics



Similar Threads
Thread Thread Starter Forum Replies Last Post
Post PCR Enrichment Gel Selection for ChIP-Seq Daytwa Sample Prep / Library Generation 7 05-10-2012 07:06 AM
ChIP-Seq: ChIP-Array: combinatory analysis of ChIP-seq/chip and microarray gene expre Newsbot! Literature Watch 0 05-19-2011 02:50 AM
ChIP-Seq: ChIP-chip versus ChIP-seq: Lessons for experimental design and data analysi Newsbot! Literature Watch 0 03-02-2011 02:50 AM
ChIP-Seq: ChIPpeakAnno: a Bioconductor package to annotate ChIP-seq and ChIP-chip dat Newsbot! Literature Watch 0 05-13-2010 02:00 AM
PubMed: An integrated software system for analyzing ChIP-chip and ChIP-seq data. Newsbot! Literature Watch 0 11-04-2008 05:03 AM

Reply
 
Thread Tools
Old 03-30-2009, 02:14 AM   #1
Inti
Member
 
Location: Mexico city

Join Date: Jul 2008
Posts: 12
Default No enrichment in ChIP-seq

Hi guys,
I'm doing some ChIP-seq for k4me3 and k27me3 using the GAII (illumina) and although I see a clear difference in my control PCR (comparing on and off genes) I can not see any difference after the sequencing! it seems that the reads are in all the genome with out a peak in the promoter of off genes (in the case of K27me3).
does anyone have an idea of what can be the reason?
Inti is offline   Reply With Quote
Old 06-10-2009, 04:09 PM   #2
picabo
Junior Member
 
Location: San Diego, CA

Join Date: Jun 2009
Posts: 5
Default

We've seen this before. In our case, it was because the ChIP part of the experiment didn't work.
picabo is offline   Reply With Quote
Old 06-30-2009, 12:33 AM   #3
mudshark
Senior Member
 
Location: Munich

Join Date: Jan 2009
Posts: 138
Default

how many reads do you get. how many mappable. how many unique sequences?
you sure your sequencing depth is sufficient?

Quote:
Originally Posted by Inti View Post
Hi guys,
I'm doing some ChIP-seq for k4me3 and k27me3 using the GAII (illumina) and although I see a clear difference in my control PCR (comparing on and off genes) I can not see any difference after the sequencing! it seems that the reads are in all the genome with out a peak in the promoter of off genes (in the case of K27me3).
does anyone have an idea of what can be the reason?
mudshark is offline   Reply With Quote
Old 03-19-2010, 01:37 PM   #4
hon
Junior Member
 
Location: san fran

Join Date: Sep 2009
Posts: 9
Default

Hi Inti, have you figured your problem yet? Recently I encountered the similar problem on one histone marks chip-seq and one transcrioptional factor HA-tagged ChIP-seq. Not many enrichments although QPCR test on ChIP DNA was OK. I wonder if the ChIP library or even the sequencing part would go wrong. I have 15 millions reads but they are all over the genome, much less peaks than I expected for K79me2 marks and RNA Pol II.
hon is offline   Reply With Quote
Old 08-31-2010, 01:40 PM   #5
superduper
Member
 
Location: asdf

Join Date: Aug 2009
Posts: 13
Default

Hey Inti,
Were you working on human cells? What genes did you use to verify your ChIP experiment? Could you share the primer info?
superduper is offline   Reply With Quote
Old 09-01-2010, 08:17 AM   #6
JHU-ChIPmaniac
Member
 
Location: Baltimore

Join Date: Jul 2010
Posts: 21
Default

I have just encountered the same problem. Been doing ChIP for years with the same antibody for a TF and it usually works beautifully. After making the libary, I validated 3-4 promoters that I know are bound by this TF and the enrichment looked pretty good (5-10 fold enrichment of TF IP over total input). However the sequening look bad. I have about 12 million mappable reads but few peaks in the TF IP samples that pass statistical significance. I'm very puzzled. Anyone have any thoughts?
JHU-ChIPmaniac is offline   Reply With Quote
Old 09-01-2010, 09:30 AM   #7
GKM
Member
 
Location: Pasadena, CA

Join Date: May 2009
Posts: 45
Default

Quote:
Originally Posted by JHU-ChIPmaniac View Post
I have just encountered the same problem. Been doing ChIP for years with the same antibody for a TF and it usually works beautifully. After making the libary, I validated 3-4 promoters that I know are bound by this TF and the enrichment looked pretty good (5-10 fold enrichment of TF IP over total input). However the sequening look bad. I have about 12 million mappable reads but few peaks in the TF IP samples that pass statistical significance. I'm very puzzled. Anyone have any thoughts?
That could be down to a number of things going wrong with library building. What is you library complexity? Gel cut size/actual input distribution?

Also, since you say you have worked with this factor for a long time and know it's behavior, this probably isn't the issue, but how exactly are you calculating the qPCR enrichment? People do this in different ways so it is not comparable, but having worked a range of factors, the way I am calculating it, 5-fold enrichment would not be considered a successful ChIP (and it would probably fail after sequencing too)
GKM is offline   Reply With Quote
Old 09-01-2010, 09:34 AM   #8
GKM
Member
 
Location: Pasadena, CA

Join Date: May 2009
Posts: 45
Default

Quote:
Originally Posted by hon View Post
Hi Inti, have you figured your problem yet? Recently I encountered the similar problem on one histone marks chip-seq and one transcrioptional factor HA-tagged ChIP-seq. Not many enrichments although QPCR test on ChIP DNA was OK. I wonder if the ChIP library or even the sequencing part would go wrong. I have 15 millions reads but they are all over the genome, much less peaks than I expected for K79me2 marks and RNA Pol II.
K79me2 and Pol2 aren't your typical point-source TF, K79me2 especially, so what is your expectation for number of peak calls based on? What are you using to call regions?
GKM is offline   Reply With Quote
Old 09-01-2010, 09:57 AM   #9
JHU-ChIPmaniac
Member
 
Location: Baltimore

Join Date: Jul 2010
Posts: 21
Default

Hi GKM,

After ChIP but before library construction, I do qPCR and calculate a %TI for my TF antibody and for an IgG control. For a number of target promoters I see anywhere from 10 to 100's of fold difference between the TF pulldown and IgG. When I make the libraries, I don't make one from the IgG control just from the TI and TF. After library contruction, I just compare the Ct values of the TI vs the TF. For one known target I didn't even see amplification in the TI library but had CT values in the high twenties for the TF library. For a few other targets, I saw anything form 2 ot 4 CT value differences. For one target region and for the negative control region the CT values were similar (less than 0.5) so I thought overall my enrichment after library construction looked good. I have done a number of ChIP-chip experiments with this antibody in several cell types and the results were good. I have to say that I have changed my ChIP protocol for ChIP-seq to a magnetic bead kit. Previously, I used StaphA (pansorbin) to pull down my antibody/chromatin complexes. Generally, I don't think my new protocol works as well as the Staph A cells but I wanted to avoid any Staph A DNA in my final ChIP product.
JHU-ChIPmaniac is offline   Reply With Quote
Old 09-01-2010, 10:00 AM   #10
GKM
Member
 
Location: Pasadena, CA

Join Date: May 2009
Posts: 45
Default

OK, so it seems like a reasonably well characterized antibody. Then library issues are the most likely source of failure. How complex is the library?
GKM is offline   Reply With Quote
Old 09-01-2010, 11:01 AM   #11
JHU-ChIPmaniac
Member
 
Location: Baltimore

Join Date: Jul 2010
Posts: 21
Default

The libraries are made from ChIP products generated from a human cell line. Myc, the TF we study is known to bind to thousands of promoters and maybe 4000-5000 intergenic sites. My chromatin was sheared to an average size of 300-400 bp with a smear ranging from ~150 to 1kb. I use invitrogen Egels to isolate library fragments around 250-300 bp following adaptor ligation. I do 18 rounds of PCR and gel purify the libraries before quantitation by Qubit and bioanalyzer. I have about 15 million reads per library and about 70% of these are mappable but are spread fairly evenly across the genome. I have purchased new antibody and am repeating the ChIP experiment. I'm also throwing in an antibody to H3K4 methyl 3 that I have used with very good success in the past.
JHU-ChIPmaniac is offline   Reply With Quote
Old 09-01-2010, 11:04 AM   #12
JHU-ChIPmaniac
Member
 
Location: Baltimore

Join Date: Jul 2010
Posts: 21
Default

Just wanted to add that this particular Myc antibody has been used by multiple labs for a number of years. There is a boat load of ChIP-chip and ChIP-seq data out there for Myc using this antibody so I'm pretty confident that this is not my problem but maybe the batch I've been using has gone bad.
JHU-ChIPmaniac is offline   Reply With Quote
Old 09-01-2010, 11:09 AM   #13
GKM
Member
 
Location: Pasadena, CA

Join Date: May 2009
Posts: 45
Default

Quote:
Originally Posted by JHU-ChIPmaniac View Post
The libraries are made from ChIP products generated from a human cell line. Myc, the TF we study is known to bind to thousands of promoters and maybe 4000-5000 intergenic sites. My chromatin was sheared to an average size of 300-400 bp with a smear ranging from ~150 to 1kb. I use invitrogen Egels to isolate library fragments around 250-300 bp following adaptor ligation. I do 18 rounds of PCR and gel purify the libraries before quantitation by Qubit and bioanalyzer. I have about 15 million reads per library and about 70% of these are mappable but are spread fairly evenly across the genome. I have purchased new antibody and am repeating the ChIP experiment. I'm also throwing in an antibody to H3K4 methyl 3 that I have used with very good success in the past.
Myc should work indeed. What I was asking though was library complexity, i.e. what portion of your reads have distinct 5' ends. Since many peak callers rely on the asymmetry of reads around the actual binding site, in an extreme version of low complexity library where you have a pile up of the same reads on the + strand and a pile up of the same reads on the - strand, they will think it is an artifact and not call it (that's dependent on the peak caller, and I have only used 3 or 4 of the 30 or so out there, but I think it's true for many of them). If your peaks look reasonably rounded and asymmetric, then this probably isn't the issue, but it's one possible reason.

If you sonicated to average size of 300-400bp and cut at 250-300 after ligation, you were getting the 150-200bp part of the original distribution, which means that you are already at the tail, so it's not at all unconceivable that you ran into complexity problems, especially if you are working with small number of cells.
GKM is offline   Reply With Quote
Old 09-01-2010, 11:14 AM   #14
JHU-ChIPmaniac
Member
 
Location: Baltimore

Join Date: Jul 2010
Posts: 21
Default

Thanks GKM,

I did not analyze the data myself so I will have to ask our informatics person these questions. He is using CisGenome since he is the one who wrote the program. Do you think I should be excising larger fragments after adaptor ligation?

Thanks for all your input.

K
JHU-ChIPmaniac is offline   Reply With Quote
Old 09-01-2010, 11:17 AM   #15
GKM
Member
 
Location: Pasadena, CA

Join Date: May 2009
Posts: 45
Default

It would be better if you sonicated more so that you get it to smaller fragments. Again, we don't know if complexity is the problem, but it is one of the possible underlying problems that could explain a good qPCR enrichment and a bad ChIP-Seq.
GKM is offline   Reply With Quote
Reply

Thread Tools

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off




All times are GMT -8. The time now is 12:36 AM.


Powered by vBulletin® Version 3.8.9
Copyright ©2000 - 2018, vBulletin Solutions, Inc.
Single Sign On provided by vBSSO