ChIPQC and broadPeak input
6
0
Entering edit mode
lindahove ▴ 10
@lindahove-8760
Last seen 6.2 years ago
Denmark

Hi All,

New to this forum - so please correct me if this Q is misplaced.

Is it possible to do ChIP-seq QC on broad peaks (diffuse histone marks and pol2) with ChIPQC?

Maybe by using the PeakCaller="macs".

Found no documentation regarding this

Kind regards,

Linda

chipseq chipqc chip-seq • 1.8k views
ADD COMMENT
2
Entering edit mode
@thomas-carroll-7019
Last seen 17 months ago
United States/New York/The Rockefeller …

hi Linda,

I think the problem lies in the sample names.

Both sample sheets (from the RData objects) contain sample names starting with numerics. Internally names for R objects which start with a numeric have a "X" attached and this then causes problems in ChIPQC plotting functions.

The second "no_consensus" RData object has "#" in the sample names which will also be a problem.

 I would recommend replacing "#" with underscore and adding "Sample" infront of your numeric IDs i.e. 1 become Sample_1.

You can use also use the make.names() function to produce names which will work in ChIPQC. If i try make.names() on a few of your sample names then you can see where the problem is.

> make.names("12")
[1] "X12"
> make.names("7856_8#7")
[1] "X7856_8.7"

If you rerun ChIPQC with updated samplenames you should proceed to a complete report.

 

In the next release I intend to impose valid names for samples in ChIPQC and warn the user that they have been replaced.

 

best,

 

tom

 

 

ADD COMMENT
1
Entering edit mode
lindahove ▴ 10
@lindahove-8760
Last seen 6.2 years ago
Denmark

Thanks for your reply!

I have called peaks with MACS2 - and are now trying to figure out how obtain the QC metrics for h3k4me3, h3k27me3 and pol2 domains. I just keep getting the following error: 

 

> ChIPQCreport(samples)

Saving 7 x 7 in image

Error in `$<-.data.frame`(`*tmp*`, "SCALE_X", value = 1L) : 

  replacement has 1 row, data has 0

Calls: ChIPQCreport ... facet_train_layout -> facet_train_layout.wrap -> $<- -> $<-.data.frame

In addition: Warning message:

In max(panels$ROW) : no non-missing arguments to max; returning -Inf

Execution halted

 

Im not at all aware of what the error message is referring to - other than some data are missing - so was just wondering if it might had something to do with MACS2 .xls peak files missing the summit column? 

Cheers,

Linda

 

ADD COMMENT
0
Entering edit mode
hi Linda, Would it be possible to share the samples R object with me as an RData object? You can email me the object directly or share somewhere else privately if you wish. Then i can take a look at what might be the problem. thank you, tom On Mon, Sep 7, 2015 at 2:55 PM, lindahove [bioc] <noreply@bioconductor.org> wrote: > Activity on a post you are following on support.bioconductor.org > > User lindahove <https: support.bioconductor.org="" u="" 8760=""/> wrote Answer: > ChIPQC and broadPeak input > <https: support.bioconductor.org="" p="" 71881="" #71890="">: > > Thanks for your reply! > > I have called peaks with MACS2 - and are now trying to figure out how > obtain the QC metrics for h3k4me3, h3k27me3 and pol2 domains. I just keep > getting the following error: > > > > > ChIPQCreport(samples) > > Saving 7 x 7 in image > > Error in `$<-.data.frame`(`*tmp*`, "SCALE_X", value = 1L) : > > replacement has 1 row, data has 0 > > Calls: ChIPQCreport ... facet_train_layout -> facet_train_layout.wrap -> > $<- -> $<-.data.frame > > In addition: Warning message: > > In max(panels$ROW) : no non-missing arguments to max; returning -Inf > > Execution halted > > > > Im not at all aware of what the error message is referring to - other than > some data are missing - so was just wondering if it might had something to > do with MACS2 .xls peak files missing the summit column? > > Cheers, > > Linda > > > > ------------------------------ > > Post tags: chipseq, chipqc, chip-seq > > You may reply via email or visit > A: ChIPQC and broadPeak input >
ADD REPLY
0
Entering edit mode

Hi Thomas -

Thanks ! I will share a google drive folder containing the Rdata objects. However what email should I share to?

It should be noted that it seems as if the data is low-quality )-:

Cheers,

Linda

ADD REPLY
0
Entering edit mode

hi Linda,

Could you share with tc.infomatics@gmail.com?

thank you.

tom

 

ADD REPLY
0
Entering edit mode
@thomas-carroll-7019
Last seen 17 months ago
United States/New York/The Rockefeller …

hi Linda,

 

ChIPQC can be used for Broad peaks. It is important to call peaks using an appropriate peak caller when working with more diffuse marks. Peak callers like Sicer or MACs2 can do a pretty good job with Pol2 and less diffuse histone marks.

For comparison there are some Pol2 and histone marks in the ChIPQC course run at BioC 2014 conference (at this link)

http://www.bioconductor.org/help/course-materials/2014/BioC2014/Bioc2014_ChIPQC_Practical.pdf

 

Hope that helps,

 

tom

Thomas Carroll
Head of Bioinformatics
MRC Clinical Sciences Centre, Faculty of Medicine,
Imperial College London,Hammersmith Hospital Campus
Du Cane Road,London,W12 0NN

 

ADD COMMENT
0
Entering edit mode
@thomas-carroll-7019
Last seen 17 months ago
United States/New York/The Rockefeller …

hi Linda,

Could you share with tc.infomatics@gmail.com?

thank you.

tom

ADD COMMENT
0
Entering edit mode
lindahove ▴ 10
@lindahove-8760
Last seen 6.2 years ago
Denmark

Have done (-: Thanks.

I only think That is a problem with the plotCC, no also the plotPeakProfile

Cheers,

 

ADD COMMENT
0
Entering edit mode
lindahove ▴ 10
@lindahove-8760
Last seen 6.2 years ago
Denmark

Hi Tom,

Thanks for your help - all the plots are working now.

I have a question regarding the peak profiles of my controls. They all seems very much enriched ... also more than many of my ip samples  - however these "enrichments" is not spotted in the control pileup files from macs2 when inspecting the controls in IGV (of cause only on selected areas - those which are the most highly reproducible between replicates AND thus would be on the consensus peak list which ChIPQC generates when using consensus=TRUE). Any thoughts on this?

How is R calculating these "peaks"?

thank you so much and kind regards, Linda

 

ADD COMMENT

Login before adding your answer.

Traffic: 312 users visited in the last hour
Help About
FAQ
Access RSS
API
Stats

Use of this site constitutes acceptance of our User Agreement and Privacy Policy.

Powered by the version 2.3.6