Visualizing ChIPSeqData

[sengoku93]

Hi,

I am currently trying to generate peakcentered ChIP seq heatmaps. However, I'm having trouble generating these heatmaps because the resulting heat map has a single value per gene (one specific color) rather than several values per gene dependent on bin size (therefore multiple colors per row). Below are the multiple commands I tried (all resulting in similar heatmap that is attached, cropped image)(I also attached a representative example of the type of heatmap I'd like to generate, called represenative heatmap):

bamCoverage:

bamCoverage -b input.bam -o output.bw
bamCoverage -b input.bam -o output.bw --centerReads

Compute Matrix:

computeMatrix reference-point -S input.bw -R gencode.v19.annotation.gtf --referencePoint center --sortRegions keep -a 1000 -b 1000 -bs 50 -o output.gz

computeMatrix reference-point -S input.bw -R gencode.v19.annotation.gtf --referencePoint center --sortRegions keep -a 1000 -b 1000 -bs 10 -o output.gz

plotHeatmap:

plotHeatmap -m p63 input.gz -out ouput_v19.png --sortRegions no --colorMap 'RdBu'

Any help would be appreciated! Thank you!

-sengoku

[dpryan79]

What version of deepTools are you using? This behavior is indeed very odd.

Devon

-- Devon Ryan, Ph.D. Email: dpryan@dpryan.com Data Manager/Bioinformatician Max Planck Institute of Immunobiology and Epigenetics Stübeweg 51 79108 Freiburg Germany devon.ryan@dzne.de

On Wed, Mar 28, 2018 at 2:33 AM, sengoku93 notifications@github.com wrote:

Hi,

I am currently trying to generate peakcentered ChIP seq heatmaps. However, I'm having trouble generating these heatmaps because the resulting heat map has a single value per gene (one specific color) rather than several values per gene dependent on bin size (therefore multiple colors per row). Below are the multiple commands I tried (all resulting in similar heatmap that is attached, cropped image)(I also attached a representative example of the type of heatmap I'd like to generate, called represenative heatmap):

bamCoverage:

bamCoverage -b input.bam -o output.bw bamCoverage -b input.bam -o output.bw --centerReads

Compute Matrix:

computeMatrix reference-point -S input.bw -R gencode.v19.annotation.gtf --referencePoint center --sortRegions keep -a 1000 -b 1000 -bs 50 -o output.gz

computeMatrix reference-point -S input.bw -R gencode.v19.annotation.gtf --referencePoint center --sortRegions keep -a 1000 -b 1000 -bs 10 -o output.gz

plotHeatmap:

plotHeatmap -m p63 input.gz -out ouput_v19.png --sortRegions no --colorMap 'RdBu'

Any help would be appreciated! Thank you!

-sengoku

[image: 10bp_v19_cropped] https://user-images.githubusercontent.com/37084775/38001840-1789406a-31fd-11e8-870b-1fa862213aea.png

[image: representative heatmap] https://user-images.githubusercontent.com/37084775/38001996-dd943f62-31fd-11e8-9259-68853b8d15b5.png

— You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub https://github.com/deeptools/deepTools/issues/689, or mute the thread https://github.com/notifications/unsubscribe-auth/AGGkGHp2oQAafPK_IlqUtDK6iGQsrADpks5titpGgaJpZM4S91wn .

[sengoku93]

Hi,

deepTools 3.0.1 installed in conda

[sengoku93]

This is also in python 2.7, I will try using a conda environment with most up to date python version. Thank you

[sengoku93]

I have tried bamCoverage using updated python version 3.5.4. An error occurs:

Segmentation fault: 11

[dpryan79]

I suspect that your conda environment is somehow broken. Can you (likely privately) make your input.bam file available? I can then have a look.

[sengoku93]

can i send via dropbox? I have free acount on github, not sure if I can share files with you/don't how to upload into repository

[dpryan79]

Yes, dropbox would be preferable. You can send the invitation to dpryan79@gmail.com

[sengoku93]

Sent. Thank you for taking your time to help me. Very much appreciated.

-sengoku

[dpryan79]

The plots accurately reflect the distribution of signal in your samples. Your data is mostly uniformly distributed across the genic and intergenic regions. There is a small bit of enrichment at the transcription start site, which you can see with:

computeMatrix reference-point -a 1000 -b 1000 -S input.bw -R gencode.v19.annotation.gtf -o output.mat.gz
plotHeatmap -m output.mat.gz -o output.png

[sengoku93]

Is this for a peak centered heat map? I am not concerned with enrichment near the tss. I want to be able to generate a heatmap with all peaks calls centered at position 0 that is sorted in a manner that I can use to compare it to other heatmaps I'd generate. Would you be able to point me to software in that direction? Below is a reference (Figure 3C) for the type of heat map I'd like to generate.

Jeselsohn, Rinath, et al. "Allele-specific chromatin recruitment and therapeutic vulnerabilities of ESR1 activating mutations." Cancer cell 33.2 (2018): 173-186.

Thank you for helping me.

-sengoku

[dpryan79]

No, but you were inputting genes not peaks. deepTools was likely used to produce the heatmaps you linked to, you just need to feed actual peaks into it rather than genes.

[sengoku93]

Therefore, instead of the gencodev19 I would use a bed file generate from Macs2 peak calling? thank you

[dpryan79]

Correct

Sent from my iPhone

On 29. Mar 2018, at 17:16, sengoku93 notifications@github.com wrote:

Therefore, instead of the gencodev19 I would use a bed file generate from Macs2 peak calling? thank you

— You are receiving this because you modified the open/close state. Reply to this email directly, view it on GitHub, or mute the thread.

[sengoku93]

Does computeMatrix accept bed files for -R input? Thank you!

[dpryan79]

It does

Sent from my iPhone

On 29. Mar 2018, at 17:20, sengoku93 notifications@github.com wrote:

Does computeMatrix accept bed files for -R input? Thank you!

— You are receiving this because you modified the open/close state. Reply to this email directly, view it on GitHub, or mute the thread.