Macs2callpeaks bedgraph/ucsd genome browser

Question: Macs2callpeaks bedgraph/ucsd genome browser

16 months ago by

kempb • 20

kempb • 20 wrote:

Hello,

I am relatively new to Chip-seq analysis and have a question regarding macs2callpeaks. Overall I managed to get peaks from my Chip-Seq data however I am a bit puzzled by the "normalization" of the peaks using a control file. We have a seq on H3K27ac with different conditions and each condition has a separate "Total". The question is; when I used a different total to normalize one of my conditions with a different "total", the bedgraph on ucsd genome browser looked exactly the same as it were if I normalized it with its correct respective "total". I did not expect much variability using the different totals but I did think there would be at least a small difference. I also tried to normalize one of my conditions(treatment) with another condition(as a control) to view it on the genome browser to see if maybe I did something wrong. However when i viewed it on the genome browser there was only a slight change in its peaks compared to the correctly normalized treatment/control. I would have thought that using another condition would brought down the peaks by a significant amount? I am just having trouble understanding how the treatment/control works in MACS2Callpeaks. Do I have to change the parameters? I apologize for the novice question.

Thank you

chip-seq • 488 views

ADD COMMENT • link •

modified 16 months ago by Jennifer Hillman Jackson ♦ 25k • written 16 months ago by kempb • 20

16 months ago by

Jennifer Hillman Jackson ♦ 25k

United States

Jennifer Hillman Jackson ♦ 25k wrote:

Hello,

This may not be such an odd result. It depends on how similar the mixed-up experiments were. If the data is targeting the same factors, the controls from different experiments may be similar enough that resulting peak differences could be very small. (You could compare these directly to see how different they are and where, if curious). That said, in the end, you still do want to use the correctly paired datasets for meaningful analysis.

In very simple terms, the controls are the "normal" background signal. The difference between treatment/condition and the control can be interpreted as the specific signal for the treatment/condition with the background removed. Peaks are MACS' way of representing the signals. The other tools are to help with refining the signal/peaks or differential analysis of those peaks.

Documentation about how MACS handles data and how to use the tools can be found here: https://github.com/taoliu/MACS/wiki. Be sure to see the publication and the Google grouped linked in the page plus there is much more discussion at various resources (message boards like http://biostars.org, other ChIP-seq publications, and the like).

And, Galaxy tutorials for ChIP-seq analysis can be found here: https://galaxyproject.org/learn/

Hopefully, this helps! Jen, Galaxy team

ADD COMMENT • link written 16 months ago by Jennifer Hillman Jackson ♦ 25k

Similar posts • Search »