Comparing two CUT&RUN (similar to ChipSeq) generated peaks

Question

[susan.kuo.run]

Hi all,

I ran a CUT&RUN experiment targeting a transcription factor in a normal control cell line and its shRNA-knockdown. I would like to prove that the antibody for the transcription factor I used worked. I used MACS2 to peak call this data and have generated 2 narrowPeak files respectively. I would like to order the peaks and select the highest peaks in the normal control dataset and see if correlational peaks are lower in the knockdown, perhaps even by generating a heatmap for qualitative confirmation.

I am attempting to use computeMatrix from deeptools. Does anybody have experience attempting a similar experiment?

Thank you!

Answer 1

[babalyigit]

I am using this guideline for ChIPseq analysis. After the peak calling step by MACS2, this tutorial can be also used for CUT&RUN analysis. Here is the link:

https://hbctraining.github.io/Intro-to-ChIPseq/lessons/10_data_visualization.html

[Friederike]

The HBC Training link gives most details you'll need. The basic steps are:

1. obtain normalized coverage files of your CUT&RUN samples (bigWig files)
2. use these bigWig files together with the MACS BED files to generate the matrices needed for heatmaps and coverage plots.

If you want to force a specific order of the regions in the BED file (e.g. from "highest" to "lowest"), you'll have to turn off the clustering for the heatmap. Or you can sort the regions by the mean expression values for either sample (as computed by deepTools rather than MACS2); check out the optional arguments for plotHeatmap: https://deeptools.readthedocs.io/en/develop/content/tools/plotHeatmap.html#Optional%20arguments