The best way to correlate some genome wide signal to histone modification signal

Question

[moxu]

Suppose I have a ChIP-seq track, and I want to see how well it correlates to histone modification tracks such as H3k4me3, what would be my best course of processing? bam => bed => bigWig and use bigWigCorrelate? Or should I take the top percentage (say 5%) of each track, use 0/1 to simplify the signal （i.e. to dichotomize) and do a chi-sq, for instance?

Thank you!

Answer 1

[Sinji]

bigWigCorrelate sounds like a good option.

Answer 2

[vivekbhr]

I can tell some deepTools solutions :

a) genome-wide : Run multiBAMSummary of both files, followed by plotCorrelation.

b) on some regions : Do peak calling from your ChIP-Seq data, take top peaks (bed) and do computematrix using histonemarks (bigWig), then plotHeatmap.