I have several chromatin modification ChIP data sets and I have RNA-seq expression data. I want to ascribe a metric to how correlated having a particular histone mark (or combo of marks) is with the expression data. I've tried to use deeptools plotCorrelation to accomplish this but I have to use the input subtracted bigwig files for everything instead of the MACS peak data (so it's not quite what I want).
I've seen papers use "concurrence frequencies" between these data types: https://bmcgenomics.biomedcentral.com/articles/10.1186/s12864-019-5609-1 (figure 5) and https://www.researchgate.net/figure/Concurrence-Frequencies-for-Histone-Modifications-DH-Sites-and-Transcript-Regions-The_fig4_235377225 but there is no clear info on how to derive these. Any suggestions or if anyone knows which tools I can use to accomplish this would be much appreciated.
This review Colocalization analyses of genomic elements: approaches, recommendations and challenges may be of use (and references several tools such as StereoGene which may be what you're looking for).
Additionally, if you're interested in combos of histone marks, it could also be useful to integrate them to chromatin states and look for the correlations there (the ChromHMM tool can be used for this).
Thanks! StereoGene does seem to be a way forward to start answering this question. I've run ChromHMM on my data and have states but I again run into the same issue where I have a state assigned to a certain positional region(s) and don't have a good way forward to correlate that with the RNA-seq expression data.