I come to you to ask for suggestions in order to do this kind of representation.
The initial situation: I have a set of specific genomic regions (for example: promoters, exons, intergenic regions, etc...) I want to characterize using some ChIPseq data. I do not want to cluster data on this set of regions, I want to see which marks are under/over represented compared to an input or total histone.
So I thought it could be interesting to represent this as a heatmap where each square has also the score in order to show overrepresentation or underrepresentation of TF or Histone Marks in certain genomic regions. For this analysis I was thinking to compute the Zscore based on reads per sample and then divide each mark (TF or Histone Mark) by the input or total desired histone; then using R try to plot the figure.
So far I looked at the literature and could not find anything like this, but I remember once I saw some paper figure showing this kind of heatmap I describe but unfortunately I cannot remember the publication.
- Do you know any paper where this has been shown?
- Or even better, could you give me any suggestion/correction to the strategy on how to do this?
- Is also possible to calculate the statistical significance of each enrichment nor depletion?
Thanks for your help once more.