Plot only the genome-wide trans contacts in a Hi-C matrix (no cis contacts)

Question

[kalavattam]

Using Hi-C contact matrices (for example, those in .cool or .hic formats), I'm interested to plot all of the trans contacts in a genome while masking or otherwise excluding cis contacts. I know this has been done before in one form or another (for example, we see this kind of analysis with observed/expected matrices in Figure 2 of Lieberman-Aiden, Van Berkum et al., Science 2009), but I'm not really hitting on a strategy to do it. Do any of the members of the community have suggestions?

A related idea I've had is to generate a matrices of only trans contacts—perhaps from matrices that have been balanced using all contacts. From there, I could do some plotting. Does anyone thoughts or suggestions related to that?

Thanks!