Add group-wise calling

[j-andrews7]

Get coverage for all peaks in groups (named list of objects/paths), optionally use recurrent peaks across peaks within groups for filtering, perform SE calling.

This could also allow multiple samples to be run at once by passing them in via a list.

Length of signal and peaks lists must match, as should the names of the elements.

[j-andrews7]

Probably use rmspc to generate consensus peak sets for a group.

Ensure it's using the appropriate column for the p-values, as mentioned here.

[j-andrews7]

Implement alternative method for iterative overlap of peaks, a la ArchR. See this discussion also.

This method would be arguably less stringent than rmspc and requires fixed-width peaks, but it may be helpful for SE calling from ATAC or TFs, etc, and would enable some other experimenting.