point of discussion for pairtools module

[sergpolly]

1. chromsizes in the stats reports - yes/no ?!
2. how "rigid" are our cis-ranges (<1kb, 1-2kb, ...>40kb) - should we skip analysis when different ones encountered ?
3. interactions by chromosome pairs (the last heatmap) - remove it for now ? how to display in the future ?

[sergpolly]

more points dug out of our slack:

• %cis in the main table is confusing ( @mimakaev @gfudenberg proposed other way to calculate it - couldn't find now )
• @gfudenberg proposed to add self-circles and dangling ends to main table - are we going to over-interpret FR RF type of pairs for that ? e.g. what's the definition of a dangling end - range <1 kb and FR orientation ? why not < 2kb ? microC vs Hi-C ? should we just name them something else e.g. "short-range FR(dangling end enriched)" ?

[sergpolly]

how else can we recycle pairwise chromosomal counts - ideas ?
right now we try to plot these counts in a form of a heatmap with rows as samples and columns as counts of inter-chromosomal interactions (sorted and truncated at some arbitrary point)

can we condense it into something smaller but still useful - e.g. cis/trans on a per chromosome level ? maybe per-chromosome coverage (this would require chromsizes though) ?

yeah - I guess ideally one would infer copy-number variations and potential transolcations from this ...