Testing related topic to domibubble

[ASLeonard]

Hey Erik,

I was looking at this because bubble gun did not seem to produce sensible bubbles from pggb output. I used a lot of networkx in a previous life, so figured I would just use their strongly connected component algorithm since that is what you were using here.

I can't believe it seems to work, but just changing nx.MultiGraph to nx.DiGraph in pygfa and then finding SCCs actually worked out the box (and only used ~5 Gb and 1 CPU hour for one 100 Mb chromosome with 8 assemblies). Analysing bubbles seems like it will be a common op soon, so just wanted to share that this worked in case it is of any help.

This is a subgraph with gray nodes and recolouring the two SCCs pink and blue.

Best,
Alex

[urbanslug]

Dear @ASLeonard the project has gone through a lot of changes since you first asked this. The problem raised here should be easily handled by now. Please try it and let me know if you encounter any issues. There's also a preprint of the work here (https://arxiv.org/abs/2410.20932), though it will be revised soon.