[Question]: Usage of Graph Algorithms and Any Slowdowns?

[nv-rliu]

What kind of feature would you like to request?

Other?

Please describe your wishes

Hi there,

I'm interested in understanding if scanpy depends on any graph algorithms from its usage of NetworkX? If so,

• What algorithms are used for what purpose?
• What graph sizes are they being used with?
• Have users experienced any slowdowns or issues with algorithms provided by NetworkX? (Speed, algorithm availability, etc)

Furthermore, would users be interested in accelerated nx algorithms via a GPU backend? This would involve zero code change.

Any insight into this topic would be greatly appreciated! Thank you.

[ilan-gold]

Hi! From what I can tell, we don't use it much aside from small (sometimes outdated) things:

https://github.com/search?q=repo%3Ascverse%2Fscanpy%20networkx&type=code

Do you have any use-cases in mind?

[nv-rliu]

I see NetworkX is used a lot to create and modify graphs. Do you ever run nx algorithms on these said graphs? e.g. PageRank, BC, etc. If so, I have an idea that could greatly speedup your workflows when working with 1M> graph sizes.

[ilan-gold]

From what I can tell, it's mainly here:

scanpy/src/scanpy/tools/_paga.py

Line 402 in 607bb9c

import networkx as nx

but I am not a scanpy algorithms expert despite my familiarity with the overall codebase. @flying-sheep anything come to mind?