What does network merge actually *do*

[Vincent Laufer]

Hi all,

I have seen many posts in this forums about how to USE network merge (difference). However, I have it working on my machine, that's not my issue - I want to know what exactly the 3 options *do*. 

The documentation 3.7.2 only gives a description of how to use the GUI - not what it does behind the scenes if executed. Do I need to go directly to the code Github or is there a technical description somewhere of what these operations do. 

If the answer is, it doesn't perform anything mathematically, but it simply removes edges, then in that case the question I'm asking is how exactly does it work. If I have 2 networks and then both contain an edge, and in one case the edge is a negative coexpression and in the other case its a positive coexpression, then how is that done? Etc.

Thank you very much for your help.

[Scooter Morris]

Hi Vincent,

   Roughly speaking, network merge proceeds in two passes.  First, it merges the nodes using the parameters you've given it.  Second, it will merge the edges between the mapped nodes.  If the edges have the same interaction, it will attempt to merge the *values* for that edges.  If the edges have different interactions, it will create two edges.

-- scooter