Incorporating the edge weights of my network into Hierarchical HotNet analysis

[Jeffrey Brabec]

I want to do network propagation using your Hierarchical HotNet pipeline. I'm using a network from HumanBase which is weighted and I also have gene "heats". I'm relatively new to python so reading through construct_similarity_matrix.py has been a little difficult but it seems like between lines 48 and 54 deals with the weighted graphs parameter. How is the directed-ness of the my network built into the output file of this function and how do the downstream scripts handle weighted networks? 

Any advice or direction you could provide would be much appreciated!

Jeff

[Matthew Reyna]

Hi Jeff,

From your question, I assume that your network is both weighted and directed. The network propagation process generalizes to weighted and direct networks, and the similarity matrix that is defined for the stationary distribution of the network propagation process captures this information. Please add your edge weights as the third column in your edge list file, and please use the -d/--directed flag for the construct_similarity_matrix.py script. The resulting command should look something like this:

    python construct_similarity_matrix.py -i input_edge_list.tsv -d -o similarity_matrix.h5

All of the following commands should work without any changes, but I would use gene score/"heat" permutations instead of network permutations.

Best,
Matt