mims-harvard / decagon

Graph convolutional neural network for multirelational link prediction
http://snap.stanford.edu/decagon
MIT License
448 stars 142 forks source link

Could you tell the meaning of variable n_drugdrug_rel_types in main.py? #5

Closed Msan1995 closed 5 years ago

Msan1995 commented 5 years ago

Could you tell the meaning of variable n_drugdrug_rel_types in main.py?

vidarmehr commented 5 years ago

I think it is the number of polypharmacy side effects (or type of edges in drug-drug network). In the paper, they say there have focused on predicting the 964 polypharmacy side effects that each occurred in at least 500 drug combinations.

Msan1995 commented 5 years ago

I think so too, wanted to confirm. Thank you for the answer

Msan1995 commented 5 years ago

Do you know how to run the code with the original dataset?

vidarmehr commented 5 years ago

No problem!

Not yet. I am working on it. I think we just need to create adjacency matrices for the protein-protein network as well as protein-drug network and create a list of adjacency matrices for each edge type in drug-drug network.