I've made an application that at this point ranks all combinations of drug pairs relevant to a biological network/graph in the order of how disruptive the outcome of deleting the targets of a given drug pair is. So, in short, my applications objective is to kill/disrupt the network, but the optimal goal would be to disrupt nodes related to disease and have minimal impact on the functional part of the cell/network. I've read a few articles about network medicine and found that a diffusion-based method for identifying disease modules seems easier to understand coming from an informatics background. I know how to implement the algorithm for random walkers with reset algorithm (RWR) but unsure how to get/ identify known disease associated genes/nodes of the network and these are used as seeds for the algorithm.
If you guys have some advice on how to do this or even know of a better way to identify disease, functional and topological modules I would appreciate it.
Some details of my project: Application developed to run in Cytoscape, programmed in Java. Using Jung and JGraphT for graph algorithms.
This is for exploring paired drug combinations for treating cancer. I retrieve drug-target relation from a refined list published in an article called: "A comprehensive map of molecular drug targets" from December.
