When I analyse my sensitive yeast deletion strains, I get a long list of different drugs/chemicals, with different scores and p value. I am unable to figure out the most similar drug/chemical by looking at that data.
That was the case. In the original dataset, the p value is given by comparing the condition/control and calculating the log ratio But upon analysis by NetwoRx, one-sided P-value was computed corresponding to the Z-score defined by the Gene Set Score. For me, its still difficult to explain.
Sorry, I am not a networks expert, but I sent your question to someone who does know that topic. However, I think she is very busy and may not have a response for you. We will see.
Best wishes - Jim
PS - Perhaps you might obtain more responses by posting under social networks and/or neural networks, as I don't see simply "networks" as a ResearchGate topic. Anyway, you might find a topic category for this or a similar question that could reach more people working in this area.