How do I compare degree and betweenness centrality for different PPI networks?

Question

[gemmalouisebaldock]

Hi everyone,

I am in the midst of writing my project for university and have managed to get to grips with Cytoscape over the last month or two. I have created PPI networks and changed the node size to correlate with the degree and the colour of the node to correlate with the betweenness centrality values. It is my goal to be able to use that data to identify hubs for example.

I am running into a slight issue because I feel that the PPI networks that I am creating are not consistent with each other. To give an example, the BC values for each data set is wildly different and so is the degree. Do I plot the top 10 or 20% nodes with a high BC value, because the BC values for each data set will be different. I am confused as to how to keep things consistent. Any information or advice would be appreciated.

Answer 1

[Istvan Albert]

The betweenness centrality and degree distributions in PPI networks are long tailed. It is normal to observe a high variability (if that is what you mean by "not consistent with each other") hence there won't be a universal cutoff that would have the same interpretation across the networks.

Plotting top 10% is a good rule of thumb, perhaps define hubs as those are in top 10% in both BC and degree.

[gemmalouisebaldock]

Thank you very much Istvan for the reply. Yes, by not consistent I mean that my results are highly variable. I have been able to change the node size and colour to correlate to degree and BC, respectively. In order to show certain nodes fall into the top 10% for both BC and degree, is there a way to denote that in Cytoscape?

[Istvan Albert]

I am not all that knowledgeable of Cytoscapes features and scripting capabilities,

but if it does not support that operation you can export your data, make a new file with additional columns where you label the rows as needed then load that back into Cytoscape.

[gemmalouisebaldock]

Thank you very much, Istvan, I believe I have found a way to rectify this issue.