emapplot and cnetplot customization (compareCluster)

[plcardozo]

Hi there, I was wondering whether it is possible to customize each arc color for the nodes (or the entire color of the larger nodes, if possible) to a colorblind friendly palette. In addition, I have also had a hard time changing the variable associated with node size from default (number of genes) to adjusted p-value. Finally, the legend related to size is misplaced in my plot. How can I fix all of this, please? Any help would be appreciated.

See an image of what I mean for reference.

[huerqiang]

Thank you for your very good suggestions. centplot() does not currently have parameters to achieve these goals, but we will improve it in future releases.

[plcardozo]

Thank you so much for your quick reply. I understand that the current built-in functions cannot deal with what I asked, but is there a way to modify these parameters either via ggplot, igraph or other packages? Alternatively, is it possible to edit the plot object after generating the plot, the same way I'd do to edit data frames, matrices, etc?

[huerqiang]

It is not easy to modify these parameters or edit the plot object, so we do not provide documentation in this area, and I will update the cnetplot() when I am busy with the work at hand.

[guidohooiveld]

@plcardozo : you indeed came up with some very useful suggestion! Hopefully @huerqiang finds some time to implement these!

FYI (since you asked for hints): please note that you can export a network to Cytoscape for further fine-tuning. See e.g. my post here: #220 (comment)

Note that I realize you generated a network with pie-plots (compareCluster result with multiple groups/clusters), and I wasn't able yet to figure out how to import/visualize such network in Cytoscape... Also a legend will be lacking in Cytoscape.

[plcardozo]

@guidohooiveld thanks for your reply. I also hope @huerqiang can implement these features whenever possible. I think I'll try to import them into Cytoscape as you suggested. If I find a way to do so using the compareCluster results, I'll let you know.

Thanks for your help =D