Dear Suoqin Jin,

When using aggregateNet() followed by netVisual_diffInteraction() for visualizing changes in cell-cell communication between two groups, it is possible only for pathways detected as overrepresented (with significant interaction) in both groups. Unfortunately, this is not always the case. In our research projects, we regularly find overrepresented pathways in only one of the disease conditions. In such cases, we still want to visualize and show the dimension of such differences.

I have introduced some slight modifications in three of your functions (aggregateNet, subsetCommunication, and subsetCommunication_internal) to calculate the aggregated network of nonsignificant pathways so they can be compared across groups. I created an example using your nonlesional (NL) and lesional (LS) data set for the TNF pathway, only detected as significant in the LS group (see zipped jupyter notebook in attachment, those lines changed are marked with #ADDED...).

It would be great if you could provide me feedback on such an approach, and if you are convinced, merge the changes included in the pull requests, so the community could also explore such functionality.

Best regards,

Alvaro
PS: I also added dplyr::summarize() in aggregateNet(). This solves an error when plyr is loaded after dplyr.

TEST_new_aggregateNet_netVisual_diffInteraction.zip