Hallo,

I’m using visANT for the first time, after performing a WGCNA. I managed to obtain modules that seem to be related to our disease of interest, and I created files to export to visANT. I have a beginner question on something I’m doing in visANT: I would like to hear if what I’m doing is acceptable to get trustable hub genes from my modules. I will explain what I do: after importing the module and having the node sizes adjusted and the layout set to elegant relaxed, I get a very dense net of edges. So I use the edge weight cut off function and increase the lowest value from 0 to one that gets me a -by eye- reasonable number of hub genes. When going to Topology>>global statistics>>degree distribution I get then a graph with red dots and a correlation value, from which I select the ones with more connections, on the right part of the graph. Does all this make sense? Or should I use a more strict approach? Thank you in advance for any comment,

Angela

I like to filter the visAnt module by edge weight in a similar way you do by simply increasing the stringency until all edges disappear.  Once they all disappear, I then step the filter down one step and then create a node from the sub-network.  I then change the layout to the spoke layout so that the hub gene is in the center.  I then incrementally step down the filter stringency and generate more nodes connected to the previous nodes to create a edge weight based interaction map.  I then GO annotate the nodes to find putative functions.  You can then add fold-change data to the genes and overlay KEGG graphs to try and make informed biological interpretations on the data.  I don't know if anyone else does it this way, but this is what I have been doing.  I'm open to other ways of going about using visANT and I'm glad you brought this up.