Changing the p-value threshold should add/remove edges immediately

[maehler]

One problem here is that if edges should be shown again we don't know what order the nodes were selected. The order will affect what orthology edges are shown, and if we just choose to show the orthology edges connected to the current selection of nodes, more edges than expected might be revealed.

[maehler]

This could possibly be solved by only showing orthology edges between pairs of selected nodes. Let's see how difficult this will be to implement.

[dilema]

Added this feature during refactoring. I used the suggested ideas : check edges between pairs of selected genes. Commit : d34a462f9506a8c2b4ee8e6f35c5102ac41c4d36

[dilema]

Since I misunderstood the problem with co-expression edges threshold (I thought some nodes would appear and disappear), we could implement the same thing for co-expression edges.

Possible issues:

• to many data (memory problem)
• to many edges (slow cytoscape rendering)

Test: I took the three leaf networks and the Arabidopsis example gene list as reference with the minimum threshold for co-expression (0.95) to see if there is really a lot of co-expression edges compared to orthology edges. Total nodes = 193 Total orthology edges = 161 Total co-expression edges = 956 Yeah so almost 6 times more co-expression edges than orthology edges...