Improve compare gene modules function.

[mutwil]

Did you not have some prototype running already, and if yes, when will you make this available for testing?

[proost]

Similar clusters can be detected and graphs plotted side by side. Linking homologs with edges.
The next step would be to include phylostrata information to edges connecting homologs.

Furthermore Expression Context Conservation can be computed between homologs, however this performed quite poorly with the hrr <30 networks. I'll have to check between two hrr <100 networks to see if this becomes usable at this network density (in the original publication pcc > 0.4 was used). This would connect neighborhoods

We need more people testing to see how much priority we give these advanced features. So far the most feedback is on the heatmap, profiles, expression specificity not on anything directly related with the co-expression.

[mutwil]

Yeah, lets go with hrr<100, first neighbours for all networks. To minimize confusion fir the users, lets remove the hrr<30 instances.

[mutwil]

So the modules could be shown like this:

Where nodes represent gene families or label co-occurrences (the user can select which).

Phylogenetic info between the genes could be shown with the colored edges between modules:

[proost]

the second I have (with the exception of some tweaks), but there is no UI to get to a meaningful comparison yet.

[proost]

First step. The interface to get to this point is however more difficult !