search

thpar / diffany

Diffany is an open-source toolbox for calculating and visualizing differential networks.
7 stars 2 forks source link

[CLOSED] Visualisation for virtual nodes / unspecific edges #132

Closed diffany-admin closed 9 years ago

diffany-admin commented 9 years ago

Issue by svlandeg Thu Apr 10 13:17:22 2014

Provide Cytoscape visualisation of the data modeled in #131

diffany-admin commented 9 years ago

Comment by svlandeg Tue May 20 13:40:22 2014

Example from part of the 1.5h osmotic data, with node fill color and node label color either purple or white:

1 5h_2ppi

diffany-admin commented 9 years ago

Comment by svlandeg Tue May 20 13:52:58 2014

PS: I really like not having lines around the nodes. Also helps with these long node names when a shorter synonym is not available...

diffany-admin commented 9 years ago

Comment by svlandeg Thu Jun 12 15:22:01 2014

Example of alternative visualisation, with #145. Now the up/down regulation is depicted as node attributes rather than edges with hidden nodes.

alternative

diffany-admin commented 9 years ago

Comment by svlandeg Wed Jun 18 20:45:31 2014

Another example, with positive (green edge) and negative (red edge) regulations & PPIs (purple), and non-DE genes (grey) as well as overexpressed (green node) as repressed (red node) genes: 1 5_pil5_neigh1

diffany-admin commented 9 years ago

Comment by svlandeg Thu Jun 19 13:31:27 2014

This might give some additional trouble in Cytoscape, as node attributes are defined per network collection. So, Diffany would need to be able to calculate differential networks across different collections...

diffany-admin commented 9 years ago

Comment by svlandeg Thu Jun 19 15:19:24 2014

A final option is to have a pre-processing step which translates the DE genes and their FC to a difference in weight of the adjacent edges... will discuss these options with Marieke and see whether she has an opinion...

diffany-admin commented 9 years ago

Comment by svlandeg Thu Aug 28 15:13:00 2014

They think the translation of this node-specific data to the edge weights might make for more intuitive, condensed networks. This will avoid the collection-specific node attribute problem in cytoscape, though at the same time we will lose some information. But I guess the user will then have to consult the raw data to fetch the FC values etc.

diffany-admin commented 9 years ago

Comment by svlandeg Fri Sep 5 08:41:32 2014

This will be further implemented in #184