Cytoscape Wiki|
Size: 1833
Comment:
|
← Revision 6 as of 2009-02-12 01:03:48
Size: 1833
Comment: converted to 1.6 markup
|
| Deletions are marked like this. | Additions are marked like this. |
| Line 8: | Line 8: |
| * KeiichiroOno - This question is about '''''interaction''''' attribute for edges. When a metanode is created, interaction type of all edges connected to the node is set to '''''metaEdge'''''. Although it's a metaEdge, at the same time, it still has original interaction type (like, protein-protein interaction). So how about create a new edge attribute called '''''edgeType''''' instead of overwriting '''''interaction''''' edge attribute? '''''EdgeType''''' attribute will have two types of data: ''normal'' or ''metaEdge''. | * KeiichiroOno - This question is about '''''interaction''''' attribute for edges. When a metanode is created, interaction type of all edges connected to the node is set to '''''metaEdge'''''. Although it's a metaEdge, at the same time, it is still original interaction type (like, protein-protein interaction). So, how about create a new edge attribute called '''''edgeType''''' instead of overwriting '''''interaction''''' edge attribute? '''''EdgeType''''' attribute will have two types of data: ''normal'' or ''metaEdge''. |
Comments on Metanode Concept Plugin
KeiichiroOno - Currently, metanode plugin uses commonname as the storage for member node names. Should we create a new attribute membernode or children to store member node names? Or just continue to use commonname ?
IlianaAvila - The best way of getting the names of the children nodes of a metanode is by using the GINY method to get the children Nodes of a Node, and then iterating through the children to get their names using Cytoscape methods. The commonname attribute is not meant to be storage for the children nodes' names, it is just a possible way of labeling a metanode, which can change. I don't like the idea of having an attribute to store the names of children nodes because we would need to keep this attribute synchronized to the actual metanode membership. So if a node is no longer a child of a metanode, or a new node is added as a child, we would need to reflect this in the attribute.
KeiichiroOno - OK, let's go with the idea without new attribute (children/membernodes). I'll modify the attribute browser to distinguish metanodes and normal nodes (maybe highlight rows with metanode). Also, the new browser will show the children of metanodes somehow.
KeiichiroOno - This question is about interaction attribute for edges. When a metanode is created, interaction type of all edges connected to the node is set to metaEdge. Although it's a metaEdge, at the same time, it is still original interaction type (like, protein-protein interaction). So, how about create a new edge attribute called edgeType instead of overwriting interaction edge attribute? EdgeType attribute will have two types of data: normal or metaEdge.