Master Maps - an implementation by Oliver
Example:
load the attached session file
- ignore the Networks EM1 - EM3 and directly go to the "Master Map"
- switch the visual Style between EM1, EM2, EM3 and Master
How to manually create a Master Map:
- create a couple of Enrichment Maps in one session
- Merge Networks to a Master Map (Merge Networks Plugin - Union)
- for each Enrichment Map:
- create Node-Attribute (String) "is_significant_in_EM#" and default it to "false" (as a flag to see in which analysis the gene set came up)
- set "is_significant_in_EM#" to "true" for all gene-sets in Enrichment Map EM# (that have non-empty attributes EM#_...)
- for each Visual Style "EM#":
- create discrete Mapping of "is_significant_in_EM#" to Node Opacity: true=255.0; false=100
- create discrete Mapping of "is_significant_in_EM#" to Node Border Opacity: true=255.0; false=100
- Master Map Style:
- create Node Attribute "Master_formatted_Name" and copy each "EM#_formatted_Name" to this Attribute
- create Node Attribute "Master_Colouring_dataset1" and copy each "EM#_Colouring_dataset1" to this Attribute
- create Node Attribute "Master_gs_size_dataset1" and copy each "EM#_gs_size_dataset1" to this Attribute
- create Node Attribute "Master_Genes" and copy each "EM#_Genes" to this Attribute
- create Edge Attribute "Master_similarity_coeffecient" and copy each "EM#_similarity_coeffecient" to this Attribute
- create Visual Style "Master Map" based on "EM#"
- map Edge Line Width with "Master_similarity_coeffecient"
- map Node Color with "Master_Colouring_dataset1"
- map Node Label with "Master_formatted_Name"
- map Node size with "Master_gs_size_dataset1"
How to do it programmatically
- same as above plus:
- re-calculate overlap to capture all gene-set combinations
- calculate an overall colouring value (e.g. most extreme, average)
associate new instance of EnrichmentMapParameters with the map, that we can use the slider-bars, too
(and maybe also run the Signature Post-Analysis)- join expression data from EM# Maps and attach it to the Mastermap (we can eliminate identical columns to reduce the size)