Hello,
I Recently performed WGCNA analysis of microarray dataset. This dataset consists of two following conditions
1) Untreated 2) treated
WGCNA analysis revealed that there 10 co-expression modules. Now I am thinking to identify the significant module among these 10. I went through the tutorials there are two measures Gene significance and Module significance. But these measures are calculated by relating each module with trait file (like phenotypic features or weight and so on). I also want to do this but I donot have any external information to related to expect two condition information like Untreated and treated. Can any body give suggestion how to determine module significance with out external data?
Thanks a lot,
Sai
•
written
by
Nitin
1 answer
Take the genes in each module and query them using a gene ontology tool such as: DAVID, GOSTAT, REACTOME, ENRICHR or PANTHER. This will give you clues and information about what each module represents biologically, as well as p-values to measure significance.
Significance for what? What is your hypothesis?
Log in to answer this question.
-
WGCNA unpaired trait/expression data
written by nika13 •
-
WGCNA (TRAIT DATA)
written by Rishabh •
-
Grey module trait correlation in WGCNA
written by christofffer •
-
Correlation of modules calculated in separate WGCNA runs
written by rodd
-
WGCNA gene significance
written by Baylie_321 •
-
Problem for getting P-Value in Module-Trait Relationship in WGCNA
written by modarzi
-
WGCNA : Module trait relationships
written by msumaira36 •
-
WGCNA heatmap modules-trait relationship
written by debitboro
-
Significance of GO similarity
written by Nitin
-
Consensus module detection using WGCNA
written by Nitin
Treated and untreated also is a trait. Your traitfile could very simply contain a 1 for treated and a 0 for untreated.
Thanks for reply. I gave a thought about it now you gave me confidence to go ahead in that direction :)
You don't even need a trait file. You could just run a simple correlation between the module eigen-values (moduleEigengenes()) and the 1 for treated and 0 for untreated "phenotype vector" (Just make sure the samples are in the same order).