wgcna modules detection and MEs

Question

[Omics data mining]

Dear all,

Recently I have gone through tutorial of WGCNA and performed Module identification using dynamic tree cut and saved output in variable dynamicMods. Then, I proceeded for Merging of modules whose expression profiles are very similar. As a result found Eigengenes of the new merged modules into mergedColors variable (following tutorial).

mergedMEs = merge$newMEs; MEs = mergedMEs;

Initially, There were 89 modules in total but number of modules is 48 in MEs.

In downstream analysis, should I proceed with modules colors shown only in MEs but number of modules is 48 ??? But during the gene count, I found it cover just 50 % of my input gene list.

All suggestions will be appreciated.

Thank you Archana

[zizigolu]

I think the genes whose expression is so that could not be grouped within a module (containing genes whose expression levels is somehow similar) have been ignored. We can define the minimum number of genes assigned to each module that the default is 30 genes.

[Omics data mining]

Dear Fereshteh

I checked moduleColors table count (also named as mergedColors) which consist of total of all genes into multiple modules where gene count matches to my input list. I missed this point earlier .

Thank you for your suggestion :)

[Kevin Blighe]

That is a very large number of modules (?). The grey module is usually the 'waste' module, I believe, i.e., the module to which unassigned genes are [ironically] assigned

[1769mkc]

@Kevin for WGCNA I m doing as such so i have 5 different lineages i have taken the feature count from all the 5 different lineages with their biological replicates , then i m running deseq2 , so after that i'm using VST data for WGCNA ,so is it the right way for WGCNA input or am I doing something wrong ?