Module merging WGCNA

Question: Module merging WGCNA

18 months ago by

Genomic Island

Bioinformatics_NewComer • 320 wrote:

Hello all,

I'm interested to perform gene co-expression using WGCNA R module. I clean data for samples and remove genes that don't form any cluster. Post cleaning I perform clustering to identify different modules (40 excluding grey). Each module consist different number of genes, I generate list of genes in a output text file.
Next, I merge module that are highly similar and left with 20 modules. Further, I perform regression to identify module significant to the trait. I see the P-value and module name, which is OK.

I'm unable to move forward here with two things:

1- I don't know which modules were merged with what. Initial module clustering and gene list isn't helpful after module merging.
2. I'd like to have gene names binned in newly merged modules.

How do I make my way with these two points?

Thanks.

microarray expression wgcna • 1.6k views

ADD COMMENT • link •

written 18 months ago by Bioinformatics_NewComer • 320

18 months ago by

Kevin Blighe ♦ 63k

Kevin Blighe ♦ 63k wrote:

Hey, you should check the output of each and every command that you have used so that you can understand how the program's functions behave. Then, you may find the answer to your question(s).

ADD COMMENT • link modified 11 months ago • written 18 months ago by Kevin Blighe ♦ 63k

I'm working step by step, looking into function parameters and reading explanations provided on the tutorial. After rummaging and lost I've posted my query. I can share code and comments if needed to show my inputs.

https://horvath.genetics.ucla.edu/html/CoexpressionNetwork/Rpackages/WGCNA/Tutorials/

http://systemsbio.ucsd.edu/WGCNAdemo/

http://pklab.med.harvard.edu/scw2014/WGCNA.html

The first URL has work flow; I've gone through different PDFs individually (not blindly/naively).

Thank you.

ADD REPLY • link modified 18 months ago • written 18 months ago by Bioinformatics_NewComer • 320

Thank you. Relating to your 2 questions, can you show the lines of code from the tutorials to which your questions relate? It can help to use the str() function on objects so that you can see all information that is held within them.

ADD REPLY • link written 18 months ago by Kevin Blighe ♦ 63k

URL: https://horvath.genetics.ucla.edu/html/CoexpressionNetwork/Rpackages/WGCNA/Tutorials/Consensus-NetworkConstruction-man.pdf

Section 2.a.7, Page 9,

merge = mergeCloseModules(multiExpr, unmergedLabels, cutHeight = 0.25, verbose = 3)

Here, code merges modules that have close/similar expression values at a height of 0.25 (in dendrogram, Fig. 4) are merged.

ADD REPLY • link written 18 months ago by Bioinformatics_NewComer • 320

After you run these:

# Numeric module labels
moduleLabels = merge$colors

# Convert labels to colors
moduleColors = labels2colors(moduleLabels)

,,,the colour vector, moduleColors, has the same order as your genes (or samples, if you are doing WGCNA sample-wise).

To see which modules have been merged, run:

cbind(unmergedColors, moduleColors)

ADD REPLY • link modified 18 months ago • written 18 months ago by Kevin Blighe ♦ 63k

Similar posts • Search »