how to filter genes in significant module?

Question

[modarzi]

Hi,

I used WGCNA for finding modules. I found 9 modules: turquoise(1501), red(173), pink(41), magenta(36), grey(6), green(446), brown(348), blue(382),black(67). Based on Univariate Cox Model just turquoise is significance. The number of 1501 genes in turquoise module is high and I have to filter more for Gene Enrichment Analysis (GEA) and finding significant pathway. So, how can I decrease the number of genes in turquoise module? Generally, is it reasonable that I run a new clustering into turquoise module?

I appreciate if anybody shares his/her comment with me.

[Kevin Blighe]

Why do you need to reduce the number of genes? Did you pre-filter your data prior to performing WGCNA? Can you modify the tree cut height in order to reduce the size of this turquoise module?

Generally, is it reasonable that I run a new clustering into turquoise module?

You can do whatever you please - this is research. Try to think through what it may mean when you do this, though.

[modarzi]

Hi, my origin number of genes in the dataset is 3000 and after constructing, I find 9 modules as I said in the privous comment. The 1501 genes are high in a turquoise module. I would like to achieve a fewer number of genes for finding better results. I use cutreeDynamic() for Module identification based on the below code:

dynamicMods = cutreeDynamic(dendro = geneTree, distM = dissTOM,
                            deepSplit = 2, pamRespectsDendro = FALSE,
                            minClusterSize = 30);

As you see, in that function I have no choice for cutting height and the value of that is calculated automatically.

Now, do you have any more comment for reducing the

[Kevin Blighe]

Try deepSplit = 4