WGCNA: many over and under-expressed features in modules of a signed network
1
0
Entering edit mode
3 months ago

In a WGCNA analysis of transcriptome and proteome of a white blood cell in development (in 6 stages), I find in most modules (especially the large ones) over, as well as underexpressed features, but I am using a signed network type.

I attach some of the graphs and results below. This is my first time running a WGCNA analysis and it's hard for me to deduct meaning from some of the graphical outputs. Would it be great if you could give me tips to better understand the outputs and explain why I get a mix of over and underexpressed features? A final question would be if it matters if some of the groups are highly similar. (eg: the last 3 stages are highly similar on proteome level). Does this throw off the analysis? Should I unite them into one group? Would it matter if that combined group is larger than the other groups? (4-5 samples per group, new group would be 14 samples).

Data: transcriptome as tpm with vsn normalisation proteome as intensity with vsn normalisation. Both data batch effect free.

Thank you very much!

Sebastian

(ps: his is a re-post from a question I posted at the bioconductor forum, where it unfortunatly did not get an answer: https://support.bioconductor.org/p/9152480/

enter image description here enter image description here

WGCNA • 318 views
ADD COMMENT
2
Entering edit mode
3 months ago

In blockwiseModules set networkType = "signed"

ADD COMMENT
0
Entering edit mode

Sorry, I forgot to add that code. I used the code below:

bwnet_mrn <- blockwiseModules(exp_4wgcna_mrn,
                              maxBlockSize = ncol(exp_4wgcna_mrn),
                              TOMType = 'signed',
                              power = soft_power_mrn,
                              mergeCutHeight = 0.25,
                              numericLabels = T,
                              randomSeed = 3234,
                              verbose = 3)

Ah, I just realized that I set the TOMType to signed. Is that something different than the network type?

ADD REPLY
1
Entering edit mode

Ah, I just realized that I set the TOMType to signed. Is that something different than the network type?

It is a setting used for the construnction of the adjacency matrix and the default value is unsigned. See the usage of the blockwiseModules() function: link

ADD REPLY

Login before adding your answer.

Traffic: 1141 users visited in the last hour
Help About
FAQ
Access RSS
API
Stats

Use of this site constitutes acceptance of our User Agreement and Privacy Policy.

Powered by the version 2.3.6