Hello,
From WGCNA documentation, I understand that a minimum of ~20 samples is required for robust WGCNA. I'm working with a set of 144 samples. These are two replicates for 72 different conditions. I wanted to make sure that using WGCNA in this case would be ok to identify coexpressed modules.
Thanks!
Manoj.
72 different conditions and only two samples (I don't know if biological replicates or technical replicates) for each condition is a low sample size for condition. You could find the co-expressed modules in each condition if you had a bigger sample size ( Ideally ~30) for condition.
If you use WGCNA with all the samples you will identify the co-expressed modules between all the 72 conditions. If you want to find co-expressed genes between classes see this answer of A: WGCNA - Compare module preservation between class, about comparing two different classes, to find out which are the methodologies available.
Use of this site constitutes acceptance of our User Agreement and Privacy Policy.
version 2.3.6
Thanks for your answer Lluís R. Indeed my objective is to find co-expressed modules between all the 72 conditions. I think I am on the right track - using all 144 samples together and then looking at patterns of coexpressed modules. I could see clear banded patterns when I look at modules with r > 0.6 and p < 0.0001.