WGCNA unpaired trait/expression data

Question

[nika13]

Hi all,

I am new to WGCNA and am learning how to relate module eigengenes to phenotypic traits.

I was able to perform correlation analysis between categorical traits and module eigengene values for RNAseq data collected from animal samples across four populations (3 replicates per population).

However, I also have an unpaired quantitative trait (offspring number) measured from different individuals within the same four populations (10 replicates per population). To include this trait in the analysis, should I use the average offspring number across the 10 replicates for each population in the correlation analysis?

Would be grateful for any tips!

Answer 1

[LChart]

Because you cannot associate offspring number with any sample in your expression data, the analysis you envision is impossible; and you cannot perform any statistics on an individual

The best you can do is to average the eigengenes within population, and average the offspring number within population, talk about the relationship between averages. However saying "the red module average eigengene trends with offspring number" will necessarily be unsatisfying. You could take a look at some of the comparative WGCNA studies (say human vs non-human primate) which describe changes in co-expression architecture. N=3 is very under-powered however.