Question about module_trait_cor

[PFRoux]

Dear BioNERO developers,

Thanks a lot for this nice implementation of WGCNA.

I wanted to know how module_trait_cor is working when the variable of interest are categorical.
In other words, how to correlate module eigengenes to categorical traits ? What is the meaning of the correlation values ?
Wouldn't it be more meaningful to run an ANOVA kind of test to compare expression of eigengenes across categories for those categorical traits ?

Thanks a lot and have a great day.

Pef

[almeidasilvaf]

Hi, @PFRoux

For categorical variables, BioNERO calculates point-biserial correlation coefficients. The design matrix in this case includes so-called 'dummy variables' (without an intercept), as you'd do for levels of a factor (see source code here).

You could also run an ANOVA to compare eigengenes across levels, but you'd have to check if assumptions are met first.

All the best,
Fabricio