Hi everyone,

I’m using the modulePreservation() function from the WGCNA package to assess module preservation across datasets. My question concerns how adjacency is calculated when I use expression data directly (i.e., dataIsExpr = TRUE).

From what I understand, modulePreservation() computes the adjacency matrices internally using a correlation function (like cor). However, I couldn’t find a way to specify the soft-thresholding power — the power parameter isn’t exposed in the function’s arguments.

In my case, I determined the power for the reference network using scale-free topology fit (SFT), so it’s not the default (e.g., 6). I want to make sure this is correctly used in the preservation step.

Is there any way to explicitly set the power used inside modulePreservation() when dataIsExpr = TRUE?

Or should I precompute the adjacency matrices using my chosen power and set dataIsExpr = FALSE?

Alternatively, is this a non-issue and WGCNA somehow reuses the power from the reference network construction?

Any guidance would be greatly appreciated.