Hello,
I am currently using the hdWGCNA package on single cell data. So I get modules on a scRNAseq of cancer cells that I project on healthy cells in order to see what are the modules that are "absent" from the healthy cells and so "exclusive" to cancer cells.
To affirm this, I proceeded to the Z-statistics analysis which gave me two types of values which are the Z-summary.qual and Z.summary.pres. I thought I understood that, unlike Zsummary.pres which measures the preservation of the modules in the other dataset, Zsummary.qual is the quality of the individual modules themselves without taking into account the projection of the modules from the cancer cells onto the healthy cell dataset.
In a paper I read (https://onlinelibrary.wiley.com/doi/epdf/10.1002/jcb.28377), he describes on page 4 a segmentation of their modules according to different thresholds:
- Z-summary > 10 : Highly preserved modules
- 5 > Z-summary > 2 :
- Moderately preserved modules 2 > Z-summary : Very variable modules
In my case, where I want to see which modules are "exclusive" to cancer cells, I should prefer the modules being the most variable in the projection on normal cells and thus keep only the modules having a Z-summary .pres < 2 and exclude those having a Z-summary .pres > 10 since these last ones would be just as present in normal cells as in cancer cells, right ?
In addition to that, concerning the Z-summary .qual which is a quality criterion of the modules, do we agree that in all cases I should keep exclusively the modules with a Z-summary .qual > 10 and possibly those with a 5 > Z-summary .qual > 2?
To summarize, the modules being "exclusive" to cancer cells would be those with a Z-summary.qual higher than 2 and a Z-summary.pres lower than 2, right?