dummy 'kid' variable for whole dataset DE

HelenaLC / muscat

Multi-sample multi-group scRNA-seq analysis tools

166 stars 33 forks source link

dummy 'kid' variable for whole dataset DE #80

Closed lindsaynhayes closed 2 years ago

lindsaynhayes commented 3 years ago

Hello muscat masters.

I would like to contrast the Cell Level mixed Analysis on all the data not on a per cluster basis with the cluster based analysis. I tried the prepSCE() without the kid term, but it didn't like that. So I added a dummy variable for kid to be all the same value for all the cells, then I ran mmDS. Does this seem like a reasonable approach or will I induce an illegal action or assumption in the mmDS function?

HelenaLC commented 2 years ago

Apologies for the very late and probably no-longer useful reply... I think this is fine to do! We have done exactly this here in order to test for global state-changes, i.e., overall differences across groups when being blinded to subpopulations.