Help in understanding condition_key and cell_type_key

[yojetsharma]

From what I understood from the manuscript is that condition_key is something that needs to be regressed out from the analyses. But what if the query and the reference dataset are entirely different?
I have snRNA scanpy object merged from patients+healthy NSCs. So the condition_key in this case is the ‘sample’ and cell_type_key ‘leiden’. But I have a reference from Linnarsson lab (trimester human fetal scRNA seq) which has ‘CellClass’, ‘CellID’, ‘Age’ and ‘Tissue’. So, what should one give as a condition_key here?