hey,
I saw that the example files for both human and mouse contain normalized gene expression. Can I also use MapMyCells with an expression matrix that only contains raw UMI counts?
many thanks
Hello @bendui ,
If you are using the online tool, you actually should use raw UMI counts. The way we wrote the code for the online tool, MapMyCells assumes that it is receiving raw counts and automatically converts them to log2(CPM+1) normalized counts. This ensures that the unlabeled and reference data are normalized in a consistent manner.
If you are running the MapMyCells code locally, you can tell the code that you are submitting raw counts by setting the type_assignment.normalization parameter to 'raw’ (see cells 14 and 15 in sections 4 and 4.1 of this Jupyter notebook).
Let us know if anything remains unclear.