Can we do a standard pre-clustering step for major cell types?
e.g. using scran::quickCluster() or standard clustering steps from OSCA.
Major cell types could then easily be included as a covariate in the model to identify non-HVG SVGs.
This requires we can easily (without tuning parameters) get a reasonable approximation for the number of "major cell types" in the dataset using the clustering on molecular profiles.
Can we do a standard pre-clustering step for major cell types?
e.g. using
scran::quickCluster()or standard clustering steps from OSCA.Major cell types could then easily be included as a covariate in the model to identify non-HVG SVGs.
This requires we can easily (without tuning parameters) get a reasonable approximation for the number of "major cell types" in the dataset using the clustering on molecular profiles.