Single-Cell RNA-seq Cell Type Annotation

Routes at a Glance

Choose the route; the tool skills hold the code for each automated route.

Route	Fits when	Needs	In default sandbox	Code
A. Manual markers	Well-characterized tissue, small dataset, or no reference matches tissue and state	`scanpy`	Installed	This skill, Workflow step 4
B. CellTypist transfer	A pre-trained model matches tissue and state	`celltypist`	Installed	`celltypist-cell-annotation`
C. Ensemble transfer	You have labelled reference data	`popv` or `scvi-tools`	Not installed - install first	`popv-cell-annotation`, `scvi-tools-single-cell`
D. Both, compared, step-wise	Labels feed a biological claim; rare populations	A plus B or C	A + B installed	Workflow steps 4 and 7

Spatial data is not covered here: Xenium/MERFISH/CosMx cells go to spatial-transcriptomics-cell-type-annotation, Visium spots to spatial-transcriptomics-deconvolution-cell2location.

Ask the route as D2 in the Pre-flight Interview and say which routes run as-is: route C needs popv or scvi-tools installed first.