For cell type annotation in SDAS (e.g., using cell2location), I understand that a single-cell reference .h5ad file is required, containing the raw expression matrix and cell type labels.
I would like to clarify whether this reference is expected to be generated from the same experiment (e.g., from SAW/STOmics outputs such as cellbin.h5ad), or if it must be obtained externally from publicly available scRNA-seq datasets.
Additionally, is it mandatory to provide such a reference for every analysis run, or can a single reference .h5ad file be reused across multiple spatial datasets, as long as it matches the same species and tissue?
Finally, could you confirm whether SDAS supports any form of cell type annotation directly from SAW outputs alone (e.g., using clustering or marker-based approaches), or if all supported methods strictly require an external scRNA-seq reference?
For cell type annotation in SDAS (e.g., using cell2location), I understand that a single-cell reference .h5ad file is required, containing the raw expression matrix and cell type labels.
I would like to clarify whether this reference is expected to be generated from the same experiment (e.g., from SAW/STOmics outputs such as cellbin.h5ad), or if it must be obtained externally from publicly available scRNA-seq datasets.
Additionally, is it mandatory to provide such a reference for every analysis run, or can a single reference .h5ad file be reused across multiple spatial datasets, as long as it matches the same species and tissue?
Finally, could you confirm whether SDAS supports any form of cell type annotation directly from SAW outputs alone (e.g., using clustering or marker-based approaches), or if all supported methods strictly require an external scRNA-seq reference?