How to perform the imputation task on a new dataset without cell types

[lkmdeer]

Hi, I now want to use spatialscope as baseline, but it hardly works on the new dataset.
On the one hand the new dataset doesn't have cell type, I tried to use the clustering results as category labels or the scmap results as labels, neither is good, is it a big influence here?
On the other hand the genes in the dataset only retain the genes shared by ST and scRNA-seq (gene num < 1000), I tried to use cross validation, does the exceptionally low gene number affect the results?

[JiaShun-Xiao]

Hi, accurate cell type labels of scRNA-seq reference are very important for SpatialScope. To have the cell types for scRNA-seq, it is better to integrate scRNA-seq with some other scRNA-seq data with known cell types.
What ST data type are you using? why only <1000 shared genes left?
If the shared genes are very informative, such as marker genes and HVGs, 1000 genes may be sufficient

[lkmdeer]

Hi, accurate cell type labels of scRNA-seq reference are very important for SpatialScope. To have the cell types for scRNA-seq, it is better to integrate scRNA-seq with some other scRNA-seq data with known cell types. What ST data type are you using? why only <1000 shared genes left? If the shared genes are very informative, such as marker genes and HVGs, 1000 genes may be sufficient

Thank you for your response. I preprocess both ST data and scRNA-seq data by retaining only the shared genes. Then, by applying a mask to the ST data and imputing the masked values, I perform cross-validation to impute all genes in the ST data.
I can now run SpatialScope correctly. Thank you very much for your assistance.