The cell projection from single cell to spatial transcriptomics is very different from Seurat

[smallfishcui]

Dear developer,

Thank you very much for developing this interesting tool that could integrate the RNA velocity analysis of single cell and spatial transcriptomics. I run this analysis to project the single cell clusters on the the spatial, but the projections are very different than the projections in Seurat (I used prediciont.assay scoring), so I am wondering is this expected? How could I improve the accurary on the cell projection?

thanks,
Cui