Spatial Transcriptomics
Assessing segmentation quality of marker cells
Pipeline via Molecular Cartography ImageJ2 plugin
Input Image and gene list from Resolve
Transferring to ImageJ2 to use the Molecular Cartography plugin
Move data to QuPath for segmentation.
Returning the segmented data to ImageJ to see cells filled with expressed genes in it
Obtaining the countmatrix from the plugin and further clustering analysis via Seurat
I produced correlation plots to see good or mixed clusters to see whether marker genes intersect in one cell.
I can say, watershed segmentation from QuPath is not good enough
I would use now for the Resolve data for another segmentation algorithm, designed as an ImageJ plugin (CellPose, Stardict, etc)