Working with Visium data deconvolved via cell2location. Need to calculate gene signature scores but cell type layers have different scales (Ductal: 0-37 vs Malignant: 0-228). Best approach? 1) Global score 2) Cell-type-specific 3) Weighted by abundance. Leaning toward option 3 but concerned about comparability across spots. Scale differences make normalization tricky. Anyone implemented signature scoring on cell2location results? #spatial_omics