Imputing genes

[thomvreven]

Great to see the Crescendo code released! We have used it to model gene expression counts for Harmonized dissociated/spatial single cell data. What we are particularly interested in is imputing genes that are present in the dissociated dataset, but are not part of the spatial gene panel. Crescendo did not do that out of the box, but when we extended the spatial count matrix to include those 'missing' genes (set to zero counts), the program did return gene expression counts for the spatial cells. Great to have these genes imputed, but since Crescendo did not do it out of the box, we are wondering whether the results actually make sense. Any thoughts?