atarashansky / SAMap

SAMap: Mapping single-cell RNA sequencing datasets from evolutionarily distant organisms.
MIT License
64 stars 19 forks source link

get_mapping_scores() for cell types within same species non-zero #121

Open dnjst opened 1 year ago

dnjst commented 1 year ago

I was wondering, get_mapping_scores() returns a table of all pairwise scores between cell types of different species, but sets scores within the same species to zero.

This may be a bit misleading, as if I wanted to try to do a downstream interpretation of the scores as a kind of distance metric, certainly, say Hs_Tcell1 and Hs_Tcell2 in the human have some relationship to each other, which may be similar to both of their relations to Mm_Tcell1 and Mm_Tcell2 in the mouse.

Is there any way to calculate mapping scores on the same scale as the cross-species scores for cell types within a data set, or is this simply impossible?

atarashansky commented 1 year ago

The mapping scores are a relative measure of the connectivity between cells across species. It's not really comparable to the connectivity strength of cells within species. As SAMap is a graph-stitching method, it doesn't offer any sort of distance metric.