How SAMap deal with species-specific cell types/clusters in integration?

[DiracZhu1998]

Dear Dr. Atarashansky,

I tried your wonderful tool with a dataset containing species-specific cell types, but the species-specific cell type can still map to another species' cell type with a mapping score above 0.5. Although the gene pairs supporting this only have around 100 gene pairs. If I understand your paper correctly, your method will force the species-specific cell types to map to a similar (but not homologous) cell type with much more weight on the small number of genes in the gene homology graph.

Could you confirm that?

Best wishes,
YZ