Hello Scanpy Team,

I am trying to understand if the matrix generated by Scanpy after scaling (sparse matrix) is similar to matrix returned by Seurat’s ScaleData (this function returns zscores residuals which means I have negative values as well). Does Scanpy do the same? I converted the sparse matrix to a dense and written it to a csv file. I don’t see any negative values here, the lowest value I see is 0.

I need to do some overlap analysis with regards to markers calculated and certain gene lists of interest. In order to do this, I am wondering should I be plotting scaled data of overlapped genes or raw data? I have expression data from 3 different papers (some using Seurat, others Scanpy and others a combination). Hence would be great to know if Scanpy approaches scaling data differently from Seurat or if they are comparable?

Thank you

Asma