scvelo.pp.filter_and_normalize n_top_genes

[xzhang08]

Hi @WeilerP, great tool for trajectory inference!!!

I have a question about n_top_genes in scvelo.pp.filter_and_normalize. Is the n_top_genes chosen based on matrix X or unspliced/spliced?

Thanks!!

[WeilerP]

@xzhang08, it's chosen based on adata.X. See here.

[xzhang08]

Thanks so much for your fast reply. I have one more question if you don't mind. I got opposite directions for certain areas on the stream plot using different number of top genes. So the question is which results I should rely on? Is there a way to evaluate the best number of top genes for a particular dataset?