Nan values after running sc.tl.filter_rank_genes_groups and plotting

[shreyii]

Hi Team,

I'm analyzing scrna seq data from different mice tissues. I'm following the standard pipeline - preprocessing, normalizing, feature selection, clustering - currently annotating cell types(manual + reference). I've observed that when I want to study the markers cluster wise based on logfcscore, after running sc.tl.filter_rank_genes_groups and plotting - I get 'nan' values in the clusters at certain positions. Kindly assist, thank you.

[ilan-gold]

Can you add a reproducer for this behavior? Are you sure your genes do not have 0 value?