"Compare to a single cluster" can not work

code:

sc.tl.rank_genes_groups(adata, 'louvain_groups', groups=['13'], reference= '18' )

Error:
ValueError: reference = 18 needs to be one of group_by = [0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23]

I think the problem is the code comfuse str(18) and int(18).

could you solve it? Thanks

[AnatoleKing]

I solved the problem by

adata.obs['louvain_groups'].cat.categories = [
    str(i) for i in adata.obs['louvain_groups'].cat.categories
]

I think you need check the function sc.tl.louvain(adata, n_neighbors=10, resolution=1.3, recompute_graph=True), why it gives int number.

[falexwolf]

Hope that these issues are solved, now…