Closed sid5427 closed 1 year ago
resolved myself -
for adding UMAP embeddings.... umap coordinates go into adata.obsm['X_umap']
sample code -
##read in UMAP coordinates
pd_cell_umap_coord = pd.read_csv('umap_coordinates.txt', sep='\t')
#where you have 3 columns - cell barcode, umap X coordinate, umap y coordinate
#numpy stack
umap_array=np.column_stack((pd_cell_umap_coord.UMAP_1,pd_cell_umap_coord.UMAP_2))
##optional change datatype to float32 for entire array
umap_array = umap_array.astype('float32')
#assign umap coordinates
adata.obsm['X_umap']=umap_array
Hi Seppe,
Is there a direct way to use our own filtered scRNAseq data, in the form of a filtered gene matrix along with a csv or similar file containing cellbarcode to cell type mapping, along with two columns for UMAP1 and UMAP2 coordinates?
I created my own adata object and loaded the filtered gene matrix using sc.read_10x_h5. However, can you suggest a way to annotate the cells using my mapping table with umap coordinates?
-- Sid