In Scanpy, how to RDS file and merge it with other Scanpy objects.

Question

[Andy]

I have an RDS file that includes several Seurat objects that I want to use. I want to read it into scanpy and merge it with another file. What I did was convert the RDS file to an h5ad file and then read it into scanpy. However, when I merged all my files together, I encountered an error where my merged object had 0 vars. I hope you could offer me some suggestions.

Thanks Andy

Answer 1

[fracarb8]

I usually export everything I need manually

# save metadata table:
[e-mail]$barcode <- rownames([e-mail])
[e-mail]$UMAP_1 <- srtObject@reductions$[e-mail][,1]
[e-mail]$UMAP_2 <- srtObject@reductions$[e-mail][,2]
write.csv([e-mail], file='/data/metadata.csv', quote=F, row.names=F)

# write expression counts matrix
library(Matrix)
counts_matrix <- GetAssayData(srtObject, assay='RNA', slot='counts')
writeMM(counts_matrix, file=paste0("/data/", 'counts.mtx'))

# write dimesnionality reduction matrix
write.csv(srtObject@reductions$[e-mail], file='data/pca.csv', quote=F, row.names=F)

# write gene names
write.table(data.frame('gene'=rownames(counts_matrix)),file='data/gene_names.csv', quote=F,row.names=F,col.names=F)

Once everything is saved, I load it into scanpy

# load sparse matrix :
X = io.mmread("data/counts.mtx")

# create anndata object
adata = anndata.AnnData(X=X.transpose().tocsr() )

# load cell metadata:
cell_meta = pd.read_csv("data/metadata.csv")

# load gene names:
with open("data/gene_names.csv", 'r') as f:
    gene_names = f.read().splitlines()

# set anndata observations and index obs by barcodes, var by gene names
adata.obs = cell_meta
adata.obs.index = adata.obs['barcode']
adata.var.index = gene_names

# load dimensional reduction:
pca = pd.read_csv("data/pca.csv")
pca.index = adata.obs.index

# set pca and umap
adata.obsm['X_pca'] = pca.to_numpy()
adata.obsm['X_umap'] = np.vstack((adata.obs['UMAP_1'].to_numpy(), adata.obs['UMAP_2'].to_numpy())).T

# check the uMAP
sc.pl.umap(adata, color='cellType_final', add_outline=True, legend_loc='on data', legend_fontsize=8, legend_fontoutline=2,frameon=False, title='')

Probably a bit old fashion and not optimal for dealing with multiple files, but it can easily be put into a function.

[Andy]

Thanks! I will try it out!