Total Variational Inference

This page is under construction. For now, please learn more about totalVI in our user guide.

# totalVI is for end-to-end analysis of CITE-seq dataimport anndataimport scvi
# read data and prepare for scvi-toolsadata = anndata.read("my_data.h5ad")scvi.data.setup_anndata(adata, protein_expression_obsm_key="proteins")model = scvi.model.TOTALVI(adata)model.train()
# get integrated low-dimensional representationadata.obsm["X_totalvi"] = model.get_latent_representation()
# normalized expressionadata.layers["totalvi_norm"], adata.obsm["totalvi_pro_norm"] = model.get_normalized_expression()
# differential expressionresults = model.differential_expression(    groupby="cell types",    group1="CD4",    group2="CD8")