scanvi_predict

besca.tl.auto_annot.scanvi_predict(adata_trains, adata_pred, celltype)

merges all datasets and predicts on testing set with scANVI. Note that unlike auto_annot predict, merging and fitting is all combined in one function.

Parameters:

• adata_trains (list) – List of training datasets

• adata_pred (AnnData) –

  AnnData object containing testing dataset celltype: string

  Column to be used as training label

• adata_train (AnnData) – Merged training datasets with new representation stored in obsm

• adata_pred – Test dataset with predictions in obs and new representation in obsm

• adata_concat (AnnData) – Both adata_train and adata_pred concatenated to one dataset