symphonypy.tl.transfer_labels_kNN

symphonypy.tl.transfer_labels_kNN(adata_query: AnnData, adata_ref: AnnData, ref_labels: list[str] | str, *kNN_args, query_labels: list[str] | str | None = None, ref_basis: str = 'X_pca_harmony', query_basis: str = 'X_pca_harmony', **kNN_kwargs) → None

Run sklearn kNN classifier for label transferring.

Parameters:

• adata_query (AnnData) – AnnData object to use for train

• adata_ref (AnnData) – AnnData object to use for prediction

• ref_labels (list[str] | str) – either a list of column names or a str of one column name from adata_ref.obs to use as labels for model training

• query_labels (list[str] | str | None, optional) – keys in adata_query.obs where to save transferred ref_labels (in corresponding to ref_labels order). If not provided, ref_labels will be used

• ref_basis (str, optional) – adata_ref.obsm[ref_basis] will be used as features for kNN training, defaults to “X_pca_harmony”

• query_basis (str, optional) – adata_query.obsm[query_basis] will be used as features for prediction, defaults to “X_pca_harmony”