omicverse.single.batch_correction

omicverse.single.batch_correction#

omicverse.single.batch_correction(adata, batch_key, use_rep='scaled|original|X_pca', methods='harmony', n_pcs=50, **kwargs)[源代码]#

Run batch-effect correction for single-cell data integration.

参数:

  • adata (anndata.AnnData) – Input data matrix. Different methods require different preprocessing: for example, scVI typically needs raw counts in adata.layers['counts'].

  • batch_key (str) – Column in adata.obs defining batch/domain labels.

  • use_rep (str, default='scaled|original|X_pca') – Embedding key used by embedding-based methods (for example Harmony).

  • methods (str, default='harmony') – Integration backend. Supported values include 'harmony', 'combat', 'scanorama', 'scVI', 'scANVI' (semi-supervised VAE; requires labels_key=), 'totalVI' (joint RNA+protein; requires protein_expression_obsm_key=), 'scPoli' (scArches conditional VAE; supports cell_type_keys= for prototype learning), 'CellANOVA', 'Concord', 'cca' / 'seurat_cca' — the pure-Python port of Seurat::RunCCA (via the pyccasc package, no R / rpy2 required) — and 'rpca' / 'seurat_rpca' — a faithful Python port of Seurat’s IntegrateLayers(method = RPCAIntegration) (reciprocal-PCA anchors, no R required; writes adata.obsm['X_rpca']). RPCA accepts features=, reference=, k_anchor=, k_weight=, sd_weight= and orig_rep= (the joint reduction to correct).

  • n_pcs (int, default=50) – Number of principal components / canonical components used by the selected backend. For 'cca' this is the number of canonical components (num_cc in pyccasc terms).

  • **kwargs – Additional method-specific keyword arguments passed to the selected backend implementation.

返回:

Returns adata for most methods, scVI model object for 'scVI', and Concord object for 'Concord'. The integrated embeddings are written to adata.obsm.

返回类型:

anndata.AnnData or object

目录