goatpy.batch_correction

Batch correction for MALDI SpatialData objects.

Supports two methods:

• Harmony

  Operates on a PCA embedding. Corrected embeddings are stored in:

  adata.obsm[“X_pca_harmony”]

  The raw intensity matrix (adata.X) is NOT modified.

• ComBat

  Operates directly on adata.X. Corrected intensities overwrite:

  adata.X

This function assumes the data has already been normalised (e.g. TIC or RMS normalisation).

Applying batch correction to raw counts will produce misleading results.

USAGE

# Option 1 — merge inside batch_correction merged = batch_correction(

sdatas=[sdata1, sdata2], batch_names=[“A”, “B”], method=”harmony”,

)

# Option 2 — use an already merged object merged = batch_correction(

pre_merged=merged_sdata, method=”combat”,

)

Functions

batch_correction(→ spatialdata.SpatialData)

Merge SpatialData objects and apply batch correction.

Module Contents

goatpy.batch_correction.batch_correction(sdatas: list[spatialdata.SpatialData] | None = None, batch_names: list[str] | None = None, pre_merged: spatialdata.SpatialData | None = None, method: Literal['harmony', 'combat'] = 'harmony', pcs: int = 30, covariates: list[str] | None = None, table_name: str = 'maldi_adata', batch_col: str = 'batch', feature_join: str = 'inner', random_state: int = 42) → spatialdata.SpatialData

Merge SpatialData objects and apply batch correction.

Parameters:

• sdatas – SpatialData objects to merge.

• batch_names – Batch labels corresponding to sdatas.

• pre_merged – Already merged SpatialData object. If provided, merge_spatialdata is skipped.

• method – “harmony” or “combat”

• pcs – Number of PCs for Harmony.

• covariates – Covariates to preserve during ComBat.

• table_name – AnnData table name.

• batch_col – obs column containing batch labels.

• feature_join – “inner” or “outer”

• random_state – Random seed for Harmony.

Return type:

SpatialData