bento.tl.flux

bento.tl.flux(sdata, points_key='transcripts', instance_key='cell_boundaries', feature_key='feature_name', method='radius', n_neighbors=None, radius=None, res=1, train_size=1, random_state=11, recompute=False, num_workers=1) → SpatialData

Compute RNAflux embeddings of each pixel as local composition normalized by cell composition.

Parameters:

  • sdata (SpatialData) – SpatialData object.

  • points_key (str, default “transcripts”) – Key for points element that holds transcript coordinates.

  • instance_key (str, default “cell_boundaries”) – Key for cell_boundaries instances.

  • feature_key (str, default “feature_name”) – Key for gene instances.

  • method (Literal[“knn”, “radius”], default “radius”) – Method to use for local neighborhood.

  • n_neighbors (Optional[int], default None) – Number of neighbors to use for local neighborhood if method is “knn”.

  • radius (Optional[float], default None) – Fraction of mean cell radius to use for local neighborhood if method is “radius”. If None, defaults to 1/3 of average cell radius.

  • res (Optional[float], default 1) – Resolution to use for rendering embedding.

  • train_size (Optional[float], default 1) – Fraction of data to use for training.

  • random_state (int, default 11) – Random state for reproducibility.

  • recompute (bool, default False) – If True, recompute flux even if it already exists.

  • num_workers (int, default 1) – Number of workers to use for parallel processing.

Returns:

Updated SpatialData object with: - .points[“{instance_key}_raster”]: pd.DataFrame containing flux values, embeddings, and colors. - .tables[“table”].uns[“flux_genes”]: List of genes used for embedding. - .tables[“table”].uns[“flux_variance_ratio”]: Array of explained variance ratio for each component.

Return type:

SpatialData

Notes

RNAflux requires a minimum of 4 genes per cell to compute all embeddings properly.