scHopfield.tools.decompose_bias_energy

scHopfield.tools.decompose_bias_energy(adata: AnnData, cluster: str, spliced_key: str = 'Ms', cluster_key: str = 'cell_type', x: ndarray | None = None) → ndarray[source]

Calculate gene-wise bias energy decomposition.

Computes the bias energy contribution for each gene separately, allowing analysis of which genes contribute most to the total bias energy.

Parameters:

  • adata (AnnData) – Annotated data object with fitted parameters

  • cluster (str) – Cluster name to use for bias vector

  • spliced_key (str, optional (default: 'Ms')) – Key in adata.layers for spliced counts

  • cluster_key (str, optional (default: 'cell_type')) – Key in adata.obs for cluster labels

  • x (np.ndarray, optional) – Optional expression data. If None, uses stored sigmoid values

Returns:

Array of shape (n_cells, n_genes) with bias energy per gene

Return type:

np.ndarray