ocelli.tl.imputation

Contents

ocelli.tl.imputation#

ocelli.tl.imputation(adata: AnnData, t: int = 5, kappa: float = 1.0, features: list | None = None, eigvals: str = 'eigenvalues', eigvecs: str = 'eigenvectors', scale: float = 1.0, copy: bool = False)#

Diffusion-based multimodal imputation.

This function performs iterative imputation of a count matrix (adata.X) using multimodal eigenvectors and eigenvalues. The imputation amplifies signals while preserving or enhancing the structure in the data.

Parameters:

  • adata (anndata.AnnData) – The annotated data matrix.

  • t (int) – Number of imputation iterations. Higher values increase imputation smoothing. (default: 5)

  • kappa (float) – Scaling factor for signal amplification while preserving maximum expression values. (default: 1.0)

  • features (list or None) – List of feature names (adata.var.index) to be imputed. If None, all features are imputed. (default: None)

  • eigvals (str) – Key in adata.uns storing eigenvalues. (default: ‘eigenvalues’)

  • eigvecs (str) – Key in adata.uns storing eigenvectors. (default: ‘eigenvectors’)

  • scale (float) – Scaling factor for signal amplification. Higher values increase imputed signal and maximum expression values. (default: 1.0)

  • copy (bool) – Whether to return a copy of adata. If False, updates are made in-place. (default: False)

Returns:

  • If copy=False: Updates adata with the imputed count matrix in adata.X.

  • If copy=True: Returns a modified copy of adata with the imputed count matrix.

Return type:

anndata.AnnData or None

Example:
import ocelli as oci
from anndata import AnnData
import numpy as np

# Example data
adata = AnnData(X=np.random.rand(100, 50))
adata.var.index = [f"gene_{i}" for i in range(50)]

# Add mock eigenvalues and eigenvectors
adata.uns["eigenvalues"] = np.random.rand(10)
adata.uns["eigenvectors"] = np.random.rand(100, 10)

# Perform imputation
oci.tl.imputation(adata, t=5, kappa=1.5, features=["gene_1", "gene_2"])
Contents