scSLAT.model.prematch.utils.perturb_data

scSLAT.model.prematch.utils.perturb_data(adata, noise='nb', inverse_noise=5)[source]

Add noise to the count matrix.

Parameters:

adata (anndata._core.anndata.AnnData) – Annotated data matrix.
noise (str) – Type of noise to add. Default is nb for negative binomial noise.
inverse_noise (float) – Inverse of noise. Default is 5.

Return type:

anndata._core.anndata.AnnData

Returns:

Perturbed AnnData object.

Note

The raw count matrix is stored in adata.layers[“counts”].