omicverse.pp.scale

Contents

omicverse.pp.scale#

omicverse.pp.scale(adata, max_value=10, layers_add='scaled', to_sparse=False, **kwargs)[source]#

Scale the input AnnData object.

Parameters:

  • adata – Annotated data matrix with n_obs x n_vars shape.

  • max_value (default: 10) – Maximum value after scaling. Default: 10.

  • layers_add (default: 'scaled') – Name of the layer to store the scaled data. Default: ‘scaled’. OOM backend only supports layers_add=’scaled’ (the underlying chunked_scale writes a lazy ScaledBackedArray to that fixed key); passing anything else raises ValueError.

  • to_sparse (default: False) – If True, convert the result to csr_matrix format. Default: False. Scaled data is 100% dense, so sparse storage only adds overhead.

  • **kwargs – Additional arguments passed to scaling functions.

Returns:

Annotated data matrix with n_obs x n_vars shape. Adds a new layer called ‘scaled’ that stores the expression matrix that has been scaled to unit variance and zero mean.

Return type:

adata

Examples

>>> import omicverse as ov
>>> # Scale data with default sparse output
>>> ov.pp.scale(adata, max_value=10)
>>> # Scale data keeping dense format
>>> ov.pp.scale(adata, max_value=10, to_sparse=False)
Contents