nichepca.workflows.nichepca

Contents

nichepca.workflows.nichepca#

nichepca.workflows.nichepca(adata, knn=None, radius=None, delaunay=False, n_comps=30, obs_key=None, obsm_key=None, sample_key=None, pipeline=('norm', 'log1p', 'agg', 'pca'), norm_per_sample=True, backend='pyg', aggr='mean', allow_harmony=True, max_iter_harmony=50, remove_graph=False, **kwargs)#

Run the general NichePCA workflow.

Parameters:

  • adata (AnnData) – The input AnnData object.

  • knn (int | None (default: None)) – Number of nearest neighbors to use for graph construction.

  • radius (float | None (default: None)) – Radius for graph construction.

  • delaunay (bool (default: False)) – Whether to use Delaunay triangulation for graph construction.

  • n_comps (int (default: 30)) – Number of principal components to compute.

  • obs_key (str | None (default: None)) – Observation key to use for generating a new AnnData object.

  • obsm_key (str | None (default: None)) – Observation matrix key to use as input.

  • sample_key (str | None (default: None)) – Sample key to use for multi-sample graph construction.

  • pipeline (tuple | list (default: ('norm', 'log1p', 'agg', 'pca'))) – Pipeline of steps to perform. Must include ‘agg’.

  • norm_per_sample (bool (default: True)) – Whether to normalize per sample.

  • backend (str (default: 'pyg')) – Backend to use for aggregation.

  • aggr (str (default: 'mean')) – Aggregation method to use.

  • allow_harmony (bool (default: True)) – Whether to allow Harmony integration.

  • max_iter_harmony (int (default: 50)) – Maximum number of iterations for Harmony.

  • remove_graph (bool (default: False)) – Whether to remove the constructed graph from adata.uns after the workflow completes.

  • **kwargs (dict) – Additional keyword arguments.

Returns:

None

Contents