SPACEL.Scube.alignment.align

SPACEL.Scube.alignment.align(ad_list, cluster_key='spatial_domain', output_path=None, raw_loc_key='spatial', aligned_loc_key='spatial_aligned', n_neighbors=15, knn_exclude_cutoff=None, p=2, a=1, bound_alpha=1, write_loc_path=None, n_threads=1, seed=42, subset_prop=None, *args, **kwargs)

Pairwise alignment.

Pairwise align the slices in ad_list. The aligned coordinates are saved in .obsm[aligned_loc_key] in each slices of ad_list.

Parameters:

  • ad_list – A list containing all slice data in AnnData object.

  • cluster_key – A string representing one column of obs in AnnData object, containing the spatial domain information used for alignment.

  • output_path – A string representing the path directory where the alignment parameters are saved. If None, it will be ‘Scube_outputs’.

  • raw_loc_key – A string representing one key of obsm in AnnData object of each slice in ad_list, containing the raw coordinates.

  • aligned_loc_key – A string written to a key of obsm in AnnData object of each slice in ad_list, containing the aligned coordinates.

  • n_neighbors – A number of neighbors in target slices considered by each spot/cell in source slices.

  • knn_exclude_cutoff – A number used to filter the neighbors in MNN. The neighbor will be exclude when the distance of neighbors larger than the median of neareast n_neighbors + knn_exclude_cutoff neighbors distance in all spots/cells in target slice. If None, it will automatically default to n_neighbors.

  • p – Degree of the penalty function.

  • a – Coefficient of the penalty function.

  • bound_alpha – For the optimized boundary, the multiplier based on the maximum and minimum values of the slice coordinates.

  • write_loc_path – A string representing the path directory where the aligned coordinates of all slices are saved. If None, it won’t be saved.

  • n_threads – The number of parallel threads for the optimization algorithm.

  • seed – Seed for the optimization algorithm.

  • subset_prop – The downsampling ratio for cells in each slice.

Returns:

None