xenium_to_domain#

xenium_to_domain(path_to_xenium_data=None, domain_name='Default_xenium_domain', cells_from_selection_csv=None, load_transcripts=False, selected_transcripts=None, load_nuclei=False, load_cells_as_shapes=True, exclude_no_nuclei_cells=True, add_transcript_counts_to_cell=True, print_summary=True)#

The function will take a xenium dataset and produce a muspan domain loaded with cell and transcript objects from an xenium output. Each data will be linked using string cell ID from Xenium and given as a label on the object. The function will also load only the cells selected in the cells_from_selection_csv file which is obtained via annotation using Xenium explorer.

Parameters:
path_to_xenium_datastr, optional

String providing the path to the directory to the Xenium data folder, defaults to None.

domain_namestr, optional

Name given to created domain, defaults to ‘Default_xenium_domain’.

cells_from_selection_csvstr, optional

Path to a region csv exported from Xenium Explorer using the ‘download cell states as csv’ functionality, defaults to None.

load_transcriptsbool, optional

Option to load transcripts into the domain, defaults to False.

selected_transcriptslist of str, optional

A list of transcript names to be loaded into the domain. If None, all transcripts will be loaded, defaults to None.

load_nucleibool, optional

Option to load cell nuclei as shape objects into the domain, defaults to False.

load_cells_as_shapesbool, optional

Option to load cells as shapes. If False, cells will be loaded as point objects, defaults to True.

exclude_no_nuclei_cellsbool, optional

Option to exclude any cell object that does not contain a nucleus following segmentation, defaults to True.

add_transcript_counts_to_cellbool, optional

If True, each cell object is annotated with a label showing the transcript count for every gene in the panel. Labels follow the format ‘Transcript Counts: XYZ’, where XYZ is a transcript in the panel. Default is True

print_summarybool, optional

If True, print a summary of the resulting domain, by default True.

Returns:
muspan.domain

A muspan domain object.

Notes

Important

To import data from a Xenium experiment, we require the following files present the Xenium data folder (path_to_xenium_data).

Xenium data folder:
├── cells.parquet
├── nucleus_boundaries.parquet
├── cell_boundaries.parquet
├── transcripts.parquet
├── cell_feature_matrix.h5
└── gene_panel.json

These files are all generated as part of the standard Xenium output following processing and segmentation. For more information, see Xenium documentation

Additional information

  • Transcript names are loaded as labels on the transcript objects with the label name ‘Transcript ID’.

  • Cell IDs are loaded as labels on the cell objects with the label name ‘Cell ID’.

  • Cell annotations (labels) are loaded as labels on the cell objects with the label name ‘Cluster ID’ and are only imported using the ‘cells_from_selection_csv’ file. If missing, no cell labels will be added.

On this page