COSIE.linkage_construction.split_raw_data
- split_raw_data(data_dict, spatial_loc_dict, n_x, n_y)[source]
Divide spatial omics data from each section into spatially partitioned subgraphs.
Each tissue section is split into a grid of n_x × n_y rectangular regions of equal size. This function is primarily used for scaling to large spatial datasets by enabling subgraph-level analysis.
Parameters
- data_dictdict
A dictionary where each key is a modality (e.g., ‘RNA’, ‘Protein’) and each value is a list of AnnData objects (one per tissue section). Each AnnData must include:
.X: Feature matrix
.obs, .var: Metadata
.obsm[‘spatial’]: 2D spatial coordinates
Use None if a modality is missing in a section.
- spatial_loc_dictdict
A dictionary mapping each section name (e.g., ‘s1’, ‘s2’, …) to a 2D NumPy array of shape (n_cells, 2), representing the spatial coordinates of each cell in the section.
- n_xint
Number of spatial partitions along the x-axis.
- n_yint
Number of spatial partitions along the y-axis.
Returns
- sub_data_dictdict
A nested dictionary organizing modality-specific AnnData subgraphs by section and region.
Format:
- {
- ‘s1’: {
0: {‘RNA’: AnnData, ‘Protein’: AnnData, …}, # First spatial region of section s1
1: {…}, …
},
- ‘s2’: {
…
}
}