scSLAT.model.loaddata.load_anndatas

scSLAT.model.loaddata.load_anndatas(adatas, feature='DPCA', dim=50, self_loop=False, join='inner', backend='sklearn', singular=True, check_order=True, n_top_genes=2500)

Transfer adatas with spatial info into PyG datasets

Parameters:

• adatas (typing.List[anndata._core.anndata.AnnData]) – List of Anndata objects

• feature (typing.Optional[str]) – use which data to build graph - PCA (default) - DPCA (For batch effect correction) - Harmony (For batch effect correction) - GLUE (NOTE: only suitable for multi-omics integration)

• dim (typing.Optional[int]) – dimension of embedding, works for [‘PCA’, ‘DPCA’, ‘Harmony’, ‘GLUE’]

• self_loop (typing.Optional[bool]) – whether to add self loop on graph

• join (typing.Optional[str]) – how to concatenate two adata

• backend (typing.Optional[str]) – backend to calculate DPCA

• singular (typing.Optional[bool]) – whether to multiple singular value in DPCA

• check_order (typing.Optional[bool]) – whether to check the order of adata1 and adata2

• n_top_genes (typing.Optional[int]) – number of highly variable genes

• Note –

• ---------- –

• yet (Only support 'Spatial_Net' which store in adata.uns) –

Return type:

typing.List[torch_geometric.data.data.Data]