Results Storage

Running artifacts

Keys of embedding results

1. The embedding results are stored in adata.obsm['X_rep'] for the scModel and crossModel objects.

2. The embedding results with maunual annotation refinement are stored in adata.obsm['X_anchord'] for the scModel and crossModel objects.

3. The embedding results with spatially smoothed are stored in adata.obsm['X_smoothed'] for the stModel objects.

4. The embedding results without spatially smoothed are stored in adata.obsm['X_unsmoothed'] for the stModel objects (only when not running on e2e mode).

Keys of spatial domain identification results

1. The spatial domain identification results are stored in adata.obs['domain'] for the stModel and crossModel objects.

2. The spatial sub-domain identification results are stored in adata.obs['sub_domain'] for the stModel and crossModel objects.

Keys of deconvolution results

1. The normalized (sum=1) deconvolution results are stored in adata.obsm['deconv'] for the crossModel objects.

2. The un-normalized deconvolution results are stored in adata.obs['deconv_abundance'] for the crossModel objects.

Keys of batch-corrected and imputed gene expression profiles

1. The batch-corrected gene expression profiles are stored in adata.layers['corrected_counts'] for the scModel and crossModel objects.

2. The imputed gene expression profiles are stored in adata.layers['imputed_counts'] for the scModel and crossModel objects.

Saving and loading artifacts

Saving artifacts

All the artifacts can be saved by the save method of the scModel, stModel, and crossModel objects. The saved artifacts include the adata object, the model weights, and the model configurations.

Loading artifacts

All the artifacts can be loaded by the load method of the scModel, stModel, and crossModel objects. The loaded artifacts include the adata object, the model weights, and the model configurations.