cpa.CPA.get_latent_representation

CPA.get_latent_representation(adata=None, indices=None, batch_size=32)

Returns All latent representations for the given dataset

Parameters:

adata Optional[AnnData], optional

[description], by default None

indices Optional[Sequence[int]], optional

Optional indices, by default None

batch_size Optional[int], optional

Batch size to use, by default None

Returns:

latent_outputs : Dict[str, anndata.AnnData] Dictionary of latent representations containing:

• ’latent_corrected’: batch-corrected (if batch_key is set) latent representation

• ’latent_basal’: basal latent representation

• ’latent_after’: final latent representation which can be used for decoding.