CelLink.model.Cellink

class CelLink.model.Cellink(full_ann1, full_ann2, shared_ann1, shared_ann2, partition=True)[source]

Core Cellink class

Methods

`alignment`	Perform cell-cell alignment (stage I and stage II).
`balanced_ot`	stage I of celLink: perform balanced optimal transport, filter out the matched cells and retain the unmatched cells.
`construct_graph`	Compute k-nearest neighbors of data and return the UMAP graph.
`iterative_unbalanced_ot`	stage II of celLink: perform iterative unbalanced optimal transport to correct alignments, filter out the matched cells and retain the unmatched cells.
`split_into_batches`	PS: put the modality that has larger cell population into the second position of the arr list.
`synchronize_imputed_to_initial`	Re-mapp the imputed feature profiles to align with the original data indices, which are shuffled when partitioning.
`visualize_integration`	Joint embed the arr and arr_imputed from alignment.