moscot.problems.space.MappingProblem.correlate

MappingProblem.correlate(var_names=None, corr_method='pearson', device=None, groupby=None, batch_size=None)

Correlate true and predicted gene expression.

Warning

Sparse matrices stored in X will be densified.

Parameters:

• var_names (Optional[Sequence[str]]) – Keys in var_names. If None, use all shared genes.

• corr_method (Literal['pearson', 'spearman']) –

  Correlation method. Valid options are:

  • 'pearson' - Pearson correlation.

  • 'spearman' - Spearman rank correlation.

• device (Union[Literal['cpu', 'gpu', 'tpu'], str, None]) – Device where to transfer the solutions, see to().

• groupby (Optional[str]) – Optional key in obs, containing categorical annotations for grouping.

• batch_size (Optional[int]) – Number of features to process at once. If None, process all features at once. Larger values will require more memory.

Return type:

Union[Mapping[Tuple[TypeVar(K, bound= Hashable), TypeVar(K, bound= Hashable)], Mapping[Any, Series]], Mapping[Tuple[TypeVar(K, bound= Hashable), TypeVar(K, bound= Hashable)], Series]]

Returns:

: Correlation for each solution in solutions.