sccross.data.metacell_regr

sccross.data.metacell_regr(*adatas, use_rep=None, n_meta=None, skeleton=None, model='Lasso', **kwargs)

Metacell-based regression

Parameters

• *adatas – Datasets to be correlated, where .X are raw counts (indexed by domain name)

• use_rep (Optional[str]) – Data representation based on which to cluster meta-cells

• n_meta (Optional[int]) – Number of metacells to use

• skeleton (Optional[DiGraph]) – Skeleton graph determining which pair of features to correlate

• model (str) – Regression model (should be a class name under sklearn.linear_model)

• **kwargs – Additional keyword arguments are passed to the regression model

Return type

DiGraph

Returns

regr – A skeleton-based graph containing regression weights as edge attribute “regr”