snapatac2.tl.add_regr_scores#
- snapatac2.tl.add_regr_scores(network, *, peak_mat=None, gene_mat=None, select=None, method='elastic_net', scale_X=False, scale_Y=False, alpha=1.0, l1_ratio=0.5, use_gpu=False, overwrite=False)[source]#
Perform regression analysis for nodes and their parents in the network.
- Parameters:
network (
PyDiGraph
) – networkpeak_mat (
Union
[AnnData
,AnnDataSet
,None
]) – AnnData or AnnDataSet object storing the cell by peak count matrix, where the.var_names
contains peaks.gene_mat (
Union
[AnnData
,AnnDataSet
,None
]) – AnnData or AnnDataSet object storing the cell by gene count matrix, where the.var_names
contains genes.select (
Optional
[list
[str
]]) – Run this for selected genes only.method (
Literal
['gb_tree'
,'elastic_net'
]) – Regresson model.scale_X (
bool
) – Whether to scale the features.scale_Y (
bool
) – Whether to scale the response variable.alpha (
float
) – Constant that multiplies the penalty terms in ‘elastic_net’.l1_ratio (
float
) – Used in ‘elastic_net’. The ElasticNet mixing parameter, with0 <= l1_ratio <= 1
. Forl1_ratio = 0
the penalty is an L2 penalty. Forl1_ratio = 1
it is an L1 penalty. For0 < l1_ratio < 1
, the penalty is a combination of L1 and L2.use_gpu (
bool
) – Whether to use gpuoverwrite (
bool
) – Whether to overwrite existing records.