scglue.models.fit_SCGLUE

scglue.models.fit_SCGLUE(adatas, graph, model=SCGLUEModel, skip_balance=False, init_kws=None, compile_kws=None, fit_kws=None, balance_kws=None)

Fit GLUE model to integrate single-cell multi-omics data

Parameters:

• adatas (typing.Mapping[str, AnnData]) – Single-cell datasets (indexed by modality name)

• graph (Graph) – Guidance graph

• model (type) – Model class, must be one of {scglue.models.scglue.SCGLUEModel, scglue.models.scglue.PairedSCGLUEModel}

• skip_balance (bool) – Skip the balancing step and use equal weight for all cells

• init_kws (typing.Optional[typing.Mapping[str, typing.Any]]) – Model initialization keyword arguments (see the constructor of the model class, either scglue.models.scglue.SCGLUEModel or scglue.models.scglue.PairedSCGLUEModel)

• compile_kws (typing.Optional[typing.Mapping[str, typing.Any]]) – Model compile keyword arguments (see the compile method of the model class, either scglue.models.scglue.SCGLUEModel.compile() or scglue.models.scglue.PairedSCGLUEModel.compile())

• fit_kws (typing.Optional[typing.Mapping[str, typing.Any]]) – Model fitting keyword arguments (see scglue.models.scglue.SCGLUEModel.fit())

• balance_kws (typing.Optional[typing.Mapping[str, typing.Any]]) – Balancing weight estimation keyword arguments (see scglue.data.estimate_balancing_weight())

Return type:

scglue.models.scglue.SCGLUEModel

Returns:

model – Fitted model object