stereo.core.StPipeline.sctransform#

StPipeline.sctransform(n_cells=5000, n_genes=2000, filter_hvgs=True, var_features_n=3000, inplace=True, res_key='sctransform', exp_matrix_key='scale.data', seed_use=1448145, filter_raw=True, **kwargs)[source]#

Normalization of scTransform, refering to Seurat [Hafemeister19].

Parameters:

  • n_cells (int) – number of cells to use for estimating parameters.

  • n_genes (int) – number of genes to use for estimating parameters. means all genes.

  • filter_hvgs (bool) – True to retain data associated with highly variable genes only while False to entire data.

  • var_features_n (int) – the number of variable features to select, for calculating a subset of pearson residuals.

  • inplace (bool) – whether to replace the previous expression data.

  • res_key (str) – the key to get targeted result from self.result.

  • exp_matrix_key (str) – which expression matrix to use for analysis.

  • seed_use (int) – random seed.

  • filter_raw (Optional[bool]) – because this function will filter data, whether to filter raw data meanwhile by setting filter_raw.

Returns:

  • An object of StereoExpData.

  • Depending on inplace, if True, the data will be replaced by those normalized.