scDesign3Py.scDesign3

class scDesign3Py.scDesign3(n_cores: int = 1, parallelization: Literal['mcmapply', 'bpmapply', 'pbmcmapply'] = 'mcmapply', bpparam: rpy2.robjects.methods.RS4 | None = None, return_py: bool = True)

Python interface for scDesign3

All functions are arranged in scDesign3 class for easy reuse generated results.

Attributes:

sce:

SingleCellExperiment R object changed from the given Anndata object

assay_use:

The name of the assay used for modeling

all_covar:

All covariates (explainary variables) used to model gene expression pattern

family_use:

The set marginal distribution of each gene.

copula:

The copula type you set.

n_cores:

The number of cores used for model fitting.

parallelization:

The parallelization method.

bpparam:

If @parallelization is ‘bpmapply’, the corresponding R object to set the parallelization parameters.

return_py:

Whether the functions will return a result easy for manipulation.

construct_data_res:

Result of calling @construct_data

fit_marginal_res:

Result of calling @fit_marginal

fit_copula_res:

Result of calling @fit_copula

model_paras:

Result of calling @extract_para

simu_res:

Result of calling @simu_new

whole_pipeline_res:

Result of calling @scdesign3

Methods

__init__([n_cores, parallelization, ...])	Decide basic settings when running the class methods.
construct_data(anndata, corr_formula[, ...])	Construct the input data
extract_para([data, new_covariate, ...])	Extract the parameters of each cell's distribution
fit_copula([input_data, copula, ...])	Fit the copula model
fit_marginal(mu_formula, sigma_formula, ...)	Fit the marginal models
scdesign3(anndata, corr_formula, mu_formula, ...)	The wrapper for the whole scDesign3 pipeline
set_r_random_seed(seed)	Set the R random seed
simu_new([mean_mat, sigma_mat, zero_mat, ...])	Simulate new data