API documentation¶

This section provides detailed API documentation for all public functions and classes in seSemiProfiler.

Initial Setup¶

scSemiProfiler_dev.initial_setup.initsetup(…)

Initial setup of the semi-profiling pipeline, including processing the bulk data, clustering for finding the initial representatives.

Get Representatives Single-cell (used in example)¶

scSemiProfiler_dev.get_eg_representatives.get_eg_representatives(name)

Used for acquiring representatives’ single-cell data in the example.

Single-cell Processing & Feature Augmentation¶

scSemiProfiler_dev.singlecell_process.scprocess(…)

Process the reprsentatives’ single-cell data, including preprocessing and feature augmentations.

Single-cell Inference¶

scSemiProfiler_dev.inference.scinfer(name, …)

Computationally infer the single-cell data of all non-representative samples (target samples) based on the cohort’s bulk data and the representatives’ single-cell data

Representatives Selection¶

scSemiProfiler_dev.representative_selection.activeselection(…)

Use active learning to select the next batch of representatives

Utils - Downstream Analysis¶

`scSemiProfiler_dev.utils.estimate_cost`(…)	Estimate the cost of semi-profiling and real-profiling.
`scSemiProfiler_dev.utils.visualize_recon`(…)	Visualize the performance of reconstruction by plotting the original and reconstructed data in the same UMAP.
`scSemiProfiler_dev.utils.visualize_inferred`(…)	Visualize the inference performance by plotting the representative, inferred target, and target ground truth in the same UMAP.
`scSemiProfiler_dev.utils.loss_curve`(name[, …])	Visualize the training loss curves
`scSemiProfiler_dev.utils.assemble_cohort`(…)	Assemble inferred sample data and representative sample data into semi-profiled cohort and annotate the celltype.
`scSemiProfiler_dev.utils.assemble_representatives`(name)	Assemble previous round of inferred representative data and annotate the cell type.
`scSemiProfiler_dev.utils.compare_umaps`(semidata)	Compare the real-profiled and semi-profiled datasets by plotting them in a same UMAP
`scSemiProfiler_dev.utils.compare_adata_umaps`(…)	Compare the real-profiled and semi-profiled datasets by plotting them in a same UMAP
`scSemiProfiler_dev.utils.celltype_proportion`(…)	Compute the cell type proportion in a dataset
`scSemiProfiler_dev.utils.composition_by_group`(adata)	Visualizing the cell type composition in each group.
`scSemiProfiler_dev.utils.geneset_pattern`(…)	Generate heatmaps for visualizing gene set activation pattern in a dataset.
`scSemiProfiler_dev.utils.celltype_signature_comparison`(…)	Use dotplot to compare the cell type signatures found using the real-profiled dataset and the semi-profiled datset.
`scSemiProfiler_dev.utils.rrho`(gtdata, …)	Use RRHO graph to compare the positive and negative markers found using real-profiled and semi-profiled datasets.
`scSemiProfiler_dev.utils.enrichment_comparison`(…)	Compare the enrichment analysis results using the real-profiled and semi-profiled datasets.
`scSemiProfiler_dev.utils.get_error`(name)	Conclude the semi-profiling history of a project and output the erros, upperbounds, and lower bounds, which are necessary for overall performance evaluation.
`scSemiProfiler_dev.utils.errorcurve`(…[, …])	Visualize the error and cost as more representatives are sequenced.

Utils - Statistics¶

`scSemiProfiler_dev.utils.comb`(a, b)	Combination number
`scSemiProfiler_dev.utils.hyperp`(N, n1, n2, k)	Returns the cdf of a hypergeometric test.
`scSemiProfiler_dev.utils.hypert`(N, n1, n2, k)	Returns the p-value of a hypergeometric test.
`scSemiProfiler_dev.utils.faiss_knn`(query, x)	Compute distances from a vector to its K-nearest neighbros in a matrix.