API

Import the spapros API as follows:

import spapros as sp

Selection

Main Class

se.ProbesetSelector(adata, celltype_key[, ...])

General class for probeset selection.

Other Functions

`se.select_reference_probesets`(adata, n[, ...])	Select reference probesets with basic selection methods.
`se.select_pca_genes`(adata, n[, ...])	Select n features based on pca loadings.
`se.select_DE_genes`(adata, n[, per_group, ...])	Select genes based on wilxocon rank genes test.

Evaluation

Main Class

ev.ProbesetEvaluator(adata[, celltype_key, ...])

General class for probe set evaluation, comparison, plotting.

Other Functions

`ev.get_metric_default_parameters`()	Get the default metric parameters.
`ev.forest_classifications`(adata, selection)	Train best trees including secondary trees.
`ev.single_forest_classifications`(adata, ...)	Compute or load decision tree classification results.

Utility functions

Expression Constraints

`ut.get_expression_quantile`(adata[, q, ...])	Compute each genes q'th quantile on normalised (and log1p) data.
`ut.transfered_expression_thresholds`(adata[, ...])	Transfer expression thresholds between different normalisations.
`ut.plateau_penalty_kernel`(var[, x_min, x_max])	Return penalty function.

Spatial Data Analysis

Plotting

The sp.pl functions are used directly only in rare cases, instead always try to use the wrapper methods from the ProbesetEvaluator and ProbesetSelector (see 2nd table below).

`pl.masked_dotplot`(adata, selector[, ct_key, ...])	Create dotplot with additional annotation masks.
`pl.clf_genes_umaps`(adata, df[, basis, ...])	Plot umaps of genes needed for cell type classification of each cell type.
`pl.selection_histogram`(adata, ...[, ...])	Plot histogram of quantiles for selected genes for different penalty kernels.
`pl.gene_overlap`(selection_df[, style, ...])	Plot overlap of gene sets
`pl.correlation_matrix`(set_ids, cor_matrices)	Plot heatmap of gene correlation matrix.
`pl.summary_table`(table[, summaries, ...])	Plot table of summary statistics
`pl.cluster_similarity`(selections_info[, ...])	Plot cluster similarity as NMI over number of clusters
`pl.knn_overlap`(selections_info[, data, ...])	Plot mean knn overlap over k
`pl.confusion_matrix`(set_ids, conf_matrices)	Plot heatmap of cell type classification confusion matrices.
`pl.marker_correlation`(marker_corr[, ...])	Plot maximal correlations with marker genes

pl module	`ProbesetSelector`	`ProbesetEvaluator`
`pl.masked_dotplot()`	see pl.masked_dotplot for now
`pl.clf_genes_umaps()`	`ProbesetSelector.plot_clf_genes()`
`pl.selection_histogram()`	`ProbesetSelector.plot_histogram()`
`pl.gene_overlap()`	`ProbesetSelector.plot_gene_overlap()`	potentially added soon
`pl.correlation_matrix()`	`ProbesetSelector.plot_coexpression()`	`ProbesetEvaluator.plot_coexpression()`
`pl.summary_table()`		`ProbesetEvaluator.plot_summary()`
`pl.cluster_similarity()`		`ProbesetEvaluator.plot_cluster_similarity()`
`pl.knn_overlap()`		`ProbesetEvaluator.plot_knn_overlap()`
`pl.confusion_matrix()`		`ProbesetEvaluator.plot_confusion_matrix()`
`pl.marker_correlation()`		`ProbesetEvaluator.plot_marker_corr()`