API

Import decoupler as:

import decoupler as dc

Preprocessing:

extract(mat[, use_raw, verbose, dtype])	Processes different input types so that they can be used downstream.
filt_min_n(c, net[, min_n])	Removes sources of a net with less than min_n targets.
match(c, r, net)	Matches mat with a regulatory adjacency matrix.
rename_net(net[, source, target, weight])	Renames input network to match decoupler's format (source, target, weight).
get_net_mat(net)	Transforms a given network to a regulatory adjacency matrix (targets x sources).

Methods:

run_aucell(mat, net[, source, target, n_up, ...])	AUCell.
run_gsea(mat, net[, source, target, times, ...])	Gene Set Enrichment Analysis (GSEA).
run_gsva(mat, net[, source, target, kcdf, ...])	Gene Set Variation Analysis (GSVA).
run_mdt(mat, net[, source, target, weight, ...])	Multivariate Decision Tree (MDT).
run_mlm(mat, net[, source, target, weight, ...])	Multivariate Linear Model (MLM).
run_ora(mat, net[, source, target, n_up, ...])	Over Representation Analysis (ORA).
run_udt(mat, net[, source, target, weight, ...])	Univariate Decision Tree (UDT).
run_ulm(mat, net[, source, target, weight, ...])	Univariate Linear Model (ULM).
run_viper(mat, net[, source, target, ...])	Virtual Inference of Protein-activity by Enriched Regulon (VIPER).
run_wmean(mat, net[, source, target, ...])	Weighted sum (WMEAN).
run_wsum(mat, net[, source, target, weight, ...])	Weighted sum (WSUM).
run_consensus(mat, net[, source, target, ...])	Consensus score from top methods.

Running multiple methods:

decouple(mat, net[, source, target, weight, ...])	Decouple function.
cons(res)	Consensus score between methods.
dense_run(func, mat, net[, source, target, ...])	Run a method without zero values.

General utils:

melt(df)	Function to generate a long format dataframe similar to the one obtained in the R implementation of decoupler.
show_methods()	Shows available methods.
check_corr(net[, source, target, weight, ...])	Checks the correlation across the regulators in a network.
get_toy_data([n_samples, seed])	Generate a toy mat and net for testing.
summarize_acts(acts, groupby[, obs, mode, ...])	Summarizes activities obtained per group by their mean or median and removes features that do not change across samples.
assign_groups(summary)	Assigns group labels based on summary activities.
dense_run(func, mat, net[, source, target, ...])	Run a method without zero values.
p_adjust_fdr(p)	Benjamini-Hochberg p-value correction for multiple hypothesis testing.
shuffle_net(net[, target, weight, seed, ...])	Shuffle network to make it random.

AnnData utils:

get_acts(adata, obsm_key)	Extracts activities as AnnData object.
get_pseudobulk(adata, sample_col, groups_col)	Generates an unormalized pseudo-bulk profile per sample and group.
get_contrast(adata, group_col, ...[, ...])	Computes Differential Expression Analysis using scanpy's rank_genes_groups function between two conditions from pseudo-bulk profiles.
get_top_targets(logFCs, pvals, contrast[, ...])	Return significant target features for a given source and contrast.
format_contrast_results(logFCs, pvals)	Formats the results from get_contrast into a long format data-frame.

Omnipath wrappers:

show_resources()	Shows available resources in Omnipath.
get_resource(name[, organism])	Wrapper to access resources inside Omnipath.
get_progeny([organism, top])	Pathway RespOnsive GENes for activity inference (PROGENy).
get_dorothea([organism, levels, weight_dict])	DoRothEA gene regulatory network.
translate_net(net[, columns, ...])	Translate networks between species by orthology.

Plotting

plot_volcano(logFCs, pvals, contrast[, ...])	Plot logFC and p-values.
plot_violins(mat[, thr, log, use_raw, ...])	Plot distribution of features' values per sample.
plot_barplot(acts, contrast[, top, ...])	Plot barplots showing the top absolute value activities.
plot_metrics_scatter(df[, x, y, groupby, ...])	Plot scatter plot of metrics across two different axes.
plot_metrics_scatter_cols(df, col[, x, y, ...])	Extension of the function plot_metrics_scatter to group metrics by two categories at the same time.
plot_metrics_boxplot(df, metric[, groupby, ...])	Plot boxplots showing the distribution of scores between methods for a metric.

Metrics

metric_auroc(y_true, y_score)	Area Under the Receiver Operating characteristic Curve (AUROC)
metric_auprc(y_true, y_score[, pi0])	Area Under the Precision-Recall Curve (AUPRC)
metric_mcauroc(y_true, y_score[, n_iter, seed])	Monte-Carlo Area Under the Receiver Operating characteristic Curve (AUROC)
metric_mcauprc(y_true, y_score[, n_iter, seed])	Monte-Carlo Area Under the Precision-Recall Curve (AUPRC)

Benchmark utils

get_toy_benchmark_data([n_samples, seed, ...])	Generate a toy mat, net and obs for testing the benchmark pipeline.
show_metrics()	Shows available evaluation metrics.

Benchmark

benchmark(mat, obs, net, perturb, sign[, ...])	Benchmark methods or networks on a given set of perturbation experiments using activity inference with decoupler.
format_benchmark_inputs(mat, obs, perturb, ...)
get_performances(res, obs, groupby, by, metrics)