API Reference

PEACH provides a scVerse-compatible API with three main modules:

peach.pp	Preprocessing functions for archetypal analysis.
peach.tl	Tools for archetypal analysis.
peach.pl	Plotting functions for archetypal analysis.

Preprocessing (pp)

peach.pp.load_data(path[, use_raw, ...])	Load AnnData for archetypal analysis.
peach.pp.generate_synthetic([n_points, ...])	Generate synthetic convex data for testing.
peach.pp.prepare_training(adata[, ...])	Create DataLoader from AnnData for training with HPC optimizations.
peach.pp.load_pathway_networks([sources, ...])	Load pathway networks from MSigDB or OmniPath.
peach.pp.compute_pathway_scores(adata[, ...])	Compute pathway activity scores using MSigDB pathways.

Tools (tl)

Training & Coordinates

peach.tl.train_archetypal(adata[, ...])	Train Deep Archetypal Analysis model to discover cellular archetypes.
peach.tl.hyperparameter_search(adata, *[, ...])	Perform cross-validation hyperparameter search for archetypal analysis.
peach.tl.archetypal_coordinates(adata, *[, ...])	Extract archetypal coordinates for all cells.
peach.tl.assign_archetypes(adata, *[, ...])	Assign cells to archetypes based on distances.
peach.tl.extract_archetype_weights(adata[, ...])	Extract cell archetype weights from trained Deep_AA model.
peach.tl.compute_conditional_centroids(...)	Compute centroid positions in PCA space for each level of a categorical condition.
peach.tl.assign_to_centroids(adata, ...[, ...])	Assign cells to nearest centroid based on distance (top bin_prop% closest).

Statistical Analysis

peach.tl.gene_associations(adata, *[, ...])	Test gene expression associations with archetypal assignments.
peach.tl.pathway_associations(adata, *[, ...])	Test pathway activity associations with archetypal assignments.
peach.tl.conditional_associations(adata, *, ...)	Test associations between archetypes and categorical metadata.

Pattern Analysis

peach.tl.pattern_analysis(adata, *[, ...])	Comprehensive archetypal pattern analysis.
peach.tl.archetype_exclusive_patterns(adata, *)	Identify features exclusively high in single archetypes.
peach.tl.specialization_patterns(adata, *[, ...])	Identify specialization features relative to centroid archetype.
peach.tl.tradeoff_patterns(adata, *[, ...])	Identify mutual exclusivity and tradeoff patterns.

CellRank Integration

peach.tl.setup_cellrank(adata[, ...])	Set up CellRank workflow for archetypal or centroid-based trajectory analysis.
peach.tl.compute_lineage_pseudotimes(adata)	Convert fate probabilities to lineage-specific pseudotimes.
peach.tl.compute_lineage_drivers(adata, lineage)	Identify genes driving commitment to a specific lineage.
peach.tl.compute_transition_frequencies(adata)	Compute frequency of transitions between archetypal states.
peach.tl.single_trajectory_analysis(adata, ...)	Analyze single archetype-to-archetype trajectory.

Plotting (pl)

Core Visualizations

peach.pl.archetypal_space(adata, *[, ...])	Visualize cells in 3D archetypal coordinate space.
peach.pl.archetypal_space_multi(adata_list, *)	Compare multiple archetypal analysis fits in 3D PCA space.
peach.pl.training_metrics(history, *[, ...])	Visualize training metrics over epochs.
peach.pl.elbow_curve(cv_summary, *[, metrics])	Plot elbow curves for hyperparameter selection.

Archetype Analysis

peach.pl.archetype_positions(adata, *[, ...])	Visualize archetype positions in PCA space with distance matrix.
peach.pl.archetype_positions_3d(adata, *[, ...])	Visualize archetype positions in 3D PCA space.
peach.pl.archetype_statistics(adata, *[, ...])	Compute and display statistics about archetype positions.
peach.pl.dotplot(results_df, *[, x_col, ...])	Create dotplot visualization for statistical results.

Pattern Visualization

peach.pl.pattern_dotplot(pattern_df[, ...])	Create dotplot for pattern analysis results.
peach.pl.pattern_summary_barplot(pattern_results)	Create summary barplot showing pattern counts across different analyses.
peach.pl.pattern_heatmap(pattern_df, adata)	Create heatmap showing pattern expression across archetypes.

Trajectory Analysis

peach.pl.fate_probabilities(adata[, ...])	Plot fate probabilities on embedding.
peach.pl.lineage_drivers(adata, lineage[, ...])	Plot heatmap of top driver genes for a lineage.

Configuration

peach.tl.SearchConfig([n_archetypes_range, ...])

Configuration for hyperparameter search space and CV settings.