🍑 PEACH: Deep Archetypal Analysis for Single-Cell Biology

PEACH is a deep learning framework for archetypal analysis of single-cell data, following scverse conventions.

Getting Started

• Installation
  • Requirements
  • Installation Options
  • Verify Installation
  • GPU Support (Optional)
  • Troubleshooting
• 🍑 PEACH Documentation
  • Quick Start
  • Tutorials
  • Workflow Scripts
  • API Structure
  • Developer Reference
  • Additional Resources

Tutorials

• Tutorials
  • Core Workflows
  • Tutorial Overview
  • Getting Started
• WORKFLOW 01: Data Loading & PCA Preprocessing
  • Configuration
  • Step 1: Load Data
  • Step 2: Quality Control
  • Step 3: Normalization
  • Step 4: Principal Component Analysis (PCA)
  • Summary
• WORKFLOW 02: Hyperparameter Search with Cross-Validation
  • Configuration
  • Step 1: Load Data with PCA
  • Step 2: Run Hyperparameter Search
  • Step 3: Analyze Results
  • Step 4: Visualize Results with Elbow Plot
  • Summary
• WORKFLOW 03: Archetypal Model Training
  • Configuration
  • Step 1: Load Data with PCA
  • Step 2: Train Archetypal Model
  • Step 3: Examine Training Results
  • Step 4: Accessing the Trained Model
  • Summary
• WORKFLOW 04: Archetype Coordinates & Cell Assignment
  • Configuration
  • Step 1: Load Data and Train Model (Prerequisites)
  • Step 2: Compute Archetype Distances (Geometric)
  • Step 3: Extract Archetype Weights (Model-based)
  • Step 4: Assign Cells to Archetypes (Categorical)
  • Step 5: Compare Distance vs Weight Assignments
  • Summary
• WORKFLOW 05: Gene & Pathway Enrichment Analysis
  • Configuration
  • Step 1: Prepare Data with Model and Assignments (Prerequisites)
  • Step 2: Compute Gene Associations
  • Step 3: Compute Pathway Associations (Optional)
  • Step 4: Export Results
  • Summary
• WORKFLOW 06: CellRank Integration for Lineage Analysis
  • Configuration
  • Step 1: Prepare Data with Model and Assignments
  • Step 2: Setup CellRank with ConnectivityKernel
  • Step 3: Compute Lineage Pseudotimes
  • Step 4: Compute Lineage Drivers
  • Step 5: Compute Transition Frequencies
  • Step 6: Visualization
  • Summary
• WORKFLOW 08: Comprehensive Visualization
  • Configuration
  • Step 1: Prepare Complete Analysis (All Prerequisites)
  • Step 2: Archetypal Space Visualizations
  • Step 3: Training Diagnostics
  • Step 4: Gene/Pattern Visualizations
  • Step 5: Archetype Statistics
  • Step 6: CellRank Visualizations (if available)
  • Summary
• Tutorial 09: scATAC-seq Archetypal Analysis
  • Step 1: Load scATAC-seq Data
  • Step 2: TF-IDF + LSI Preprocessing
  • Step 3: Hyperparameter Search
  • Step 4: Train Final Model
  • Step 5: Archetype Distances, Positions & Assignments
  • Step 6: Characterization
  • Summary
• Tutorial 10: Spatial Archetypal Analysis
  • Step 1: Load Spatial Data
  • Step 2: Preprocessing
  • Step 3: Hyperparameter Search
  • Step 4: Train Final Model
  • Step 5: Distances, Positions & Assignments
  • Step 6: Spatial Neighbor Graph
  • Step 7: Neighborhood Enrichment
  • Step 8: Distance-Dependent Co-occurrence
  • Step 9: Spatial Archetype Map
  • Step 10: Spatial Autocorrelation
  • Step 11: Interaction Boundaries
  • Step 12: Gene Characterization
  • Summary

API Reference

• API Reference
  • peach.pp
  • peach.tl
  • peach.pl
  • Preprocessing (pp)
  • Tools (tl)
  • Plotting (pl)
  • Configuration

Development

• Contributing

Indices and tables

• Index

• Module Index

• Search Page