Tutorials#
Learn how to use PEACH for archetypal analysis of single-cell data through comprehensive tutorials.
Core Workflows#
- WORKFLOW 01: Data Loading & PCA Preprocessing
- WORKFLOW 02: Hyperparameter Search with Cross-Validation
- WORKFLOW 03: Archetypal Model Training
- WORKFLOW 04: Archetype Coordinates & Cell Assignment
- WORKFLOW 05: Gene & Pathway Enrichment Analysis
- WORKFLOW 06: CellRank Integration for Lineage Analysis
- WORKFLOW 08: Comprehensive Visualization
- Tutorial 09: scATAC-seq Archetypal Analysis
- Tutorial 10: Spatial Archetypal Analysis
Tutorial Overview#
01 - Data Loading & Preprocessing#
Load single-cell data, perform quality control, and prepare for archetypal analysis.
02 - Hyperparameter Search#
Optimize model architecture using cross-validation and grid search.
03 - Model Training#
Train the Deep Archetypal Analysis model with optimal parameters.
04 - Archetype Coordinates#
Extract cell-archetype distances, weights, and assign cells to archetypes.
05 - Gene Enrichment#
Perform differential expression and pathway enrichment analysis per archetype.
06 - CellRank Integration#
Integrate with CellRank for lineage tracing and trajectory analysis.
08 - Visualization#
Create publication-ready 3D archetypal space plots and statistical visualizations.
09 - scATAC-seq Analysis#
TF-IDF + LSI preprocessing for chromatin accessibility data, then full archetypal analysis pipeline with pca_key='X_lsi'.
10 - Spatial Analysis#
Archetypal analysis on spatial transcriptomics data (Slide-seq, MERFISH, Visium) with neighborhood enrichment and distance-dependent co-occurrence testing via squidpy.
Getting Started#
Each tutorial is self-contained and can be run as a Jupyter notebook. Start with Tutorial 01 for a complete walkthrough, or jump to specific topics as needed.
# Install peach
pip install peach
# Run tutorials
jupyter lab docs/tutorials/