PICASSO: Phylogenetic Inference of Copy number Alterations in Single-cell Sequencing data Optimization

PyPI version Python versions Documentation Status License

PICASSO is a Python package for reconstructing tumor phylogenies from noisy, inferred copy number alteration (CNA) data derived from single-cell RNA sequencing (scRNA-seq). Unlike methods designed for direct single-cell DNA sequencing data, PICASSO is specifically optimized to handle the uncertainty and noise inherent in CNAs inferred from gene expression profiles.

Key Features

Noise-Aware Phylogeny Reconstruction

  • Handles uncertainty in scRNA-seq-inferred CNA data

  • Probabilistic assignment with confidence thresholds

  • Robust to technical artifacts and dropout events

Flexible Tree Building

  • Iterative binary splitting with categorical mixture models

  • Multiple termination criteria (BIC, confidence-based, chi-squared)

  • Customizable depth and clone size constraints

Comprehensive Analysis

  • Clone aggregation and modal profile generation

  • Evolutionary change inference along tree branches

  • Integration with iTOL for publication-ready visualizations

Designed for Single-Cell Data

  • Optimized for the specific challenges of scRNA-seq CNA inference

  • Handles variable clone sizes and imbalanced datasets

  • Prevents over-fitting to noise patterns

Quick Start

Installation

Install PICASSO from PyPI:

pip install picasso-phylo

Basic Usage

from picasso import Picasso, load_data

# Load example CNA data
cna_data = load_data()

# Initialize PICASSO with noise-appropriate parameters
picasso = Picasso(cna_data,
                 min_clone_size=10,  # Larger for noisy data
                 assignment_confidence_threshold=0.8)

# Reconstruct phylogeny
picasso.fit()

# Get results
phylogeny = picasso.get_phylogeny()
clone_assignments = picasso.get_clone_assignments()

# Analyze results
from picasso import CloneTree
tree_analyzer = CloneTree(phylogeny, clone_assignments, cna_data)
tree_analyzer.plot_alterations()

Algorithm Overview

PICASSO uses an iterative binary splitting approach:

  1. Initialization: All cells start in a single root clone

  2. Iterative Splitting: At each depth level:

    • Fit Categorical Mixture Models with k=1 and k=2 components for each clone

    • Evaluate splitting criteria (BIC or assignment confidence)

    • Split clones that meet the criteria into two daughter clones

  3. Termination: Stop when no clones meet splitting criteria or constraints are reached

  4. Tree Construction: Build phylogenetic tree from the clone hierarchy

The algorithm is specifically designed to handle:

  • Noise and artifacts in scRNA-seq-inferred CNAs

  • Uncertainty in copy number state assignments

  • Variable clone sizes and imbalanced data

  • Over-fitting to noise patterns through confidence-based termination

Citation

If you use PICASSO in your research, please cite our preprint while our manuscript is under review:

@article{picasso2025,
title={Transcriptomic plasticity is a hallmark of metastatic pancreatic cancer},
author={Jim{'e}nez-S{'a}nchez, Alejandro and Persad, Sitara and Hayashi, Akimasa and Umeda, Shigeaki and Sharma, Roshan and Xie, Yubin and Mehta, Arnav and Park, Wungki and Masilionis, Ignas and Chu, Tinyi and Zhu, Feiyang and Hong, Jungeui and Chaligne, Ronan and O'Reilly, Eileen M. and Mazutis, Linas and Nawy, Tal and Pe'er, Itsik and Iacobuzio-Donahue, Christine A. and Pe'er, Dana},
journal={bioRxiv},
year={2025},
month={February},
day={28},
doi={10.1101/2025.02.28.640922},
url={https://www.biorxiv.org/content/10.1101/2025.02.28.640922v1},
note={Preprint}
}

Documentation Contents

User Guide

API Reference

Development

Support

License

PICASSO is released under the MIT License. See the LICENSE file for details.

Indices and tables