TieDIE: Tied Diffusion for Subnetwork Discovery
A network analysis algorithm that finds subnetworks connecting genomic perturbations to transcriptional changes in large gene interaction networks.
Authors
Evan O. Paull, Daniel Carlin and Joshua M. Stuart (UC Santa Cruz)
Additional Contributors
- Srikanth Bezawada (TieDIE Cytoscape Plugin)
- Josh L. Espinoza (Quick kernel loading feature)
- Dana Silverbush (MATLAB kernel generation code updates)
- Denes Turei (modern tooling and packaging)
Requirements
- Python >= 3.9
- numpy >= 1.20
- scipy >= 1.0
- networkx >= 2.0
Installation
Using pip (from GitHub)
Using uv (recommended)
uv is a fast Python package manager. To install TieDIE with uv:
# Install uv if you don't have it
curl -LsSf https://astral.sh/uv/install.sh | sh
# Create a new project with TieDIE
uv init my-project
cd my-project
uv add git+https://github.com/saezlab/tiedie.git
# Or add to an existing project
uv add git+https://github.com/saezlab/tiedie.git
Development installation
git clone https://github.com/saezlab/tiedie.git
cd tiedie
# Using uv (recommended)
uv venv
source .venv/bin/activate
uv pip install -e ".[tests]"
# Or using pip
python -m venv .venv
source .venv/bin/activate
pip install -e ".[tests]"
Usage
Command Line Interface
For full options:
Python API
from tiedie import ScipyKernel
from tiedie.util import parse_heats, parse_net, normalize_heats
# Parse input files
network = parse_net('pathway.sif')
upstream_heats, upstream_signs = parse_heats('upstream.input')
downstream_heats, downstream_signs = parse_heats('downstream.input')
# Normalize heats
upstream_norm = normalize_heats(upstream_heats)
downstream_norm = normalize_heats(downstream_heats)
# Create diffusion kernel and run diffusion
diffuser = ScipyKernel('pathway.sif')
up_diffused = diffuser.diffuse(upstream_norm, reverse=False)
down_diffused = diffuser.diffuse(downstream_norm, reverse=True)
Input File Formats
Network file (.sif)
Tab-separated: source \t interaction \t target
Interaction types:
- -a> : activation
- -t| : inhibition
- -component> : component relationship
Heat file (.input)
Tab-separated: gene \t heat \t sign
Project Structure
tiedie/
├── __init__.py # Public API
├── cli.py # Command-line interface
├── kernel.py # Pre-computed kernel diffusion
├── kernel_scipy.py # On-the-fly kernel generation (scipy)
├── ppr.py # Personalized PageRank diffusion
├── permute.py # Permutation testing
├── util.py # Utility functions
└── ...
tests/ # Test suite
docs/ # Documentation (Tutorial.pdf, FAQ.txt)
Publications
TieDIE was first featured in the 2013 Nature paper "Comprehensive molecular characterization of clear cell renal cell carcinoma" (TCGA Network). In this publication, TieDIE analysis connected frequently mutated genes involving the SWI/SNF chromatin remodelling complex to gene expression changes characteristic of tumor development and progression. The TieDIE network solution is shown in Figure 4: http://www.nature.com/nature/journal/v499/n7456/full/nature12222.html
License
GPL-3.0-or-later
Contact
Feature requests, comments and requests for clarification should be sent to epaull@soe.ucsc.edu.
Development
Running tests
# Install test dependencies
uv pip install -e ".[tests]"
# Run tests
pytest
# Run tests with coverage
pytest --cov=tiedie
Linting and formatting
The project uses ruff for linting and formatting:
Pre-commit hooks
Pre-commit hooks are configured for automated checks:
# Install pre-commit hooks
uv pip install -e ".[dev]"
pre-commit install
# Run hooks manually on all files
pre-commit run --all-files
Building documentation
Documentation is built with MkDocs:
# Install docs dependencies
uv pip install -e ".[docs]"
# Serve docs locally
mkdocs serve
# Build static docs
mkdocs build
Version 2.0 Notes
This version represents a major modernization of the TieDIE codebase:
- Python 3 compatibility: Ported from Python 2.7 to Python 3.9+
- Modern packaging: Converted to a proper Python package with
pyproject.tomland hatchling build backend - Test suite: Added pytest-based test suite with integration tests
- uv support: Compatible with modern Python tooling including uv
- CLI improvements: Migrated from optparse to argparse
This modernization was done at Saez Lab by Dénes Türei (turei.denes@gmail.com).