RISK Network Tutorial offers a detailed guide on using RISK (Regional Inference of Significant Kinships), a next-generation tool for biological network annotation and visualization. This guide details key functionalities and practical applications for network analysis using RISK.
- Full Documentation: riskportal.github.io/network-tutorial
- Try in Browser (Binder):
- Source Code: github.com/riskportal/network
Launch RISK tutorial instantly in Binder to interactively explore the workflow and analyses without any installation required.
- Clustering Algorithms: Implements Louvain, Leiden, Markov Clustering, Greedy Modularity, Label Propagation, Spinglass, and Walktrap to identify structured network regions.
- Statistical Analysis: Includes hypergeometric, permutation, binomial, chi-squared, Poisson, and z-score tests to assess annotation overrepresentation and significance.
- Supports multiple file formats: Compatible with JSON, CSV, Cytoscape, and GPickle formats.
- Publication-Ready Visualizations: Customizable figures with kernel density estimate overlays, node/edge adjustments, and high-resolution exports in multiple formats, including SVG, PNG, and PDF.
In this demonstration, we use the yeast PPI network published by Michaelis et al. (2023), which contains 3,927 proteins and 31,004 interactions. We filtered the dataset to retain proteins with six or more interactions, emphasizing core functional interactions. The final dataset contains 2,059 nodes and 27,690 edges. RISK is applied to perform analyses such as network clustering and statistical overrepresentation of annotations.
Below is a visual representation of the yeast PPI interaction network, annotated with Gene Ontology Biological Processes (GO BP) terms:
RISK also identifies party hubs and date hubs, two distinct types of central nodes in the network. Party hubs (orange) are highly connected proteins that interact with many partners simultaneously, often forming stable protein complexes. In contrast, date hubs (cyan) connect with different partners at different times or conditions, playing roles in dynamic regulatory interactions. This visualization highlights party and date hubs in the yeast network:
The GO BP contour overlay highlights the spatial organization of biological processes, providing additional context for interpreting these hubs. For a full demonstration of this analysis, please refer to the tutorial notebook, which can be downloaded and run locally.
Step 1: Install Python 3.8+
Download and install Python 3.8 or later (recommended) from the official Python website. RISK is tested on Python 3.8–3.12.
- On Windows, ensure you check the box that says Add Python to PATH during installation. If you missed this step, follow this guide to manually set the PATH.
Step 2: Create a Virtual Environment (Optional but Recommended)
Set up a virtual environment to manage dependencies for RISK:
-
For Windows:
python -m venv risk-env risk-env\Scripts\activate
-
For macOS/Linux:
python3 -m venv risk-env source risk-env/bin/activate
Step 3: Configure Jupyter to Use the Virtual Environment
Follow this guide to link your virtual environment with Jupyter.
After setup, select the virtual environment in Jupyter Notebook:
- Click Kernel in the menu.
- Choose Change kernel.
- Select your virtual environment (e.g.,
risk-env).
Step 4: Install Jupyter
With your virtual environment activated, install Jupyter Notebook:
pip install jupyterStep 5: Clone Repository
To install the dependencies for this tutorial, run:
git clone https://github.com/riskportal/network-tutorial.git
cd network-tutorialStep 6: Install Dependencies
To install the dependencies for this tutorial, run:
pip install -r requirements.txtStep 7: Launch Jupyter Notebook
Start Jupyter Notebook to run the RISK tutorial:
jupyter notebookIf you use RISK in your research, please reference the following:
Horecka et al., "RISK: a next-generation tool for biological network annotation and visualization", 2025.
DOI: 10.5281/zenodo.xxxxxxx
This tutorial follows the GPLv3 License.