proteINET

proteINET is a small toolkit to fetch, analyze and visualize protein–protein interaction (PPI) networks using the STRING database. It builds PPI networks from a list of query proteins, computes network metrics, runs enrichment (GSEA), exports Cytoscape-ready files and generates an interactive HTML report.

Features

• Query STRING-DB (functional or physical networks)
• Compute global, node and edge network metrics
• Export nodes/edges/graph (CSV, GraphML) for Cytoscape
• Run GSEA (via gseapy) on ranked gene lists
• Generate an interactive HTML report (Cytoscape.js + plots)

Requirements

• Python 3.11 (recommended)
• Conda (recommended) or pip
• Network access to the STRING API

A curated environment is provided in environment.yml (includes networkx, pandas, matplotlib, gseapy, requests, pyvis, etc.).

Installation (recommended)

1. Clone the repository:

git clone https://github.com/GiuseppeBocci/proteINET.git
cd proteINET

2. Create and activate the conda environment:

conda env create -f environment.yml
conda activate proteINET

If not using conda, install the pip packages listed in environment.yml.

Quick example

Run the example pipeline shipped with the repo: python main.py The example pipeline queries STRING for proteins (TP53, BRCA1), builds the network, computes metrics, runs GSEA, exports Cytoscape files to cytoscape_export/ and generates protein_network_report.html.

API usage snippets

Fetch and standardize STRING network:

from graph.StringLoader import StringLoader

oader = StringLoader(protein_query=["TP53","BRCA1"], species=9606, add_nodes=20) 
data = loader.retrieve_data().standardize_data_format().get_data()

Build network and compute metrics:

from graph.PPINetwork import PPINetwork

ppin = PPINetwork(data=data) ppin.compute_all_global_metrics(weighted=False) ppin.compute_all_node_metrics(weighted=False) ppin.compute_all_edge_metrics(weighted=False)

Export for Cytoscape:

exported = ppin.export_for_cytoscape(output_dir="cytoscape_export", basename="ppi_network_TP53-BRCA1")

Generate interactive HTML report:

from visual.ReportGenerator import ReportGenerator

report_generator = ReportGenerator()
report_generator.add_cytoscape_html_report(ppin, title="STRING Protein Network").generate_report_file("protein_network_report.html")

Output files

• protein_network_report.html — interactive HTML report
• cytoscape_export/ — nodes CSV, edges CSV, GraphML, metrics JSON
• gsea_results/ — prerank files and enrichment reports

Configuration & tuning

StringLoader parameters to tune:

• required_score (int): minimum confidence score (default 400)
• add_nodes (int): number of additional nodes to include
• timeout (int): HTTP request timeout (seconds)

For large networks increase timeouts and consider filtering by required_score.

Troubleshooting

• Ensure network access and that STRING API is reachable.
• If mapping fails, confirm input protein identifiers are correct.
• Respect STRING usage policies and rate limits.
• If HTML report fails to load Cytoscape locally, it falls back to the CDN.

Project structure (key files)

• main.py - example pipeline that ties components together
• graph/ - core PPI classes (Loader, StringLoader, PPINetwork)
• visual/ - ReportGenerator and plotting helpers
• environment.yml - conda environment and pip deps
• cytoscape_export/, gsea_results/ - example outputs (created when running pipeline)

License

See LICENSE.md (Copyright (C) 2026 Giuseppe Bocci)