🧪 Explore interactively: https://cancerhubs.app
📄 Published article: Briefings in Bioinformatics (2025)
📦 Shiny App Source: cancerhubs_shiny
📘 Reproducibility Package: cancerhubs_paper
CancerHubs is a computational method that identifies cancer-relevant protein hubs by integrating:
- Somatic mutation data
- Prognostic gene expression scores (PRECOG)
- Protein–protein interaction networks (BioGRID)
By ranking genes based on the number of mutated interactors, it prioritises central players in cancer-related pathways—beyond mutation frequency alone. This network-centric perspective allows the discovery of potential driver genes otherwise overlooked by traditional methods.
Each gene is assigned a Network Score defined as:
Network Score = (# Mutated Interactors)^2 / (# Total Interactors)
This formula reflects how strongly a gene is embedded in a network of cancer-mutated interactors. Higher scores indicate greater centrality and functional relevance in tumour contexts.
Note: The Network Score is computed exclusively from somatic mutation data. It does not take into account copy-number variations or structural alterations.
-
1. Data Retrieval
Mutation data are collected for each tumour type. Genes are filtered for coding/non-coding status. -
2. Clinical Outcome Correlation
Gene expression–survival associations are obtained from PRECOG and integrated as meta-Z scores. -
3. Gene Classification
Genes are labelled as:- MUT – mutated only
- PRECOG – prognosis-associated only
- MUT + PRECOG – both
-
4. Network Construction
BioGRID interactions define the gene interactome. -
5. Network Scoring
Genes are scored and ranked by the fraction and number of mutated interactors.
This repository contains the core pipeline used to process mutation data, integrate clinical scores, and calculate network-based gene rankings.
-
/scripts
Core R scripts for mutation integration and scoring. -
/data
Input datasets (formatted mutation and interactome data). -
/result/all_results.rds
Main output object containing ranked genes, used by the Shiny app. -
/docs
Supplementary data and mutation annotation summaries.
An interactive Shiny web application is now available:
👉 https://cancerhubs.app
Use it to:
- Search gene rankings across cancers
- Visualise pan-cancer hub scores
- Browse interactors and 3D networks
- Download gene-specific data and networks
App source code:
🔗 https://github.com/ingmbioinfo/cancerhubs_shiny
To run the full pipeline locally:
- Install required packages (see headers in
functions.r) - Clone this repository and set the working directory
- Execute the pipeline:
source("workflow.r")
For full reproducibility with frozen data versions, refer to the companion repository:
📘 cancerhubs_paper
Mutational Data: Derived from 14 published cancer genomic studies encompassing both adult and paediatric datasets, including WES, WGS, or mixed approaches. Tumour types include Multiple Myeloma, Breast Cancer, Prostate Cancer, Pancreatic Cancer, Colorectal Cancer, Ovarian Cancer, Gastric Cancer, Small Cell Lung Cancer, B-ALL, Burkitt Lymphoma, and Melanoma.
Clinical Outcome Predictions: PRECOG meta-Z scores were used to assess correlation between gene expression and patient survival.
Protein Interactions: BioGRID protein–protein interaction data provided the reference interactome, restricted to physical interactions involving the selected gene sets.
All processed datasets and scripts are included in this GitHub repository. The versions and formats used are documented below.
all_results.rds: Gene-wise rankings for each tumour type, including Network Scores and annotation categories.genes_interactors_list.rds: Curated interactome data mapping each gene to its interaction partners.formatted_datasets: Input gene tables formatted for pipeline ingestion.biogrid_interactors: Full PPI interaction records derived from BioGRID.Mutational Data Summary: PDF with curation criteria and source references for mutation datasets.
This research was funded by Associazione Italiana per la Ricerca sul Cancro (AIRC), under MFAG 2021 ID 26178 project to Nicola Manfrini.
For questions or support, please contact:
📧 manfrini@ingm.org
📧 ferrari@ingm.org
If you use CancerHubs in your research, please cite:
Ivan Ferrari, Federica De Grossi, Giancarlo Lai, Stefania Oliveto, Giorgia Deroma, Stefano Biffo, Nicola Manfrini.
CancerHubs: a systematic data mining and elaboration approach for identifying novel cancer-related protein interaction hubs.
Briefings in Bioinformatics, Volume 26, Issue 1, January 2025.
https://doi.org/10.1093/bib/bbae635