Reproducible processing of TCGA regulatory networks

Area of Science:

• Computational Biology
• Bioinformatics
• Cancer Genomics

Background:

• Advances in sequencing and computation enable deep molecular disease analysis.
• Biological networks are crucial for omics data analysis and modeling gene/protein interactions.
• The Cancer Genome Atlas (TCGA) provides extensive data but lacks a unified workflow for end-to-end regulatory network analysis.

Purpose of the Study:

• To develop a coherent and reusable workflow for reproducible inference and analysis of cancer regulatory networks.
• To integrate existing computational tools into a robust pipeline for TCGA data.

Main Methods:

• Developed tcga-data-nf, a Nextflow workflow for reproducible regulatory network inference from TCGA data.
• The workflow integrates multiomic data (RNA sequencing, methylation) downloading, preprocessing, and analysis using Network Zoo.
• Powered by the NetworkDataCompanion R package for TCGA data management.

Main Results:

• tcga-data-nf enables reproducible regulatory network inference from thousands of TCGA samples via a single command.
• Demonstrated pipeline utility in investigating epigenetic differences in colon cancer subtypes.
• Generated a public database of pre-computed networks for 10 common cancer types.

Conclusions:

• tcga-data-nf offers a complete, flexible, and extensible framework for cancer regulatory network analysis.
• Addresses the need for reproducible workflows in analyzing TCGA data.
• Bridges a gap in current software tools for TCGA data analysis.