Integrative multi-omics module network inference with Lemon-Tree.

Area of Science:

• Bioinformatics
• Computational Biology
• Genomics

Background:

• Module network inference reconstructs gene co-expression and regulatory networks from multi-omics data.
• Existing methods require robust and scalable software for accurate analysis.

Purpose of the Study:

• To develop and evaluate Lemon-Tree, an open-source software for module network inference.
• To assess Lemon-Tree's performance against state-of-the-art methods using large-scale tumor datasets.

Main Methods:

• Lemon-Tree implements ensemble methods for module network inference.
• Benchmarking involved large-scale tumor datasets and analysis of glioblastoma data from The Cancer Genome Atlas (TCGA).

Main Results:

• Lemon-Tree algorithms demonstrated favorable performance compared to existing software.
• The tool successfully identified known glioblastoma oncogenes and tumor suppressors.
• Novel candidate driver genes were predicted and validated through pathway and survival analyses.

Conclusions:

• Lemon-Tree is a powerful and accurate tool for module network inference.
• It aids in identifying master regulators and novel driver genes in cancer, particularly glioblastoma.
• The software is open-source, modular, and extensible for broader applications.