HotNet

HotNet identifies significantly altered subnetworks within gene interaction networks to detect groups of functionally related somatic mutations across cancer samples.

Key Features:

• Genome-scale gene interaction network analysis: Operates on genome-scale gene interaction networks to contextualize mutated genes within signaling and regulatory pathways.
• Diffusion-based influence mapping: Uses a diffusion process on the interaction network to define a local neighborhood of "influence" for each mutated gene.
• Statistical identification of mutated subnetworks: Performs de novo identification of subnetworks that are statistically significantly altered across patient samples via a two-stage multiple hypothesis testing framework that controls the false discovery rate (FDR).
• Computational efficiency: Implements a computationally efficient strategy for de novo subnetwork discovery.
• Application to cancer genomics datasets: Was applied to large human protein-protein interaction networks with somatic mutation data from glioblastoma and lung adenocarcinoma, recovering known pathways and identifying additional pathways previously associated with other cancers.
• Scalability: Scales to large cancer genome studies to assist interpretation of complex mutational landscapes.

Scientific Applications:

• Pathway-focused subnetwork analysis: Provides an alternative to traditional pathway enrichment by identifying altered subnetworks rather than individual pathways.
• Discovery of novel cancer pathways: Identifies additional pathways altered in cancer samples that can suggest novel therapeutic targets.
• Large-scale cancer genomics interpretation: Supports interpretation of mutational landscapes in large-scale cancer genomics studies.

Methodology:

Applies a diffusion process on genome-scale (human) protein-protein interaction networks and detects altered subnetworks using a two-stage multiple hypothesis testing framework that controls the FDR, with inputs including somatic mutation data from glioblastoma and lung adenocarcinoma.