NetCore

NetCore applies node coreness–aware random walk with restart network propagation to prioritize genes and identify phenotype-associated modules within protein–protein interaction (PPI) networks.

Key Features:

• Node Coreness Integration: NetCore incorporates node coreness to mitigate node degree bias in PPI networks.
• Coreness-weighted Random Walk with Restart: It incorporates node coreness into the random walk with restart propagation procedure.
• Post-propagation Gene Re-ranking: NetCore re-ranks genes after propagation to reflect coreness-adjusted significance.
• Semi-supervised Module Identification: The method uses a semi-supervised approach anchored on known phenotype-associated genes to identify network modules.
• Integration of Known and Novel Genes: NetCore integrates established and newly identified candidate genes into coherent modules.
• Evaluation on GWAS Traits: The approach was evaluated across gene sets from 11 GWAS traits using cross-validation and compared against degree-based propagation methods.
• Application to Disease Datasets: NetCore has been applied to Schizophrenia GWAS data and pan-cancer mutation datasets to identify disease-related genes and modules.
• High-confidence PPI Source: The method uses a high-confidence PPI network extracted from ConsensusPathDB.

Scientific Applications:

• Phenotype–genotype association studies: Prioritizes genes in PPI networks to support phenotype-genotype association analyses.
• Module identification in PPI networks: Identifies phenotype-associated network modules anchored on known genes.
• Disease gene and module discovery: Applied to Schizophrenia GWAS and pan-cancer mutation datasets to discover disease-related genes and modules.
• Benchmarking network propagation: Enables comparative evaluation of propagation methods across 11 GWAS-derived gene sets using cross-validation.

Methodology:

Random walk with restart incorporating node coreness; post-propagation gene re-ranking; semi-supervised module identification anchored at known phenotype-associated genes; cross-validation evaluation on GWAS-derived gene sets; use of a high-confidence PPI network from ConsensusPathDB.