rcNet

rcNet associates gene sets with disease phenotypes by integrating disease phenotype similarity networks and gene-gene interaction networks to improve prediction of phenotype–gene set relationships.

Key Features:

• Network-Based Approach: Utilizes topological information from disease phenotype similarity networks and gene-gene interaction networks to enhance prediction accuracy of associations between gene sets and disease phenotypes.
• Learning Framework: Employs a learning framework that maximizes coherence between predicted phenotype–gene set relations and known disease phenotype–gene associations, enabling robust predictions for poorly annotated genes.
• Algorithmic Design: Combines ridge regression with label propagation and provides two algorithmic variants to optimize the objective functions within the learning framework.
• Validation and Performance: Evaluated using leave-one-out cross-validation on Online Mendelian Inheritance in Man (OMIM) data and an independent test set of recently discovered disease–gene associations, achieving superior rankings compared to baseline methods.
• Application in Genomic Studies: Applied to Genome-Wide Association Study (GWAS) results, DNA copy number variation analyses, and gene expression profiling to rank target diseases for candidate genes in case studies.

Scientific Applications:

• Disease Gene Identification: Predicts and prioritizes associations between candidate genes and specific disease phenotypes to support validation of disease genes from genomic studies.
• Enhanced Annotation: Integrates network-derived relationships to mitigate incomplete gene annotations and provide more comprehensive disease–gene relationship insights.
• Cross-Disciplinary Utility: Applicable across diverse genomic data types including GWAS, DNA copy number variation, and gene expression profiling.

Methodology:

Uses topological information from disease phenotype similarity networks and gene-gene interaction networks; applies a learning framework that maximizes coherence between predicted phenotype–gene set relations and known associations; optimizes objective functions via ridge regression combined with label propagation (with two variants); and evaluates performance with leave-one-out cross-validation on Online Mendelian Inheritance in Man (OMIM) and an independent test set of recently discovered disease–gene associations.