DisSNPNet: Predicting disease-associated single-nucleotide polymorphisms using linkage disequilibrium, disease...

Area of Science:

• Genetics and Bioinformatics
• Computational Biology
• Systems Biology

Background:

• Identifying disease-associated single-nucleotide polymorphisms (SNPs) is crucial for understanding complex genetic diseases.
• Genome-wide association studies (GWAS) are effective but costly and data-intensive.
• Network-based methods offer a complementary strategy for prioritizing candidate variants by leveraging linkage disequilibrium (LD) and disease relationships.

Purpose of the Study:

• To introduce DisSNPNet, a heterogeneous network-based framework for prioritizing disease-associated SNPs.
• To integrate chromosome-specific SNP LD networks, disease similarity networks, and known disease-SNP associations.
• To evaluate the performance and biological relevance of DisSNPNet compared to existing methods.

Main Methods:

• Constructed DisSNPNet by integrating SNP LD networks (1000 Genomes Project data), MeSH-based disease similarity networks, and CAUSALdb disease-SNP associations.
• Applied random walk with restart algorithm to rank SNPs for each disease.
• Evaluated predictive performance using 3-fold cross-validation (AUROC, AUPR) and assessed biological plausibility via GWAS resources and KEGG pathway enrichment.

Main Results:

• DisSNPNet consistently outperformed SNP-only LD networks, with heterogeneous networks achieving higher AUROC and AUPR.
• Strong LD networks (r² ≥ 0.8) enhanced precision, especially in imbalanced datasets.
• Top-ranked SNPs demonstrated significantly greater GWAS evidence than random expectation, indicating nonrandom enrichment and revealing coherent biological mechanisms across various disease types.

Conclusions:

• DisSNPNet provides a robust and interpretable framework for prioritizing disease-associated SNPs.
• The method offers a scalable, evidence-supported approach for SNP prioritization and hypothesis generation.
• DisSNPNet complements experimental and population-based studies, enhancing the understanding of complex disease genetics.