Abstract
Pathway analyses of genome-wide association studies aggregate information over sets of related genes, such as genes in common pathways, to identify gene sets that are enriched for variants associated with disease. We develop a model-based approach to pathway analysis, and apply this approach to data from the Wellcome Trust Case Control Consortium (WTCCC) studies. Our method offers several benefits over existing approaches. First, our method not only interrogates pathways for enrichment of disease associations, but also estimates the level of enrichment, which yields a coherent way to promote variants in enriched pathways, enhancing discovery of genes underlying disease. Second, our approach allows for multiple enriched pathways, a feature that leads to novel findings in two diseases where the major histocompatibility complex (MHC) is a major determinant of disease susceptibility. Third, by modeling disease as the combined effect of multiple markers, our method automatically accounts for linkage disequilibrium among variants. Interrogation of pathways from eight pathway databases yields strong support for enriched pathways, indicating links between Crohn's disease (CD) and cytokine-driven networks that modulate immune responses; between rheumatoid arthritis (RA) and "Measles" pathway genes involved in immune responses triggered by measles infection; and between type 1 diabetes (T1D) and IL2-mediated signaling genes. Prioritizing variants in these enriched pathways yields many additional putative disease associations compared to analyses without enrichment. For CD and RA, 7 of 8 additional non-MHC associations are corroborated by other studies, providing validation for our approach. For T1D, prioritization of IL-2 signaling genes yields strong evidence for 7 additional non-MHC candidate disease loci, as well as suggestive evidence for several more. Of the 7 strongest associations, 4 are validated by other studies, and 3 (near IL-2 signaling genes RAF1, MAPK14, and FYN) constitute novel putative T1D loci for further study.
Introduction
Standard Genome-Wide Association Studies (GWAS) typically test hundreds of thousands of genetic variants (SNPs) individually. While successful, this "one-SNP-at-a-time" method often fails to detect variants with small individual effects that do not reach the stringent threshold for genome-wide significance. Existing pathway analysis methods attempt to aggregate these SNPs into functional groups, but they often struggle with Linkage Disequilibrium (LD), the masking effect of the Major Histocompatibility Complex (MHC), and identifying specific driver genes within a pathway.
Methods
The authors propose a Bayesian variable selection framework that treats disease risk as the combined effect of multiple markers. This approach incorporates an "enrichment parameter" ($\theta$) directly into a multivariate regression model, quantifying the increase in probability that a SNP in a specific pathway is associated with the disease.
By analyzing all SNPs simultaneously, the model naturally accounts for LD. To handle the dominant MHC signal in autoimmune diseases, a two-stage procedure was developed to first account for the MHC effect and then search for additional pathway enrichment elsewhere in the genome.
Key Results
The analysis of data from the Wellcome Trust Case Control Consortium (WTCCC) across seven common diseases identified several highly significant pathways. The most striking finding was a massive enrichment of Type 1 Diabetes (T1D) risk factors within the IL-2 signaling pathway, showing a Bayes factor of $1.2 \times 10^{12}$.
Prioritizing variants in these enriched pathways led to the identification of several novel loci:
- Type 1 Diabetes: Identified 7 additional non-MHC candidate loci, including three novel putative loci near genes RAF1, MAPK14, and FYN.
- Crohn’s Disease: Confirmed strong links to cytokine-driven networks, specifically IL-23 and IL-12 signaling.
- Rheumatoid Arthritis: Highlighted the "Measles" pathway, suggesting viral triggers may play a role in susceptibility.
Crucially, for Crohn's Disease and Rheumatoid Arthritis, 7 of the 8 additional associations found through this prioritization were validated by independent meta-analyses, confirming the method's ability to uncover real biological signals missed by standard GWAS.
Discussion and Future Directions
While powerful, the model is sensitive to the "prior" effect sizes chosen for variants. Additionally, the results are dependent on the quality and completeness of the pathway databases used. As these databases become more refined and GWAS datasets grow, this integrated approach will become increasingly valuable for fine-mapping the genetic causes of complex diseases.
Summary updated