A hidden two-locus disease association pattern in genome-wide association studies

Area of Science:

• Genetics
• Bioinformatics
• Computational Biology

Background:

• Genome-wide association studies (GWAS) typically use single-locus or two-locus tests.
• These methods may miss association patterns involving groups of correlated single-nucleotide polymorphisms (SNPs).
• SNP correlation can cause 'unfaithfulness,' where marginal effects don't reflect joint effects.

Purpose of the Study:

• To develop a computational method for detecting SNP association patterns masked by unfaithfulness.
• To identify novel genetic associations in GWAS data that are missed by existing methods.

Main Methods:

• Developed a novel computational approach to identify complex SNP association patterns.
• Applied the method to analyze seven datasets from the Wellcome Trust Case Control Consortium (WTCCC).
• Examined all pairs of SNPs within each dataset, with analysis per dataset taking approximately one week.

Main Results:

• Demonstrated that the newly identified association pattern, masked by unfaithfulness, is prevalent in GWAS data.
• Successfully applied the computational method to real GWAS datasets.
• The analysis of WTCCC data confirmed the widespread existence of these hidden associations.

Conclusions:

• The developed method enriches GWAS discoveries by identifying previously missed associations.
• Findings may inform tagSNP analysis and future GWAS experimental design.
• Further investigation is needed to explore the biological implications of these newly identified associations, with replication challenges due to limited independent datasets.