Gene-Gene Interactions Detection Using a Two-stage Model.

Area of Science:

• Genetics
• Bioinformatics
• Statistical genomics

Background:

• Genome-wide association studies (GWAS) identify genetic loci for traits using single nucleotide polymorphisms (SNPs).
• Complex traits are likely influenced by interactions among multiple SNPs.
• Current brute-force methods for detecting SNP interactions are computationally intensive due to large datasets.

Purpose of the Study:

• To propose an efficient two-stage method, Threshold-based Efficient Pairwise Association Approach (TEPAA), for detecting SNP interactions in GWAS.
• To reduce computational burden while maintaining statistical power compared to brute-force approaches.

Main Methods:

• Implemented a two-stage approach: initial single marker tests followed by pairwise association tests on selected SNPs.
• Derived the joint distribution between single SNP and pairwise SNP association statistics.
• Applied the TEPAA method to the Northern Finland Birth Cohort data.

Main Results:

• Achieved a 63-fold speedup in analysis.
• Maintained 99% of the statistical power of the brute-force method.
• Demonstrated the computational feasibility and high power of TEPAA.

Conclusions:

• TEPAA offers a computationally efficient and powerful alternative for detecting SNP interactions in GWAS.
• The method effectively reduces the number of necessary tests while preserving statistical power.
• This approach can accelerate genetic research into complex traits.