Kernel-machine testing coupled with a rank-truncation method for genetic pathway analysis.

Area of Science:

• Genetics and Bioinformatics
• Complex Disease Research
• Statistical Genomics

Background:

• Traditional genome-wide association studies (GWASs) often rely on single-marker analysis, limiting insights into complex disease architecture.
• Pathway analysis offers a more comprehensive view by considering gene marker hierarchical structures within biological pathways.
• Existing pathway analysis methods have limitations in integrating common and rare variants effectively.

Purpose of the Study:

• To develop a novel pathway analysis approach for complex diseases.
• To effectively integrate gene-level and pathway-level analyses.
• To account for both common and rare genetic variants across the entire allelic frequency spectrum.

Main Methods:

• Utilizes the sequence kernel association test (SKAT) for gene-level effects.
• Employs an extended adaptive rank truncated product statistic for pathway-level effects.
• Introduces a new weighting scheme for genetic variants, unifying common and rare variant analysis without frequency cutoffs.

Main Results:

• The proposed method demonstrates high statistical power across various simulated scenarios.
• Maintains correct type I error rates, ensuring reliable findings.
• Successfully applied to analyze associations between bipolar disorder and candidate pathways using WTCCC data.

Conclusions:

• The novel pathway analysis approach provides a powerful and flexible tool for complex disease genetics.
• Effectively integrates diverse genetic variant frequencies, enhancing discovery potential.
• Applicable to various data types including GWAS, exome-sequencing, and deep resequencing data.