Statistical inference of genetic pathway analysis in high dimensions

Area of Science:

• Genetics
• Statistical Genetics
• Bioinformatics

Background:

• Genetic pathway analysis is crucial for understanding trait associations.
• High-density genotyping and imputation increase genetic variants, often exceeding sample size, posing analytical challenges.

Purpose of the Study:

• To develop a statistical approach for genetic pathway analysis that accommodates high-dimensional genetic data (variants > sample size).
• To enable detection of pathways with nonsparse weak signals and sparse strong signals.

Main Methods:

• Introduced a novel statistical method designed for high-dimensional genetic pathway analysis.
• Established the asymptotic distribution and conducted power analysis for the proposed statistic.
• Validated the approach through simulation studies and application to a genome-wide association study.

Main Results:

• The proposed method effectively handles pathways where the number of genetic variants exceeds the sample size.
• Demonstrated correct Type I error control and superior statistical power compared to existing methods in simulations.
• Successfully applied the approach to identify relevant pathways in a high-density lipoprotein genome-wide association study.

Conclusions:

• The developed method offers a robust solution for genetic pathway analysis in high-dimensional settings.
• This approach enhances the ability to discover genetic pathways influencing complex traits.
• The findings have implications for genetic association studies and understanding disease mechanisms.