HiGwas: how to compute longitudinal GWAS data in population designs

Area of Science:

• Genetics
• Statistical Genetics
• Bioinformatics

Background:

• Genome-wide association studies (GWAS) face the "big p, small n" challenge (many SNPs, fewer subjects).
• Integrating longitudinal data in GWAS improves power but introduces autocorrelation challenges.
• Existing methods struggle with high-dimensional genetic data and temporal correlations.

Purpose of the Study:

• To develop and provide a computational tool for analyzing longitudinal GWAS datasets.
• To address the "big p, small n" problem and autocorrelation in genetic association studies.
• To enhance the understanding of genetic architecture for complex traits and diseases using longitudinal data.

Main Methods:

• Developed statistical models incorporating dimension reduction and longitudinal data analysis.
• Created the R package HiGwas for computational accessibility to applied geneticists.
• Implemented functions for single SNP analysis, significance adjustment, preconditioning, and model selection.

Main Results:

• The HiGwas package offers a practical solution for analyzing high-dimensional longitudinal GWAS data.
• The software provides estimates of genetic parameters and their confidence intervals.
• Demonstrated the utility of HiGwas through real data analysis and package vignettes.

Conclusions:

• HiGwas effectively handles the complexities of longitudinal GWAS, enabling more powerful genetic analyses.
• The package facilitates the study of genetic influences on complex traits over time.
• HiGwas makes advanced statistical methods accessible for broader application in genetic research.