SCEBE: an efficient and scalable algorithm for genome-wide association studies on longitudinal outcomes with...

Area of Science:

• Genetics and Genomics
• Statistical Bioinformatics
• Neurodegenerative Diseases

Background:

• Genome-wide association studies (GWAS) using longitudinal data enhance statistical power for genetic discovery.
• Computational complexity of longitudinal data modeling poses a significant challenge in large-scale GWAS.
• Existing empirical Bayesian estimate (EBE) methods for longitudinal GWAS exhibit bias in association testing and effect size estimation.

Purpose of the Study:

• To develop a computationally efficient and statistically unbiased method for large-scale GWAS with longitudinal phenotypes.
• To correct for biases inherent in current EBE-based approximation methods for longitudinal GWAS.
• To provide accurate P-values and effect size estimates in longitudinal genetic association studies.

Main Methods:

• Proposed a novel method, simultaneous correction for EBE (SCEBE), to rectify bias in the naive EBE approach.
• Implemented SCEBE for unbiased P-value calculation and effect size estimation in longitudinal GWAS.
• Applied SCEBE to analyze Alzheimer's Disease Neuroimaging Initiative data, including over 6.4 million single nucleotide polymorphisms.

Main Results:

• SCEBE demonstrated significant computational efficiency, achieving nearly 10,000-fold improvement over existing methods.
• Computation time for large-scale longitudinal GWAS was reduced from months to minutes using SCEBE.
• The method provides unbiased estimates and P-values, crucial for reliable genetic association findings.

Conclusions:

• SCEBE offers a fast, unbiased, and efficient solution for conducting large-scale GWAS with longitudinal outcomes.
• The developed method accelerates genetic discovery in complex diseases by overcoming computational bottlenecks.
• The SCEBE package is publicly available, facilitating its adoption in genetic research.