gsSKAT: Rapid gene set analysis and multiple testing correction for rare-variant association studies using weighted...

Area of Science:

• Genetics
• Statistical Genetics
• Bioinformatics

Background:

• Next-generation sequencing (NGS) enables detailed characterization of rare genetic variants.
• Single-variant tests lack power for rare variants; aggregative methods are used within genes.
• Rare variants may collectively impact biological pathways, necessitating pathway-level analysis.

Purpose of the Study:

• To develop a computationally efficient statistical strategy for rare-variant association testing across multiple genes within biological pathways.
• To address challenges in multiple testing correction for large gene sets with overlapping definitions.
• To provide a method applicable to existing rare-variant analysis results.

Main Methods:

• Developed a statistical strategy using gene-level linear kernel score test summary statistics for aggregative rare-variant analysis.
• Derived estimators for the effective number of tests to control family-wise error rate.
• Conducted extensive simulation studies to compare the proposed method with existing kernel and adaptive methods.
• Applied the method to case-control studies for hereditary prostate cancer and schizophrenia.

Main Results:

• The proposed method demonstrates robust performance in simulations, offering an alternative to computationally intensive direct methods.
• The approach effectively handles correlated test statistics arising from overlapping gene sets, improving Type I error rate control.
• Successful application to real-world datasets for hereditary prostate cancer and schizophrenia.

Conclusions:

• The developed statistical strategy provides an efficient and accurate approach for pathway-level rare-variant association testing.
• The method facilitates better control of Type I error rates in large-scale genetic analyses.
• Open-source R code is provided to enable widespread adoption and application of the method in genetic research.