A weighted U-statistic for genetic association analyses of sequencing data

Area of Science:

• Genetics
• Bioinformatics
• Statistical Genetics

Background:

• Next-generation sequencing generates vast data, enabling rare variant analysis in complex diseases.
• High-dimensional sequencing data presents significant statistical challenges.
• Traditional methods lose power with low-frequency variants and high dimensionality.

Purpose of the Study:

• To develop a novel statistical test for high-dimensional sequencing data association analysis.
• To address the power limitations of existing methods in complex disease genetics.
• To introduce a flexible test applicable to various phenotypes and disease models.

Main Methods:

• Developed the Weighted U Sequencing test (WU-SEQ), a nonparametric U-statistic-based method.
• WU-SEQ makes no assumptions on disease models or phenotype distributions.
• Evaluated performance through simulations and an empirical study, comparing with SKAT.

Main Results:

• WU-SEQ demonstrated superior performance over SKAT when assumptions were violated (e.g., heavy-tailed distributions).
• WU-SEQ achieved comparable performance to SKAT even when assumptions were met.
• Applied WU-SEQ to Dallas Heart Study data, identifying an ANGPTL4 association with very low-density lipoprotein cholesterol.

Conclusions:

• WU-SEQ is a powerful and flexible tool for rare variant association analysis in high-dimensional sequencing data.
• The method provides a robust alternative to existing approaches, particularly under violated assumptions.
• WU-SEQ successfully identified a novel genetic association relevant to lipid metabolism.