A powerful and adaptive association test for rare variants.

Area of Science:

• Statistical Genetics
• Genomic Association Studies
• Bioinformatics

Background:

• Existing genetic association tests struggle with rare variants (RVs) when many non-associated RVs are present, leading to reduced statistical power.
• The performance of burden tests can degrade as the proportion of non-associated RVs increases within a tested group.

Purpose of the Study:

• To develop novel statistical tests for global association between traits and sets of rare variants (RVs).
• To address the performance limitations of current methods in the presence of numerous non-associated RVs.
• To propose an adaptive test that maintains high power across diverse scenarios.

Main Methods:

• Proposed a class of sum of powered score (SPU) tests based on score vectors from general regression models.
• SPU tests generalize existing burden tests (sum test) and variance component tests (sum of squared score test).
• Developed an adaptive SPU (aSPU) test to optimize power by approximating the most powerful SPU test for specific scenarios.

Main Results:

• Extensive simulations showed the aSPU test significantly outperforms existing state-of-the-art association tests, especially with many non-associated RVs.
• SPU tests can adjust for covariates, such as principal components for population stratification.
• The aSPU test demonstrated high power and adaptability across various simulated genetic association scenarios.

Conclusions:

• The proposed SPU and aSPU tests offer a versatile and powerful approach for genetic association studies involving rare variants.
• The aSPU test provides a robust solution for detecting associations in the presence of confounding factors like numerous non-associated RVs.
• Application to GAW17 mini-exome data confirmed the practical utility and improved performance of the aSPU test compared to existing methods.