Rare-variant association studies: When are aggregation tests more powerful than single-variant tests?

Area of Science:

• Genetics
• Statistical Genetics
• Bioinformatics

Background:

• Single-variant tests are less effective for rare genetic variants compared to common ones.
• Aggregation tests pool rare variants within genomic regions to improve association detection.
• Large biobank studies increasingly utilize aggregation tests, yielding significant findings.

Purpose of the Study:

• To determine the genetic models where aggregation tests outperform single-variant tests for rare variant association studies.
• To investigate the factors influencing the power of aggregation tests.

Main Methods:

• Analytic calculations were performed assuming an additive genetic model for a normally distributed trait.
• Power was assessed based on the number of causal variants (c), total rare variants (v), region heritability (h²), and sample size (n).
• Simulations utilized data from 378,215 UK Biobank participants.

Main Results:

• Aggregation tests are more powerful than single-variant tests only when a substantial proportion of variants are causal.
• Statistical power is highly dependent on the underlying genetic model and the specific set of aggregated rare variants.
• Aggregation tests showed superiority in over 55% of genes under specific conditions (e.g., aggregating protein-truncating variants and deleterious missense variants with high causal probabilities).

Conclusions:

• The choice between single-variant and aggregation tests depends critically on the proportion of causal variants and the genetic architecture of the trait.
• Understanding these dependencies is crucial for optimizing rare variant association studies in large, growing datasets.