A comparison study of multivariate fixed models and Gene Association with Multiple Traits (GAMuT) for next-generation...

Area of Science:

• Statistical genetics
• Genomic association studies
• Quantitative trait analysis

Background:

• Analyzing multiple correlated quantitative phenotypes is crucial for understanding complex diseases.
• Existing methods vary in their suitability for different genetic architectures.

Purpose of the Study:

• To compare the performance of approximate F-distributed tests of multivariate functional linear models (MFLM) and multivariate analysis of variance (MANOVA) against Gene Association with Multiple Traits (GAMuT).
• To determine the optimal method for different genetic scenarios, including major gene effects and polygenic effects.

Main Methods:

• Extensive simulations were conducted to evaluate statistical power.
• Methods compared include MFLM, MANOVA, and GAMuT for high-dimensional genotype data association testing.
• Real-world data from European cohorts and the Trinity Students Study were analyzed.

Main Results:

• Approximate F-tests (MFLM/MANOVA) demonstrated higher power for major gene association analysis (large variant effects).
• GAMuT showed higher power for polygenic effects (many variants with small effects).
• Both MFLM/MANOVA and GAMuT can analyze rare, common, or combined variants, with GAMuT performing well in large-scale polygenic analyses.

Conclusions:

• The choice of statistical method depends on the underlying genetic architecture.
• MFLM and MANOVA are robust for major gene detection, while GAMuT is advantageous for polygenic trait analysis.
• These findings aid in selecting appropriate methods for genetic association studies involving multiple phenotypes.