GWAR

GWAR implements robust statistical tests and meta-analysis procedures for genome-wide association studies (GWAS) to maximize analytical power while maintaining nominal type I error rates.

Key Features:

• Robust statistical tests: Implements the Cochran-Armitage trend test (CATT) under recessive, additive, and dominant genetic models, the Maximum Efficiency Robust Test statistic (MAX), and the MIN2 statistic.
• Asymptotic null distributions: Computes asymptotic null distributions for the MAX and MIN2 statistics using numerical integration to improve computational efficiency and accuracy.
• Type I error control: Maintains nominal type I error rates while maximizing analytical power in association testing.
• Meta-analysis capabilities: Performs fixed- and random-effects meta-analysis on summary data with weights equal to the reciprocal of combined cases and controls.
• Stata implementation: Implemented as a program within the Stata environment.

Scientific Applications:

• Association testing under model uncertainty: Detects genetic associations when the underlying genetic model (recessive, additive, or dominant) is unknown.
• GWAS meta-analysis: Synthesizes summary data across multiple studies using fixed- and random-effects models to increase statistical power and generalizability.
• Control of false positives: Provides analyses that preserve nominal type I error rates in genome-wide association studies.

Methodology:

Implements CATT (recessive, additive, dominant), MAX and MIN2 statistics, derives asymptotic null distributions for MAX and MIN2 via numerical integration, and supports fixed- and random-effects meta-analysis using summary-data weights equal to the reciprocal of combined cases and controls.