EpipwR

EpipwR estimates statistical power for epigenome-wide association studies (EWAS) with continuous or binary outcomes using a quasi-simulation approach guided by empirical EWAS datasets.

Key Features:

• Quasi-simulation approach: Data is generated only for CpG sites with methylation associated with the outcome, while P-values are generated directly for those with no association.
• Outcome flexibility: Supports power analysis for both continuous and binary outcomes.
• Multiple statistical tests: Can be run with a variety of standard statistical tests.
• Error rate control: Controls for either a false discovery rate or a family-wise type I error rate.
• Empirical dataset guidance: Uses reference datasets of empirical EWAS to guide the data generation process.
• Computational efficiency: Keeps computational cost low while determining power at specific sample sizes.

Scientific Applications:

• EWAS power estimation: Aids researchers in planning epigenome-wide association studies by estimating statistical power at specific sample sizes.
• Study design planning: Expands power calculation functionality beyond case-control studies to additional study designs.
• Method comparison: Enables comparison of power estimation accuracy against existing case-control alternatives.

Methodology:

EpipwR uses a quasi-simulated approach where data is generated only for CpG sites with methylation associated with the outcome, while P-values are generated directly for those with no association. Reference datasets of empirical EWAS guide the data generation process.