GAMP

GAMP performs global analysis of DNA methylation distributions across individuals using high-throughput methylation data such as the Illumina 450k Infinium platform.

Key Features:

• Functional regression with B-spline basis: Approximates each individual's methylation value density or cumulative distribution function (CDF) using B-spline basis functions.
• Spline-coefficient summarization: Summarizes an individual's overall methylation profile through spline coefficients.
• Variance component score test: Tests associations between overall methylation distributions and continuous or dichotomous outcome variables using a variance component score test.
• Correlation accommodation: Accounts for correlations between spline coefficients within the association testing framework.
• Statistical performance: Simulations demonstrate desirable statistical power while controlling type I error rates.
• High-throughput data support: Applies to high-resolution methylation data generated by platforms such as the Illumina 450k Infinium array.

Scientific Applications:

• Global methylation studies: Investigation of global methylation changes linked to environmental factors, clinical outcomes, or experimental conditions.
• Locus-specific discovery: Detection of genome-wide and locus-specific methylation differences, such as differential methylation at LINE1 elements between blood samples from rheumatoid arthritis patients and healthy controls.
• Disease-associated epigenetics: Identification of epigenetic alterations associated with human hepatocarcinogenesis in the context of alcohol abuse and hepatitis C virus infection.

Methodology:

GAMP approximates individual methylation densities or CDFs using B-spline functional regression to obtain spline coefficients and evaluates associations with continuous or dichotomous outcomes via a variance component score test that accounts for correlations among coefficients.