ACEt: An R Package for Estimating Dynamic Heritability and Comparing Twin Models.

Area of Science:

• Behavioral Genetics
• Biostatistics
• Quantitative Genetics

Background:

• Estimating dynamic effects of age on genetic and environmental variance components in twin studies is crucial for understanding gene-environment interactions and improving heritability estimation.
• Existing parametric models have limitations, such as predefined function constraints, hindering accurate dynamic variance component analysis.

Purpose of the Study:

• To introduce ACEt, an R package for efficiently estimating dynamic variance components and heritability that vary with age or other moderators.
• To provide a flexible framework for modeling age-dependent heritability using penalized splines in twin studies.

Main Methods:

• Developed ACEt, an R package utilizing penalized splines within an ACE (Additive, Common, Unique) model framework.
• Implemented resampling methods for hypothesis testing of different variance functions (splines, log-linearity, constancy) to validate model assumptions.
• Assessed robustness, type I error rates, statistical power, numerical issues, and computational costs through simulations.

Main Results:

• ACEt demonstrates robustness to spline knot misspecification and provides a refined resolution of dynamic genetic and environmental components.
• Age-specific variance components for body mass index and height in a Finnish twin cohort showed distinct patterns, despite similar overall heritability estimates.
• The package successfully applied to real-world data, revealing age-dependent heritability patterns.

Conclusions:

• The ACEt R package is a valuable tool for exploring age-dependent heritability and performing model comparisons in twin studies.
• It offers a unified and flexible approach to modeling dynamic variance components, overcoming limitations of traditional parametric models.
• ACEt facilitates a more detailed estimation of age-specific heritability and the investigation of gene-environment interactions over the lifespan.