A Bayesian model for the analysis of transgenerational epigenetic variation

Area of Science:

• Quantitative genetics
• Epigenetics
• Statistical modeling

Background:

• Epigenetics is a growing field, but its contribution to phenotypic variation is not fully understood.
• Estimating variance components in quantitative genetics relies on resemblance between relatives.
• Previous work modeled resemblance using epigenetic variance and a reset coefficient for epigenetic mark dissipation.

Purpose of the Study:

• To develop a Bayesian mixed model methodology for estimating epigenetic variance from genealogical and phenotypic data.
• To introduce a T matrix for epigenetic relationships dependent on a reset coefficient.
• To provide a Gibbs sampler for estimating model unknowns and assess transgenerational epigenetic variability.

Main Methods:

• Developed a Bayesian mixed model incorporating an epigenetic relationship matrix (T).
• Created a procedure for calculating the inverse of the T matrix (T^-1).
• Employed a Gibbs sampler algorithm for posterior estimation of model parameters.

Main Results:

• The proposed methodology successfully estimated population parameters in simulated datasets.
• Analysis of a beef cattle dataset revealed a very low posterior estimate of transgenerational epigenetic variability.
• Model comparison favored a model excluding epigenetic variance for the cattle data.

Conclusions:

• The Bayesian methodology provides a robust framework for estimating epigenetic variance.
• Transgenerational epigenetic variability appears minimal in the studied beef cattle population.
• Further research is needed to clarify the precise role of epigenetics in phenotypic variation.