Multisite Population Epigenetic Model of Passive Demethylation

TL;DR

This paper introduces a mathematical model for passive DNA demethylation in dividing cell populations, revealing how site correlations and selection influence methylation heterogeneity and providing a baseline to identify active demethylation mechanisms.

Contribution

The model uniquely incorporates site-site correlations and population selection effects, advancing understanding of passive methylation dynamics and aiding in distinguishing active mechanisms.

Findings

Passive demethylation leads to site-site methylation correlations.

Selection can stabilize heterogeneous methylation states.

States inaccessible by passive mechanisms indicate active processes.

Abstract

The methylation of DNA regulates gene expression. On cell division the methylation state of the DNA is typically inherited from parent to daughter cells. While the chemical bond between the methyl group and the DNA is very strong, changes to the methylation state do occur and are observed to occur rapidly in response to external stimulus. The loss of methylation can be active where enzyme physically breaks the bond, or passive where on cell division the newly constructed strand of DNA is not properly inherited. Here we present a mathematical model of single locus passive demethylation for a dividing population of cells. The model describes the heterogenity in the population expected from passive mechanisms. We see that even when the site specific probabilities of passive demethylation are independent, conservation of methylation on the inherited strand gives rise to site-site correlations of the methylation state. We then extend the model to incorporate correlations between sites in the locus for demethylation rates. Biologically, correlations in demethylation rates might correspond to locus wide changes such as the inability of methyltransferase to access the locus. We also look at the effects of selection on the multicellular population. The model of passive demethylation not only provides a tool for measurement of parameters in loci-specific cases where passive demethylation is the dominant mechanism, but also provides a baseline in the search for active mechanisms. The model tells us that there are states of methylation inaccessible by passive mechanisms. Observation of these states constitutes evidence of active mechanisms, either de novo methylation or enzymatic demethylation. We also see that selection and passive demethylation combined can give rise to a stable heterogeneous distribution of gene methylation states in a population.