Hidden Markov Modelling Reveals Neighborhood Dependence of Dnmt3a and 3b Activity

Abstract

DNA methylation is an epigenetic mark whose important role in development has been widely recognized. This epigenetic modification results in heritable information not encoded by the DNA sequence. The underlying mechanisms controlling DNA methylation are only partly understood. Several mechanistic models of enzyme activities responsible for DNA methylation have been proposed. Here, we extend existing Hidden Markov Models (HMMs) for DNA methylation by describing the occurrence of spatial methylation patterns over time and propose several models with different neighborhood dependences. Furthermore, we investigate correlations between the neighborhood dependence and other genomic information. We perform numerical analysis of the HMMs applied to comprehensive hairpin and non-hairpin bisulfite sequencing measurements and accurately predict wild-type data. We find evidence that the activities of Dnmt3a and Dnmt3b responsible for de novo methylation depend on 5' (left) but not on 3' (right) neighboring CpGs in a sequencing string.

Generalized Method of Moments Estimation for Stochastic Models of DNA Methylation Patterns

“…Data from ESCs with Dnmt3a/b DKO, for instance, gives ψ L ≈ ψ R ≈ 1 when fitted to this model because maintenance through Dnmt1 occurs independent of the state of neighboring CpGs. In contrast, calibration of the model to Dnmt1 KO data shows a clear dependence of the methylation activity of Dnmt3a/b to the left [17].…”

“…We focus on Dnmt1 KO data, i.e. only Dnmt 3a/b is active, since previous findings suggest, that in general only Dnmt 3a/b shows a dependence on the left neighbor, while Dnmt1 acts independent of the neighborhood [17].…”

“…A similar model has been used to gain insights into the detailed molecular mechanisms underlying passive and active demethylatyion [14]. Moreover, for genome-wide data, parameter values that describe the efficiency of epigenetic modifications in such models can be clustered and correlated with data from enrichment analysis [13,17].…”

“…Several mechanistic models have been proposed that consider methylation patterns of a number of successive CpGs and their spatial relationships [4,15,22]. Here, we consider a spatial extension of the models considered in [2,14], which has been proposed recently [17,18]. Its main strength compared to other models is that for each locus, it considers methylation efficiencies and dependency parameters.…”

DNA Methylation May be Involved in the Analgesic Effect of Hyperbaric Oxygen via Regulating FUNDC1

“…Generally, DNMT1, a most abundant isoform in proliferating cells, could maintain the methyltransferase and copy methylation targets with replication forks during replication at sites of DNA repair. DNMT3a and DNMT3b methylate specific DNA targets in response to environmental factors [22,23]. It has been observed that persistent pain can alter expression of DNMTs.…”