Estimating DNA methylation potential energy landscapes from nanopore sequencing data

Jordi Abante,Andrew P Feinberg,John Goutsias,Sandeep Kambhampati

doi:10.1038/s41598-021-00781-x

Abstract

High-throughput third-generation nanopore sequencing devices have enormous potential for simultaneously observing epigenetic modifications in human cells over large regions of the genome. However, signals generated by these devices are subject to considerable noise that can lead to unsatisfactory detection performance and hamper downstream analysis. Here we develop a statistical method, CpelNano, for the quantification and analysis of 5mC methylation landscapes using nanopore data. CpelNano takes into account nanopore noise by means of a hidden Markov model (HMM) in which the true but unknown (“hidden”) methylation state is modeled through an Ising probability distribution that is consistent with methylation means and pairwise correlations, whereas nanopore current signals constitute the observed state. It then estimates the associated methylation potential energy function by employing the expectation-maximization (EM) algorithm and performs differential methylation analysis via permutation-based hypothesis testing. Using simulations and analysis of published data obtained from three human cell lines (GM12878, MCF-10A, and MDA-MB-231), we show that CpelNano can faithfully estimate DNA methylation potential energy landscapes, substantially improving current methods and leading to a powerful tool for the modeling and analysis of epigenetic landscapes using nanopore sequencing data.

Highlights

High-throughput third-generation nanopore sequencing devices have enormous potential for simultaneously observing epigenetic modifications in human cells over large regions of the genome
This additional step allows CpelNano to account for nanopore noise and is carried out via a data-generative model expressed in terms of an Ising model for the methylation landscape and emission probabilities computed by N anopolish[4]
In contrast to existing methods of methylation analysis, CpelNano addresses this problem by using a data-generative hidden Markov model (HMM) that employs a previously introduced Ising model to characterize the true DNA methylation state as a “hidden” state and appropriate emission probabilities, computed via Nanopolish[4], to account for the presence of noise

Summary

Introduction

High-throughput third-generation nanopore sequencing devices have enormous potential for simultaneously observing epigenetic modifications in human cells over large regions of the genome. CpelNano models the hidden state through a previously developed parametric model for noiseless data, which leverages an Ising-like correlated potential energy landscape (CPEL) model[13,14] that is consistent with methylation means and pairwise correlations at each CG dinucleotide (CpG site) This model, which has been successfully used for studying the effect of DNA methyltransferase activity in human embryonic stem cells[17] and dysregulation of epigenetic landscapes in cancer[18,19], takes into account evidence suggesting that the likelihood of a given CpG site to be methylated strongly depends on the fraction of CpG sites in a local neighborhood, as well as on the methylation status at nearby CpG sites whose influence diminishes as their nucleotide distance from the given CpG site increases. Simulations and real data analysis demonstrate the accuracy, effectiveness, and superiority of the proposed statistical method, and show that it can provide a comprehensive and robust framework for the statistical analysis of epigenetic information using nanopore sequencing

Methods

Results

Conclusion