Epigenetic Exploration with Nanopore Sequencing

Nanopore sequencing is an disruptive DNA sequencing technology that operates via a similar principle to a Coulter counter, using a measurement of the current through a hole in a membrane to characterize the sample passing through the hole. For nanopore sequencing, the hole is nanometers in diameter, and the DNA molecule passing through the pore influences the current in a way characteristic of the local base sequence. In the case of the commercialized nanopore instrument from Oxford Nanopore (ONT) , the pore is large enough only for a single stranded DNA molecule; multiple bases are within the central constriction of the pore at a given time (“k-mers”) and have a significant influence on the current.

Nanopore sequencing has enormous potential in epigenetic applications; unlike traditional sequencing-by-synthesis technologies, it can distinguish covalently modified nucleotides directly through their modulation of the electrolytic current. We can take advantage of the long read lengths (>10kb) generated by nanopore sequencing to precisely call methylation patterns, to obtained phased methylation information across the genome. We have already demonstrated the accuracy and some of the promise of nanopore sequencing in calling 5-methylcytosine using a hidden Markov model.

We have now trained our model on dam-mediated adenosine methylation, which has applications in bacterial epigenetics as well as exogenous labeling for other components of human epigenetics. N6-mA can be measured simultaneously with 5mC, providing two distinct epigenetic signatures along ~10kb single DNA molecules. To that end, we sequenced both completely unmethylated, fully dam-methylated (N6-mA) E. Coli, and dam (N6-mA) and M.SssI (5-mC) methylated E. Coli genomic DNA, where dam methylates in a 5’-GATC-3’ context and M.SssI in a 5’-CG-3’ context. Using the same hidden Markov model method, we trained to distinguishing methylated adenosine from unmethylated adenosine. We applied this model on cells treated by a damID protocol, a technique to distinguish lamin-associated domains (LADs). The damID uses genetically modified cells that fuses the dam enzyme to lamina proteins which in part make up nuclear membrane, resulting in adenosine methylation only at the nuclear periphery. Applying this model we can phase nuclear positioning information (LADs) with DNA cytosine methylation (5-mC) for the first time in long reads.