NEW YORK – A trio of methods based on nanopore sequencing are enabling researchers to find open chromatin regions in eukaryotic genomes, along with certain methylation and nucleosome positioning patterns.
"We now have the prospect for multi-omics at the single-molecule level," said Georgi Marinov, a postdoc in William Greenleaf's lab at Stanford University, who developed one of the new methods, single-molecule long-read accessible chromatin mapping sequencing (SMAC-seq).
Another method, methyltransferase treatment followed by single-molecule long-read sequencing (MeSMLR-seq), from Kin Fai Au's lab at Ohio State University, was published in Genome Research in June 2019, following a BioRxiv preprint earlier that year.
A third method, dubbed nanopore nucleosome occupancy and methylome sequencing (nanoNOMe-seq), from Winston Timp's laboratory at Johns Hopkins University, was described in a BioRxiv preprint first posted in December 2018 and updated in February 2019 and is under revision at a journal.
All three methods are similar in principle. Nucleosomes and DNA-binding proteins protect DNA from being modified by DNA modification enzymes such as methyltransferases, Au explained, allowing accessible and inaccessible genomic regions to be mapped. Au's and Timp's teams used only one methyltransferase, which makes GpC m5c modifications, while Greenleaf's team used two additional enzymes, for CpG m5c and non-specific m6A modifications.
All three proof-of-concept studies build on the nucleosome occupancy and methylome sequencing method for analyzing open chromatin (NOMe-seq), developed by researchers at the University of Southern California Keck School of Medicine, by leveraging Oxford Nanopore Technology's platform for reading DNA sequence and DNA modifications simultaneously.
Like NOMe-seq, MeSMLR-seq only focuses on GpC methylation. But the SMAC-seq method takes advantage of the Oxford Nanopore technology's ability to read GpC, CpG, and m6A modifications. Timp's team was the only group to demonstrate their method in human cells (the others used yeast, which does not have endogenous methylation) and chose the GpC modification to avoid "messing with the native [CpG] methylation signal," Timp said.
Greenleaf's lab also developed a short-read sequencing method for analyzing open chromatin, called assay for transposase-accessible chromatin (ATAC-seq,) but SMAC-seq offers a different view of the genome.
"With ATAC-seq you are enriching" for open chromatin, Marinov said, which SMAC-seq does not. His co-first author, Zohar Shipony, added that the fragmentation step in preparing the short-read sequencing library loses links between different regions, including enhancers and promoters. "You're looking at accessible and non-accessible regions at the same time," he said.
Au said that the enzyme his method uses, called M.CviPI, is almost 100 percent efficient, while the m6A methyltransferase EcoGII, available from New England Biolabs, was only 50 percent efficient. Marinov denied this, saying the enzyme was more efficient and that efficiency increased with incubation time. Additionally, Au and Timp said the m6A signal is weaker than the GpC signal with Oxford Nanopore's platform.
But using the m6A nonspecific methyltransferase "greatly increases your resolution," Marinov said, because it modifies every fourth base, on average, while the GpC enzyme modifies every thirtieth, on average. He added that SMAC-seq can work with any dense, non-specific modification.
The different types of data captured all at once will allow researchers to start associating different features with each other. Coupled with long reads, the single-molecule information can "show us how multiple adjacent genes, regulatory elements and different epigenetics layers — methylation, nucleosome occupancy, and chromatin accessibility — cooperate," Au said.
Shipony said their paper provides an example that associates certain regulatory elements with stretches of ribosomal DNA, which often contain many repeats, making them tough to analyze with short reads. Some regions "are accessible all the time and some are not accessible and not expressing any genes," he said. "We can actually see the regulatory elements that are different between the two groups."
Both the Au and Greenleaf labs are pursuing enhancements to their methods, especially getting the method to work in higher-order eukaryotes, including humans and mice. Au added that Pacific Biosciences' platform is better suited for m6A detection and he's developing a method to take advantage of that.
Marinov and Shipony said they're looking at other epigenetic modifications, but did not specify which ones. They thought about pursuing a patent but decided not to, for now, although they might for future versions, Marinov said.
Timp also said he is not pursuing a patent but added that the software used in each method differs and that his was "unique, as well as the claims we can make with it."
Both Au and Timp said they were glad to see they weren't alone in developing a nanopore-based method for studying chromatin accessibility. "We stood at the tops of different mountains and saw the same beautiful lake," Au said. "We walked through forests and deserts and finally reached the lake … I am so happy that I am not alone enjoying the beautiful sights of the lake."