NEW YORK (GenomeWeb) – Seeking a lower-cost, higher-throughput alternative to next-generation sequencing-based protocols for profiling 5- hydroxymethylcytosine (5hmC), a team of researchers has developed a new microarray-based approach based on one of the most widely used chips in epigenetic studies.
As detailed in a recent Methods paper, the approach, called oxBS-450K, relies on a new chemistry to profile 5hmC, an epigenetic modification thought to contribute unique regulatory functions to the epigenome.
According to the authors, the protocol for the Illumina Infinium HumanMethylation450K BeadChip relies on sodium bisulfite conversion of DNA, which discriminates between methylated and unmethylated cytosine bases within the genome, but cannot differentiate between 5hmC and 5-methylcytosine, the epigenetic modification traditionally surveyed using the array.
As such, most profiling of 5hmC to date has been done using sequencing-based approaches. However, as they wrote in the paper, sequencing remains "costly and time-consuming," creating the need for an alternative approach.
"We felt a new method was needed to profile 5hmC with single-base resolution on a genome-wide scale that was less costly and time-consuming to perform and analyze than the current sequencing-based methods available," lead author Sabrina Stewart told BioArray News in an email this week. "The 450K BeadChip has become an incredibly useful tool for DNA methylation studies, so it was hoped it could do the same for hydroxymethylation analysis."
Stewart and colleagues at University College London worked with researchers at the University of Cambridge and the Scottish Center for Regenerative Medicine to develop the new approach, which relies on oxidative bisulfite, a new chemistry developed in the lab of Shankar Balasubramanian, a University of Cambridge professor who helped to develop the Solexa technology at the core of Illumina's offerings, as well as oxBS-seq, a sequencing application based on the new chemistry.
Balasubramanian, a co-author on the new Methods paper, started a company called Cambridge Epigenetix to commercialize the oxBS-seq application under the name TrueMethyl.
Still, given the "higher cost and extensive analysis needed for sequencing data, which makes the technology out of reach for many researchers," Stewart and colleagues decided to apply the chemistry to a protocol involving the 450K, which interrogates 480,000 CpG sites in the human genome, and has been widely adopted for epigenetic studies. A PubMed search of the term "HumanMethylation450," for example, yields 115 papers that have referenced the chip.
"The Illumina 450K BeadChips provide a relatively simple, fast, and cost-effective way of generating genome-wide methylation profiles in comparison to alternative sequencing-based methods, so they've become very popular for DNA methylation research, especially when analyzing large cohorts of samples," said Stewart. "There are also ever-increasing numbers of data analysis tools being made available specifically for the 450K array, which have made the whole process simpler and more accessible to researchers who may not have a strong background in bioinformatics," she noted.
In the paper, the authors describe both low- and high-input protocols for genome-wide profiling of 5hmC by coupling the oxBS chemistry and the Infinium 450K BeadChip.
According to the authors, oxBS involves a selective oxidation step prior to bisulfite conversion of genomic DNA that converts 5hmC residues to 5-formylcytosine. The resulting 5fC behaves similarly to unmodified cytosine under bisulfite conditions, "allowing only true 5mC positions to be detected as cytosine after oxidative bisulfite conversion and PCR amplification of the DNA sample," they wrote.
Subtraction of oxBS-generated methylation profiles from standard BS-only methylation profiles therefore allows for the detection of hydroxymethylated cytosine positions within the genome. Though traditional bisulfite conversion leads to the potential overestimation of 5mC levels in a sample because of the presence of 5hmC, the researchers' oxBS-450K method generates a more accurate 5mC profile by removing 5hmC.
In the paper, they demonstrated their approach by using the Infinium 450K to successfully detect 5hmC in three distinct genomic DNA samples, comparing BS- and oxBS-450k datasets, and validating the reported 5hmC signals using mass spectrometry and pyrosequencing.
"It definitely has the potential to help researchers who are interested in analyzing 5hmC alongside 5mC in their samples, but it will also be useful for researchers who may want to continue to focus on traditional 5mC analysis but will now be able to remove the confounding factor of 5hmC," Stewart said of the new approach.
"5hmC cannot be distinguished from 5mC under standard sodium bisulfite conversion of DNA and so contributes to the overall methylation signal in any given sample, potentially leading to misinterpretation of the methylation data," she added.