Next-generation sequencing could play a major role in the National Institutes of Health’s decision to include epigenomics as one of four new “major” initiatives under its updated Roadmap for Medical Research, dubbed Roadmap 1.5.
An epigenetics project was one of only two initiatives slated for “immediate implementation” as a five-year program under the Roadmap, with the other being the Human Microbiome Project [In Sequence 06-19-07]. The NIH announced its Roadmap 1.5 last month.
While the NIH hasn’t called for a formal human epigenome project under its revamped Roadmap — and the agency has not yet disclosed funding for any of its Roadmap 1.5 initiatives — some researchers said the NIH’s new stance could propel independent epigenomic projects.
Rob Martienssen of Cold Spring Harbor Laboratory, a co-founder of the American Association for Cancer Research’s Human Epigenome Task Force, told In Sequence via e-mail that the NIH’s recognition of the field as a Roadmap initiative could help accelerate a pilot project the AACR group has proposed called the Alliance for the Human Epigenome and Disease, or AHEAD.
“With the announcement of the Roadmap, I personally think this pilot phase could be completed in a short time, allowing the full-blown project to advance substantially,” he said.
The AACR Task Force published a paper in Cancer Research in 2005 laying out a “blueprint” for a full-scale Human Epigenome Project.
While that publication was heavily weighted toward so-called ChIP-chip or chromatin immunoprecipitation-on-chip technology, methylation arrays, and serial expression of gene expression, times have changed, Martienssen said.
“High-throughput sequencing has revolutionized the construction of epigenomic maps,” he said. “The products of chromatin immunoprecipitation can be sequenced directly and matched to the genome. Similarly, treatment of the genome with methylation-dependent and -independent enzymes, or with bisulphate, allows sequencing of the methylated and unmethylated portions of the genome.”
Martienssen noted that tiling arrays still offer lower cost, easier availability, and more advanced data-handling capabilities than new sequencing technologies, and will therefore likely be used in the “short term” for epigenomics, “eventually high-throughput sequencing is likely to provide epigenomic maps of unrivalled quality and resolution,” he said.
A recent study supports that belief. Last week, a team of researchers from the Broad Institute and Massachusetts General Hospital published a paper in Nature describing the use of ChIP-seq – chromatin immunoprecipitation followed by single-molecule sequencing – to create genome-wide chromatin maps for embryonic stem cells and two cell types derived from them.
The researchers, who used the Illumina Genome Analyzer, noted in their paper that the approach offered a number of benefits over ChIP-chip, which “suffers from inherent technical limitations.”
Bradley Bernstein, an associate professor of pathology at Mass General and the Broad and a co-author on the paper, told In Sequence that ChIP-chip requires more DNA per experiment than ChIP-seq — on the order of micrograms as opposed to nanograms — which can make a large-scale experiment using arrays much more expensive. In addition, he said, there is the risk of cross-hybridization with a chip-based approach, which makes it difficult to characterize repeats and alleles.
“The array can’t discriminate between things if there is a one-base difference,” he said.
“The problem with sequencing has always been that you couldn’t get enough throughput,” said Bernstein. “The challenge has been that you actually need to sequence millions of different fragments of DNA and even some of the relatively recent technologies could only sequence a few hundred thousand fragments.”
Now, however, the single-molecule sequencing has enabled the “transformative step” of sequencing up to 40 million fragments in a single run, Bernstein said. “So that completely changes what we’re able to do. Now we can get genome-wide data by sequencing out the DNA in the chromatin preparations.”
“Now we can get genome-wide data by sequencing out the DNA in the chromatin preparations.”
Even more significant for the purposes of a large-scale epigenome project, however, is the fact that the ChIP-seq approach uses much less sample than previous methods, which will enable researchers to study much smaller populations and rare sample sets, Bernstein said.
“Most of the studies that have been done so far have focused on cell lines that can expand to very high numbers, but the real benefit of an epigenome project would be to look in vivo at the true state of the different lineages and how they vary in terms of their epigenetic or chromatin state in vivo,” he said. By comparison, in vitro studies can cause “really widespread changes in epigenetic markings and chromatin state.”
Bernstein estimated that typical ChIP-chip studies require between 10 million and 50 million cells, whereas ChIP-seq takes that down to ”about half a million cells and [we’re] trying to push that lower.”
A pilot Human Epigenome Project is already underway in Europe, led by a public/private consortium that includes researchers from the Wellcome Trust Sanger Institute and the firm Epigenomics.
In October, the consortium published in Nature the results of a pilot study that used bisulfite DNA sequencing to profile the methylation patterns for three human chromosomes.
Stephan Beck, principal investigator of the epigenome effort at the Sanger Institute, said that bisulfite sequencing using Sanger sequencing, while highly sensitive, is "prohibitively expensive" for a full-scale epigenome project and that his team has been evaluating different platforms for future work.
The team is currently using an array-based platform, which is cheaper than the bisulfite/Sanger method, but "the jury is still out" regarding quality, he said.
Beck said that his group is also "planning" to look again at bisulfite and ChIP sequencing on a next-generation platform, but has not yet decided on what type of instrument it would use.
The NIH's decision to fund future epigenomic studies under its Roadmap program is "perfect timing," Beck said, because the field has not yet been able to attract funding at the same level of other large-scale, high-profile genomic projects.
Epigenomics has been "moving very slowly," he said, and has needed "either a boost in technology or much more funding." Now, he noted, "it looks like those two things are coming together: There is more funding that will hopefully become available, and there is a lot of new technology becoming available."
CSHL’s Martienssen said he expects that any NIH-funded epigenomic project would be coordinated with the ongoing European project. He and others expect the NIH to begin issuing requests for applications under the Roadmap initiative in the fall.
— Julia Karow contributed to this article