A recently published study has found that epigenomic markers produce a better signal-to-noise ratio than some researchers had believed, a discovery that could help convince hesitant diagnostics researchers to investigate the epigenomics space.
The findings, derived from the first major data release of the Human Epigenome Project, could also take epigenomic markers one step closer to becoming components in pharmacogenomics applications.
The Human Epigenome Project is a public-private consortium involving the German company Epigenomics, the Wellcome Trust Sanger Institute, and the French Centre National de Génotypage. Its goal is to examine differences in DNA methylation among 12 types of tissue samples from 43 human subjects. The HEP kicked off immediately after a pilot project examining chromosome 6 ended in October 2003.
DNA methylation is a chemical modification of CpG DNA sequences within or near genes that usually shuts them off. So far, HEP has compiled data on 1.9 million CpG sites.
None of the research findings, which were published in the Oct. 29 issue of Nature Genetics, directly addresses pharmacogenomic issues. But so far, the research has supported the widely shared belief among life scientists that these epigenomic changes play a role in regulating which genes are active in specific tissues.
And after the HEP and other projects have established the so-called baseline methylation profiles for specific tissues, researchers can compare them to their own results from studies of disease or drug effects.
Thus, the HEP represents a potential boon to researchers who are interested in using epigenomics as diagnostic biomarkers. And some results reported in the study show that methylation profiles of specific tissues do not appear to vary significantly between samples from people of different genders or ages.
“The less noise that you have between different populations [and] different age groups … [and] the [fewer] differences you observe, the more likely it is to find a biomarker that works across a population to monitor a certain property that is generic to a subpopulation,” such as one with a particular disease, said Kurt Berlin, Epigenomics’ chief scientific officer, in an interview this week.
Other factors uncovered by the study also seem to increase the usefulness of methylation biomarkers. “Our data show DNA methylation to be stable, specific, and essentially binary (that is, on or off) – all key hallmarks of informative clinical markers,” Stephan Beck, the project leader at the Wellcome Trust Sanger Institute, said in a statement last week.
As it gradually maps the methylation of chromosomes beyond numbers 6, 20, and 22, the HEP is showing researchers what they should examine and what to ignore. “It certainly helps in the design of those regions that are likely candidates to provide biomarkers that would also be useful for pharmacogenomic” studies, said Berlin.
Berlin and colleagues in the HEP have so far located methylation-variable positions, or MVPs, associated with 873 genes. “If those [regions] are not differentially methylated [in different tissues] in those 43 different [patient] samples, then we assume that it’s a position that’s always methylated or unmethylated, and therefore uninteresting for further research,” he said.
The project will not investigate every CpG DNA sequence in the three chromosomes, however; the HEP has concentrated its effort on the regions surrounding those 873 genes, and their exons and introns, as well as so-called CpG islands in evolutionarily conserved regions.
Many of these islands are located in non-coding DNA, but there is still the possibility that the HEP is overlooking some important regulatory epigenomic regions in junk DNA.
Epigenomics’ role in the collaboration has been to provide technology and data interpretation and the company has the right to license data coming out of the collaboration for commercial development, according to Epigenomics CFO Oliver Schacht.
“The less noise that you have between different populations [and] different age groups … [and] the [fewer] differences you observe, the more likely it is to find a biomarker that works across a population to monitor a certain property that is generic to a subpopulation.”
“We’re focused on cancer diagnostics and pharmacodiagnostics, so as far as those things go we’ve taken a good look and looked at the intellectual property that’s relevant to us, and everything else ought to be in the public domain,” he said.
There will probably be plenty of opportunity for companies looking to sublicense biomarkers from the consortium. “DNA methylation is such a big topic that we’re pretty busy covering the cancer” biomarkers, said Berlin.
Epigenomics has previously announced that it is planning
to launch a tissue-based prostate cancer prognostic assay on an Affymetrix array platform in 2007, and a tissue-based breast cancer prognostic assay soon afterward. The company is also working with Roche Diagnostics on blood-based epigenomic tests for colorectal, breast, prostate, and lung cancer.
The HEP is not related to the Alliance for Human Epigenomics and Disease, which is a larger project of the American Association for Cancer Research, although Schacht credits the HEP with catapulting epigenomic cataloguing into the realm of large academic research.
The AACR’s AHEAD effort may be more informative still for pharmacogenomics researchers. According to documents from its founding Human Epigenome Task Force, AHEAD will examine additional tissue types and diseased tissue, such as tumor tissue, and it will include an analysis of a different kind of epigenomic modification — acetylation of histone proteins.
Histone proteins are DNA-packaging and -regulating elements.
The HEP’s Nature Genetics paper mentions that the AHEAD project is already underway, but spokespeople for the project could not be contacted to confirm or to provide further information before deadline.